Clinical ophthalmic oncology : retinoblastoma [3 ed.] 9783030111236, 3030111237, 9783030111243, 3030111245

Citation preview

Clinical Ophthalmic Oncology Retinoblastoma Jesse L. Berry Jonathan W. Kim Bertil E. Damato Arun D. Singh  Editors Third Edition

123

Clinical Ophthalmic Oncology

Jesse L. Berry  •  Jonathan W. Kim Bertil E. Damato  •  Arun D. Singh Editors

Clinical Ophthalmic Oncology Retinoblastoma Third Edition

Editors Jesse L. Berry Associate Professor of Ophthalmology Children’s Hospital Los Angeles The USC Roski Eye Institute Keck School of Medicine Los Angeles, CA USA Bertil E. Damato University of Oxford Nuffield Department of Clinical Neurosciences John Radcliffe Hospital Oxford UK

Jonathan W. Kim Children’s Hospital Los Angeles The USC Roski Eye Institute Keck School of Medicine Los Angeles, CA USA Arun D. Singh Department of Ophthalmic Oncology Cole Eye Institute, Cleveland Clinic Cleveland, OH USA

ISBN 978-3-030-11122-9    ISBN 978-3-030-11123-6 (eBook) https://doi.org/10.1007/978-3-030-11123-6 Library of Congress Control Number: 2019934958 © Springer Nature Switzerland AG 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

Ophthalmic tumors are rare and diverse, so their diagnosis can be quite complex. Treatment usually requires special expertise and equipment and, in many instances, is controversial. The field is advancing rapidly, because of accelerating progress in tumor biology, pharmacology, and instrumentation. Increasingly, the care of patients with an ocular or adnexal tumor is provided by a multidisciplinary team, consisting of ocular oncologists, general oncologists, radiotherapists, pathologists, psychologists, and other specialists. For all these reasons, we felt that there was a need for the new edition of the textbook providing a balanced view of current clinical practice. Although each section of Clinical Ophthalmic Oncology now represents a standalone volume, each chapter has a similar layout with boxes that highlight the key features, tables that provide comparison, and flow diagrams that outline therapeutic approaches. The enormous task of editing a multi-author, multivolume textbook could not have been possible without the support and guidance by the staff at Springer: Caitlin Prim, Melanie Zerah, ArulRonika Pathinathan, and Karthik Rajasekar. Michael D.  Sova kept the pressure on to meet the production deadlines. It is our sincere hope that our efforts will meet the high expectation of the readers. Los Angeles, CA, USA Los Angeles, CA, USA Oxford, UK Cleveland, OH, USA

Jesse L. Berry Jonathan W. Kim Bertil E. Damato Arun D. Singh

v

Acknowledgments

I want to acknowledge my friend and mentor, Linn Murphree; my grandmother, Jeannette, for always believing in me; my husband, Paul, and our growing family. (JB) I want to acknowledge and thank my teachers, A. Linn Murphree, Bertil Damato, and David Abramson, for their wonderful mentorship over the years. I would also like to thank my parents, Heja and Jinku, for inspiring a young boy to become a physician. To Diana and Devin, I dedicate all of my work here and forever to both of you. (JWK) To my family, Frankanne, Erika, Stephen, and Anna. (BED) To my parents who educated me beyond their means, my wife Annapurna, and my children, Nakul and Rahul, who make all my efforts worthwhile. (ADS)

A. Linn Murphree, MD, Professor of Ophthalmology and Pediatrics at the Keck School of Medicine, University of Southern California (USC), and former Director of the USC Ocular Oncology Service and the Children’s Hospital Los Angeles (CHLA) Retinoblastoma Program

vii

Acknowledgments

viii

Following his training as a Fulbright Fellow in Human Genetics at the University of Copenhagen, Dr. Murphree began his medical training at Baylor College of Medicine with an interest in human genetics. He discovered a passion for both ophthalmology and pediatrics in medical school and ­subsequently combined those three interests by focusing his career on ophthalmic genetic diseases including retinoblastoma. Dr. Murphree assumed the position of Division Head, Pediatric Ophthalmology, at CHLA upon completion of his fellowship in pediatric ophthalmology at Johns Hopkins Hospital. With his first NIH grant, he was one of the pioneers in discovering the location of the retinoblastoma gene on chromosome 13 by performing detailed deletion mapping. Subsequently, he developed a clinical referral practice focused on ocular oncology and developed the largest retinoblastoma referral center in the western USA. In addition to the discovery of the retinoblastoma gene, Dr. Murphree’s contributions to the field of pediatric ocular oncology are numerous and groundbreaking. In his clinical practice, Dr. Murphree recognized an unmet need for a wide-field retinal camera to document the intraocular findings associated with retinoblastoma. He recruited a team of engineers and collaborated with optical engineers in private industry to develop the RetCam, which is the most widely used retinal camera in the world to document pediatric retinal abnormalities. Dr. Murphree’s work on systemic chemotherapy in the 1990s caused a paradigm shift in the treatment of intraocular retinoblastoma away from enucleation and external beam radiation. Dr. Murphree also created the International Classification system for retinoblastoma, which is still the most popular method for diagnosing retinoblastoma for clinicians worldwide. He is the author or coauthor of more than 70 major papers on retinoblastoma genetics and treatment. Dr. Murphree’s work over four decades revolutionized the field of retinoblastoma and improved the lives of countless children afflicted with retinoblastoma. Dr. Murphree was the former editor of the Retinoblastoma volume of Clinical Ophthalmic Oncology, and we are indebted to him for his mentorship during the writing of this current edition. He is universally respected in the field of ocular oncology for his ingenuity, expertise, kindness, and generous spirit. As the current editors of the Retinoblastoma sections, we honor his legacy and thank him for all of his previous and current contributions. Jesse L. Berry Jonathan W. Kim Bertil E. Damato Arun D. Singh

Contents

1 Retinoblastoma: Evaluation and Diagnosis����������������������������������   1 Brian Marr and Arun D. Singh 2 Differential Diagnosis of Leukocoria����������������������������������������������  11 Jonathan W. Kim and Arun D. Singh 3 Retinoblastoma: Staging and Grouping����������������������������������������  27 Jesse L. Berry and A. Linn Murphree 4 Retinoblastoma: Incidence and Etiologic Factors������������������������  39 Manuela Orjuela-Grimm, Nakul Singh, Silvia Bhatt-Carreño, and Arun D. Singh 5 Retinoblastoma: An International Perspective ����������������������������  57 Guillermo L. Chantada and Carlos A. Leal 6 Retinoblastoma Tumorigenesis ������������������������������������������������������  67 Rachel C. Brennan and Michael A. Dyer 7 Animal Models in Retinoblastoma Research��������������������������������  79 Thomas A. Mendel and Anthony B. Daniels 8 Retinocytoma or Retinoma ������������������������������������������������������������  99 Randy C. Bowen, Christina Stathopoulos, Francis L. Munier, and Arun D. Singh 9 Retinoblastoma: Genetic Counseling and Testing������������������������ 107 Meghan J. DeBenedictis and Arun D. Singh 10 Retinoblastoma: Treatment Options���������������������������������������������� 119 Jonathan W. Kim, A. Linn Murphree, and Arun D. Singh 11 Retinoblastoma: Focal Therapy: Laser Treatment and Cryotherapy�������������������������������������������������������������������������������������� 141 Jesse L. Berry and A. Linn Murphree 12 Retinoblastoma: Focal Therapies: Brachytherapy ���������������������� 149 Jose J. Echegaray, Arun D. Singh, and Bertil E. Damato 13 Retinoblastoma: Intravenous Chemotherapy ������������������������������ 155 Rachana Shah, Rajkumar Venkatramani, and Rima Fuad Jubran

ix

x

14 Intra-ophthalmic Artery Chemotherapy for Retinoblastoma�������������������������������������������������������������������������� 169 Benjamin C. King, Brian C. Tse, Matthew W. Wilson, and Rachel C. Brennan 15 Retinoblastoma: Intravitreal Chemotherapy�������������������������������� 179 Christina Stathopoulos and Francis L. Munier 16 Retinoblastoma: External Beam Radiation���������������������������������� 193 Kenneth K. Wong, Jesse L. Berry, and Jonathan W. Kim 17 Retinoblastoma: Enucleation���������������������������������������������������������� 205 Jonathan W. Kim and A. Linn Murphree 18 Retinoblastoma: Evolving Therapies �������������������������������������������� 213 Junyang Zhao and Honggai Yan 19 Histopathologic Features and Prognostic Factors������������������������ 221 Patricia Chévez-Barrios, Ralph C. Eagle, and Eduardo F. Marback 20 Orbital Retinoblastoma: Diagnosis and Management ���������������� 239 Bhavna Chawla and Maya Hada 21 Retinoblastoma: Metastatic Disease���������������������������������������������� 249 Ira J. Dunkel and Guillermo L. Chantada 22 Non-ocular Tumors and Other Long-Term Complications���������������������������������������������������������������������������������� 255 Benjamin C. King, Brian C. Tse, Rachel C. Brennan, and Matthew W. Wilson 23 Trilateral Retinoblastoma �������������������������������������������������������������� 265 Jonathan W. Kim and Ira J. Dunkel 24 Screening Children at Risk for Retinoblastoma �������������������������� 271 Dan S. Gombos and Alison H. Skalet 25 Children’s Oncology Group (COG) Trials for Retinoblastoma���������������������������������������������������������������� 275 Dan S. Gombos, Anna T. Meadows, Murali Chintagumpala, Ira J. Dunkel, Debra Friedman, Julie Ann Stoner, Rima Fuad Jubran, and Judith Grob Villablanca 26 Social Aspects, Advocacy and Organizations�������������������������������� 285 Helen Dimaras Index���������������������������������������������������������������������������������������������������������� 297

Contents

Contributors

Jesse  L.  Berry, MD Department of Ophthalmology, USC Roski Eye Institute, Keck School of Medicine of USC, Children’s Hospital Los Angeles, Los Angeles, CA, USA Silvia  Bhatt-Carreño, MPH Department of Epidemiology, Mailman School of Public Health, Columbia University Medical Center, New York, NY, USA Randy  C.  Bowen, MD Department of Ophthalmology, University of Wisconsin, Madison, WI, USA Rachel  C.  Brennan, MD Department of Oncology, St. Jude Children’s Research Hospital, Memphis, TN, USA Guillermo  L.  Chantada, MD  Hemato-oncology Department, Hospital JP Garrahan, Buenos Aires, Argentina Bhavna Chawla, MBBS, MS (Ophthalmology)  Ocular Oncology Service Dr. Rajendra Prasad Centre for Ophthalmic Sciences, New Delhi, India Patricia  Chévez-Barrios, MD Departments of Pathology and Genomic Medicine and Ophthalmology, Weill Cornell Medical College, New York, NY, USA Murali  Chintagumpala, MD Texas Children’s Cancer Center at Baylor College, Houston, TX, USA Bertil E. Damato, MD, PhD, FRCOphth  University of Oxford, Nuffield Department of Clinical Neurosciences, John Radcliffe Hospital, Oxford, UK Anthony B. Daniels, MD, MSc  Department of Ophthalmology and Visual Sciences, Vanderbilt Eye Institute, Nashville, TN, USA Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN, USA Meghan  J.  DeBenedictis, MS, LCGC, MEd  Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA Helen  Dimaras, PhD  Department of Ophthalmic Oncology, The Hospital for Sick Children; The University of Toronto, Toronto, ON, Canada

xi

xii

Ira  J.  Dunkel, MD Department of Pediatrics, Memorial Sloan Kettering Cancer Center, New York, NY, USA Michael  A.  Dyer, PhD Department of Developmental Neurobiology, St. Jude Children’s Research Hospital, Memphis, TN, USA Ralph C. Eagle, MD  Department of Pathology, Wills Eye Hospital Thomas Jefferson University, Philadelphia, PA, USA Jose  J.  Echegaray, MD  Departments of Vitreoretinal Surgery and Ocular Oncology, Bascom Palmer Eye Institute, University of Miami, Miami, FL, USA Debra  Friedman, MD Division of Hematology-Oncology, Vanderbilt University School of Medicine, Nashville, TN, USA Dan  S.  Gombos, MD FACS MD Anderson Cancer Center, The Retinoblastoma Center of Houston (MD Anderson/Texas Children’s/Baylor/ Methodist Hospital), Section of Ophthalmology-Department of Head and Neck Surgery, Houston, TX, USA Maya Hada, MD  Department of Ophthalmology, SMS Medical College & Hospital, Jaipur, Rajasthan, India Rima  Fuad  Jubran, MD, MPH, MA, CM Division of Hematology, Oncology and Blood and Marrow Transplantation, Children’s Center for Cancer and Blood Diseases, Children’s Hospital Los Angeles, Los Angeles, CA, USA Department of Pediatrics, Keck School of Medicine, University of Southern California, Children’s Hospital Los Angeles, Los Angeles, CA, USA Jonathan  W.  Kim, MD Department of Ophthalmology, USC Roski Eye Institute, Keck School of Medicine of USC, Children’s Hospital Los Angeles, Los Angeles, CA, USA Benjamin  C.  King, MD Hamilton Eye Institute, Department of Ophthalmology, University of Tennessee Health Sciences Center, Memphis, TN, USA Carlos  A.  Leal, MD Department of Oncology, Instituto Nacional de Pediatria, Mexico City, Mexico Eduardo  F.  Marback, MD Department of Ophthalmology, Federal University of Bahia, Rio Vermelho, Salvador, Bahia, Brazil Brian  Marr, MD Department of Ophthalmology, Columbia University Medical Center, New York, NY, USA Anna  T.  Meadows, MD The Children’s Hospital of Philadelphia, Philadelphia, PA, USA Thomas A. Mendel, MD, PhD  Department of Ophthalmology and Visual Sciences, Vanderbilt Eye Institute, Nashville, TN, USA

Contributors

Contributors

xiii

Francis  L.  Munier, MD  Department of Ophthalmology, Jules-Gonin Eye Hospital, University of Lausanne, Lausanne, Vaud, Switzerland A. Linn Murphree, MD  Department of Ophthalmology, Children’s Hospital Los Angeles, University of Southern California, Los Angeles, CA, USA USC Roski Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA Manuela  Orjuela-Grimm, MD, ScM Department of Epidemiology, Mailman School of Public Health, Columbia University Medical Center, New York, NY, USA Rachana Shah, MD, MS  Division of Hematology, Oncology and Blood and Marrow Transplantation, Children’s Center for Cancer and Blood Diseases, Children’s Hospital Los Angeles, Los Angeles, CA, USA Department of Pediatrics, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA Arun  D.  Singh, MD Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA Nakul Singh, MS  School of Medicine, Case Western University, Cleveland, OH, USA Alison  H.  Skalet, MD, PhD Casey Eye Institute, Department of Ophthalmology, Oregon Health and Science University, Portland, OR, USA Department of Radiation Medicine, Oregon Health and Science University, Portland, OR, USA Christina Stathopoulos, MD  Department of Ophthalmology, Jules-Gonin Eye Hospital, University of Lausanne, Lausanne, Vaud, Switzerland Julie Ann Stoner, PhD  University of Nebraska Medical Center, Nebraska Medical Center, Omaha, NE, USA Brian  C.  Tse, MD Department of Ophthalmology, University of Miami Bascom Palmer Eye Institute, Miami, FL, USA Rajkumar  Venkatramani, MD, MS Texas Children’s Cancer Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX, USA Judith  Grob  Villablanca, MD Children’s Hospital of Los Angeles, Los Angeles, CA, USA Matthew  W.  Wilson, MD, FACS  Hamilton Eye Institute, Department of Ophthalmology, University of Tennessee Health Sciences Center, Memphis, TN, USA Kenneth  K.  Wong, MD  Department of Ophthalmology, USC Roski Eye Institute, Keck School of Medicine of USC, Children’s Hospital Los Angeles, Los Angeles, CA, USA

xiv

Honggai  Yan, MD Department of Ophthalmology, Beijing Children’s Hospital, Capital Medical University, National Center for Children’s Health, Beijing, China Junyang  Zhao, MD Department of Ophthalmology, Beijing Children’s Hospital, Capital Medical University, National Center for Children’s Health, Beijing, China

Contributors

1

Retinoblastoma: Evaluation and Diagnosis Brian Marr and Arun D. Singh

Historical Background

Clinical Presentation

In 1809 a Scottish surgeon named James Wardrop wrote a monograph where he described a subset of “fungus haematodes” cases distinguishing them from other cases of “soft cancer,” “medullary sarcoma,” or “spongiod inflammation.” He was the first to recognize retinoblastoma (RB) as a discrete tumor arising primarily from the retina [1]. Virchow in 1864 used the name of glioma retinae because of retinoblastoma’s similarity to the intracranial glial tumors. Verhoeff, in 1922, observed the retinal origin and the presence of immature, embryonic cells that formed the tumor and coined the term retinoblastoma. In 1926 the American Ophthalmological Society accepted the term retinoblastoma, and the older terms, such as glioma retinae and fungus haematodes, were abandoned [2]. In 1809 it was the astute clinical observations and descriptions of the disease that made the diagnosis of what we now know as retinoblastoma.

The symptoms of retinoblastoma are most often first detected by a parent or family member directly or occasionally from an abnormal light reflex in a photograph. To a lesser extent, sporadic cases of retinoblastoma are first discovered by a routine pediatric exam or screening, less commonly by pediatric ophthalmologists and rarely incidentally on imaging for other conditions. In the United States and other developed nations, the most common presenting findings in intraocular retinoblastoma are leukocoria or cat’s eye reflex (45%) (Chap. 2), strabismus (25%), inflammatory symptoms (pseudo-preseptal cellulitis) (10%), and poor vision (10%) (Table 1.1) [3]. For several reasons discussed elsewhere in developing nations, retinoblastoma tends to be more advanced at presentation with greater proportion of cases with extraocular disease (Chap. 5). One of the major limitations to prompt treatment of retinoblastoma worldwide

B. Marr (*) Department of Ophthalmology, Columbia University Medical Center, New York, NY, USA e-mail: [email protected] A. D. Singh Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_1

Table 1.1  Presenting features of retinoblastoma (United States) Leukocoria or cat’s eye reflex Strabismus Inflammatory symptoms (preseptal cellulitis) Poor vision Screening due to family history Incidental detection

45% 25% 10% 10% 5% 5%

Based on data from Abramson et al. [14] 1

B. Marr and A. D. Singh

2

is access and availability to healthcare. As retinoblastoma care providers, it is important for us to increase accessibility for our patients into a system that is equipped to treat this condition adequately. Community education and awareness and training of ancillary staff that are able to triage and arrange prioritized evaluations are some of the important components of this approach (Chap. 5).

Misdiagnosis Histopathological studies of eyes enucleated report misdiagnosis rates from 11% to 40%, and clinical studies of referral patterns report misdiagnosis rates from 16 to 53% [3]. This may be attributed to many factors including rare incidence of retinoblastoma, multiple conditions that simulate retinoblastoma, the unfamiliarity of the primary healthcare providers, the age of presentation, and the difficulty in examining children (Chap. 2). Consequently, a thorough and detailed assessment should be done on patients suspected of having retinoblastoma.

Stepwise Evaluation for Retinoblastoma A practical stepwise approach specifically to evaluate a child suspected to have retinoblastoma includes detailed history taking, initial office examination, and focused ophthalmic ultrasonography, followed by examination under anesthesia and neuroimaging, if necessary (Fig. 1.1). This approach is merely a guide that can be modified as needed based upon clinical setting.

History For a child suspected of having retinoblastoma, it is important to examine the patient and family promptly upon referral, and the initial consultation may be performed in an office setting (Table  1.2). The story of how and over what time course the condition was noted, the health-

care professionals that saw the patient, and what was done to the child before they arrived must be recorded. A birth history including the pre- and perinatal history is important. Typically the gestational age at birth, type of delivery, birth weight, and any delivery or pregnancy complications, including infections or medications taken during the pregnancy, are noted. It is also important to inquire if any abnormalities were noted on the eye screening exam after birth or if there were any unusual birthmarks or malformations. The current history should include the child’s health, any medical conditions, and environment including pets, recent trauma, or illness. For retinoblastoma suspects, the family history should include number of siblings, their health and ocular history, and any family medical disorders. It should be noted if there was any poor vision, blindness, or loss of an eye in the family. Both parents should be questioned about their ocular health and examined if no recent dilated exam has been performed. A small subset of parents of children with RB will have evidence of retinoma/retinocytoma and even unknown treated retinoblastoma (Chap. 8) [4].

Initial Examination The initial examination of the child can be started in the office while taking the history, by observing the comfort and behavior of the child, and noting any size, proportion, or facial abnormalities (Table 1.3). It may be possible to observe leukocoria, strabismus, or periorbital swelling and visual behavior before initiating the formal examination. Assessing the vision is dependent on the age of the patient and the amount of cooperation; however, the condition of each eye should be assessed and recorded along with the pupillary response and the presence or absence of heterochromia of the irises. A brief observation of the periorbital tissues, cornea, conjunctiva, and sclera should be performed before administrating dilation drops. Using a direct ophthalmoscope, the pupillary light reflex can be noted in both eyes.

3

1  Retinoblastoma: Evaluation and Diagnosis Fig. 1.1 Stepwise evaluation for retinoblastoma. This approach is merely a guide that can be modified as needed based upon clinical setting

Detailed history

Initial office examination

Focused ophthalmic ultrasonography

Examination under anesthesia+ Ancillary testing

Retinoblastoma

Other diagnosis

Neuro imaging

Counselling Nature of retinoblastoma Genetic aspects (and testing) Available therapeutic options

Initiate treatment

Upon completion of this portion of the examination, drops for pupillary dilation can then be administered (tropicamide 0.5% and ophthalmic phenylephrine 2.5%). It is worth emphasizing that both eyes should be examined in equal detail. The examination of the posterior pole is best done with an indirect ophthalmoscope. Depending on the age, the child may cooperate, or parents may be needed to help secure the patient while lying supine on a table or chair

(Fig.  1.2). Younger children can be swaddled with a blanket or sheet. The goal of the indirect examination at this point is to confirm the suspicion of retinoblastoma and determine whether further evaluation is necessary with an exam under anesthesia (EUA). It may be necessary to place an eyelid speculum in for proper visualization of the posterior pole; appropriate topical anesthesia such as ophthalmic proparacaine 0.5% solution should be administered

4

B. Marr and A. D. Singh

Table 1.2  Elements of medical history in a child suspected of having retinoblastoma Time since onset Prior evaluation

Perinatal history

Personal history

Family history

Duration Prior diagnosis Prior treatment Prior surgical procedure Prior biopsy Pregnancy complications Prematurity Birth weight Type of delivery Use of oxygen Malformations Exposure to pets Recent trauma Systemic illness Genetic disease Blindness Enucleation Amblyopia Retinoblastoma

Table 1.3  Elements of initial examination (office) in a child suspected of having retinoblastoma External examination Facial abnormalities (13q deletion syndrome) Strabismus Periorbital swelling Presence of heterochromia Visual acuity Pupillary response Pupillary light reflex Normal Abnormal Leukocoria absent Leukocoria present Anterior segment May be limited examination Indirect May be limited ophthalmoscopy Ultrasonography Mass Calcification Retinal detachment Other abnormalities

before placing the speculum. A detailed fundus examination with scleral depression may be performed with anesthetic, eyelid speculum, and restraint; however, this is fairly traumatic for both the child and the family and is generally unnecessary if a planned exam under anesthesia is possible.

Fig. 1.2  An indirect ophthalmoscopic examination being performed in an office setting with the mother helping to hold the child

Ophthalmic Ultrasonography A limited ophthalmic ultrasonography can be done in A/B scan mode using a 10 Hz transducer to visualize the presence of a mass, calcification, retinal detachment, or abnormalities of the posterior pole. If retinoblastoma is recognized and further examination is necessary, ideally the child is scheduled for a EUA, and neuroimaging is ordered (MRI of the brain and orbit with and without contrast) to visualize the orbit and posterior portion of the optic nerve and assess for pinealoblastoma (Chap. 22).

Examination Under Anesthesia The type and method of general anesthesia vary depending on institution and availability. Safe anesthesia methods can range from mask anesthesia or laryngeal mask airway (LMA) using inhaled anesthetics, with or without intravenous anesthesia to using intravenous anesthetics alone [5]. As with all anesthesia, children must limit intake of food and liquids before the procedure. Guidelines suggest all food, milk, or formula be discontinued 8  hours prior to the exam. Breast milk is allowed up to 4 hours before the exam and clear liquids up to 2  hours before; however,

1  Retinoblastoma: Evaluation and Diagnosis

requirements vary by institution and are ­determined by the anesthesiologist and type of anesthesia used. Some younger infants require extended observation after anesthesia to be monitored for apnea. Current recommendations are that preterm infants less than 36 weeks must be at least 55  weeks post-conceptual age to go home after anesthesia without extended monitoring; otherwise an overnight stay is recommended. Full-term infants must be 50  weeks post-­ conceptual age to go directly home, and full-term infants between 40 and 50 weeks post-conceptual age require 6  hours of observation before discharge. Family members should be made aware of these recommendations so they can make arrangements for the examination. Once the patient is asleep, a full ophthalmic examination that includes all components of the initial office examination repeated in greater detail of both eyes is performed (Table 1.4). Table 1.4  Elements of initial examination (office) in a child suspected of having retinoblastoma External examination Facial abnormalities (13q deletion syndrome) Strabismus Periorbital swelling Presence of heterochromia Intraocular pressure Corneal diameter Pupillary response Prior to dilation Pupillary light reflex Normal Abnormal Leukocoria absent Leukocoria present Anterior segment Conjunctiva/sclera examination Cornea Anterior chamber Iris Lens Retrolental (anterior) vitreous Indirect Vitreous ophthalmoscopy Optic disk Macula Peripheral retina Pars plana Ultrasonography Mass Calcification Retinal detachment Other abnormalities

5

External Examination The overall appearance of the patient should be assessed by looking at the face for any abnormalities that may aid in diagnosis or that are associated with retinoblastoma such as 13q deletion syndrome. As an example, a patient with 13q deletion syndrome may have hypertelorism, a flat nasal bridge, small mouth and nose, high arched or cleft pallet, micrognathia, and/or microcephaly which may be noted during this part of the examination (Chap. 9).

Anterior Segment Examination Intraocular pressure should be measured using a Schiotz tonometer, tonopen, Perkins tonometer, or pneumotonometer. Substantially elevated intraocular pressure in retinoblastoma patients due to iris neovascularization or angle closure has been associated with higher risk of optic nerve involvement and metastatic disease [6]. Next using a caliper, the horizontal and vertical corneal diameters (CD) are measured. Simulating conditions such as persistent fetal vasculature (PFV) can have significant discrepancies between the eyes (Fig. 1.3), and the eyes with chronically elevated intraocular pressure can have increased corneal diameters. A handheld slit lamp or illuminated magnification system should be used to assess the anterior segment. Care should be taken to look for any shallowing of the anterior chamber, neovascularization of the iris, iris atrophy, cataract, retinoblastoma seeding of anterior segment, or hyphema. It is important to evaluate the conjunctiva and sclera as well as the anterior vitreous and posterior portion of the lens. It may be possible to see the underlying retina or tumor against the posterior portion of the lens or a retrolental mass or persistent tunica vasculosa lentis in simulating conditions. As an example, observation of the blood vessel branching patterns behind the lens can give a clue to their origin and help differentiate certain entities. Retinal vessels will have a branching pattern opening toward the periphery

6

B. Marr and A. D. Singh

a

a

b

b

Fig. 1.3 (a) A patient with persistent fetal vasculature showing the discrepancy between the corneal diameters. (b) A patient with advanced retinoblastoma showing increased corneal diameter and heterochromia from iris neovascularization

Fig. 1.4  Anterior segment photograph of a patient with advanced retinoblastoma (a). Note the branching patterns of the retinal blood vessels toward the periphery of the lens. Anterior segment photograph of the patient with persistent fetal vasculature (b). Note the retrolental vascular mass

of the lens, whereas persistent tunica vasculosa lentis in PFV will have a branching pattern toward the center of the lens, or a retrolental mass will have disorganized vessels (Fig. 1.4).

One eye at a time, the vitreous should be examined for the presence or absence of retinoblastoma seeding, hemorrhage, presence of abnormal vessels, fibrous membranes, inflammatory cells, or other abnormalities. If the optic disk and macula are visible, the size and presence of any abnormalities should be noted. Continued examination of the periphery can be done by working in a clockwise fashion and scleral depressing the ora serrata and then looking along that longitudinal segment to the posterior pole until the whole 360 degrees of the eye is covered. The appearance of retinoblastoma lesions can vary depending upon the size and location of the tumor; smaller tumors are round glazed e­ levations of the retina; as they grow, they acquire large

Posterior Segment Examination Indirect ophthalmoscopy is used to evaluate the fundus. An organized systematic approach to thoroughly assess the posterior pole is recommended to prevent overlooking important findings. This examination can be broken down into four parts to evaluate the vitreous, optic disk, macula, and peripheral retina including pars plana.

7

1  Retinoblastoma: Evaluation and Diagnosis

feeder vessels and have a gray white hew and develop surrounding serous retinal detachments. The larger tumors develop intrinsic calcification and a whiter color with seeding into the subretinal and or the vitreous space. Specifically for retinoblastoma, the size and number of all tumors should be documented noting any associated retinal detachment or subretinal fluid, the presence of subretinal seeds and vitreous seeds, and their location and pattern of distribution incorporated into a detailed fundus drawing (Table 1.4). This information should be used to make group and stage the eyes according to the classification systems (Chap. 3).

a

b

Ancillary Testing Photography It is useful to document both the anterior segment and the posterior segment findings with a photograph. A wide-angle handheld fundus camera is useful for taking photos of the front and back of the eye using different lenses (Fig. 1.5). Fundus photos should be taken at each EUA to aid in assessing the response to treatments. Care should be taken to standardize the orientation and position of the photographs to help with future comparisons. Fluorescein Angiography Fluorescein angiography (FA) can be a useful tool during a EUA to differentiate retinoblastoma from simulating lesions. The FA vascular pattern of retinoblastoma shows normal filling of enlarged dilated vessels diving in and through a hyper- and hypo-fluorescent tumor mass that stains and leaks depending on its size. FA is especially useful in differentiating RB from advanced Coats’ disease. In contrast to RB, Coats’ disease has large dilated telangiectatic vessels that remain in the plane of the retina and have marked areas of peripheral capillary nonperfusion (Fig. 1.6). Ophthalmic Ultrasonography During the EUA it is useful to obtain ultrasound imaging on both eyes to assess the orbit,

Fig. 1.5  Photography of a patient during an examination under anesthesia. The external lens used to photograph the anterior segment (a). The wide-angle fundus lens is used to take photographs of the posterior pole (b)

measure the thickness of lesions, and obtain the axial lengths of the eyes. Historically ultrasound has been useful in the diagnosis and treatment of retinoblastoma by providing information of the size and extent of the disease as well as differentiating it from simulating lesions [7, 8]. Ultrasound can be done in A and/or B scan mode using a 10  MHz transducer to image the posterior pole and visualize the size and location of disease, the presence of a retinal detachment, or extraocular extension. Ultrasound is specifically useful for evaluating lesions inside the eye when there is a limited view with ophthalmoscopy. Larger retinoblastoma lesions have a characteristic appearance on ultrasound because they produce calcium that is easily detected by ultrasound showing

8

B. Marr and A. D. Singh

a

a

b

b

Fig. 1.6  Fluorescein angiograms taken during an exam under anesthesia. A fluorescein angiogram of a patient with retinoblastoma demonstrating irregular vessels within the retina and slower filling vessels within the tumor inferiorly (a). Fluorescein angiogram of a patient with Coats’ disease demonstrating light bulb telangiectasia and peripheral non-perfusion (b)

multiple areas of hyper-reflectivity with acoustic shadowing (Fig. 1.7a).

Ultrasound Biomicroscopy Ultrasound biomicroscopy (UBM) also can be performed during a EUA and is useful in visualizing the pars plana, pars plicata, and ciliary body. In advanced cases, areas of anterior seeding can be detected using the UBM as well as extension of the tumor into the ciliary body or against the lens. This technique is important particularly for cases that are being considered for intravitreal chemotherapy injection (Chap. 15).

Fig. 1.7  Calcification within retinoblastoma. Ultrasonography of an eye with retinoblastoma in B scan mode showing a hyperreflective mass and acoustic shadowing (a). A CT scan of a patient with retinoblastoma demonstrating the intraocular calcification seen within the tumor in the right eye (b)

Electroretinogram An electroretinogram (ERG) has been used to monitor retinal function prior to, during, and after treatment of retinoblastoma particularly with intra-arterial chemotherapy (Chap. 14). It is a useful surrogate for obtaining information about visual potential in preverbal children and the effect of treatment toxicity on retinal function. During the EUA, a photopic 30 Hz flicker can be performed prior to the examination in the standard fashion [9]. It is preferable to perform the ERG before any physical manipulation, ophthalmoscopic examination, or photography is performed because such manipulations can affect the reliability of the readings [10].

9

1  Retinoblastoma: Evaluation and Diagnosis Fig. 1.8 Magnetic resonance imaging (MRI) of a patient with retinoblastoma. A T1-weighted image demonstrating an intraocular mass in the right eye (a). On T2-weighted image, the tumor is darker than the adjacent vitreous (b). A T1-weighted image following administration of contrast demonstrating enhancement of the tumor (c). With fat suppression, enhancement of the tumor is highlighted (d)

a

b

c

d

Neuroimaging Neuroimaging is ordered on all patients diagnosed with retinoblastoma at time of diagnosis to assess the orbits and optic nerves and to screen for pinealoblastoma. Repeat imaging may be performed every 6 months basis for all germline cases up to the age of 6 (+/−1) years for pineal screening (Chap. 23) [11]. Computed tomography (CT) scans historically had been very useful in identifying intraocular calcified lesions of retinoblastoma; however, it is currently not recommended in children with retinoblastoma in order to limit their exposure to ionizing radiation (Fig. 1.7b) [12]. MRI of the brain and orbits with and without contrast is currently the preferred initial study. Intraocular retinoblastoma on T1-weighted images appears hyperechoic compared to vitreous and enhances with contrast. On T2-weighted images, the RB lesions appear hypoechoic compared to vitreous. There should be no significant enhancement of the optic nerves post contrast (Fig. 1.8).

Counseling After taking the detailed history, performing a thorough examination, and reviewing the ancillary studies, a detailed discussion regarding the nature

of retinoblastoma, genetic aspects (and testing), the need for screening of family members and relatives (Chap. 9), and of the available therapeutic options (Chap. 10) can be held with the family and patient so as to devise and initiate a treatment plan [13].

References 1. Kivelä T. 200 years of success initiated by James Wardrop’s 1809 monograph on retinoblastoma. Acta Ophthalmol. 2009;87:810–2. 2. Dunphy EB.  The story of retinoblastoma. Am J Ophthalmol. 1964;58:539–52. 3. Maki JL, Marr BP, Abramson DH.  Diagnosis of retinoblastoma: how good are referring physicians? Ophthalmic Genet. 2009;30:199–205. 4. Shields CL, Shields JA, Baez K, et al.  Optic nerve invasion of retinoblastoma. Metastatic potential and clinical risk factors. Cancer. 1994;73:692–8. 5. Gallie BL, Ellsworth RM, Abramson DH, et al. Retinoma: spontaneous regression of retinoblastoma or benign manifestation of the mutation? Br J Cancer. 1982;45:513–21. 6. Lin BA, Messieha ZS, Hoffman WE. Safety and efficacy of pediatric general anesthesia by Laryngeal Mask Airway without Intravenous Access. Int J Clin Med. 2011;2:328–31. 7. Sterns GK, Coleman DJ, Ellsworth RM. Characterization and evaluation of retinoblastoma by ultrasonography. Bibl Ophthalmol. 1975;83:125–9. 8. Abramson DH, Ellsworth RM. Ancillary tests for the diagnosis of retinoblastoma. Bull N Y Acad Med. 1980;56:221–31.

10 9. Liu CY, Jonna G, Francis JH, et al.  Non-selectivity of ERG reductions in eyes treated for retinoblastoma. Doc Ophthalmol. 2014;128:13–23. 10. Francis JH, Abramson DH, Marr BP, et al. Ocular manipulation reduces both ipsilateral and contralateral electroretinograms. Doc Ophthalmol. 2013;127:113–22. 11. Albert DM.  Trilateral retinoblastoma: insights into histogenesis and management. Surv Ophthalmol. 1998;43:59–70. 12. de Graaf P, Göricke S, Rodjan F, et al. European Retinoblastoma Imaging Collaboration (ERIC).

B. Marr and A. D. Singh Guidelines for imaging retinoblastoma: imaging principles and MRI standardization. Pediatr Radiol. 2012;42:2–14. 13. Skalet AH, Gombos DS, Gallie BL, et al. Screening children at risk for retinoblastoma: consensus report from the American Association of Ophthalmic Oncologists and Pathologists. Ophthalmology. 2018;125:453–8. 14. Abramson DH, Frank CM, Susman M, et al. Presenting signs of retinoblastoma. J Pediatr. 1998;132(3 Pt 1): 505–8.

2

Differential Diagnosis of Leukocoria Jonathan W. Kim and Arun D. Singh

Introduction Leukocoria is the most common presenting sign of intraocular retinoblastoma in developed countries [1]. The asymmetric white pupil light reflex may be noted on photographs, in dimly lit environments by the family, or by a general pediatrician at a well-child visit [2]. An abnormal pupil reflex is also frequently observed in several pediatric ocular conditions including cataract (Fig. 2.1), and it is important to clinically differentiate retinoblastoma from simulating diagnoses (Table  2.1). Directed by the available demographic and historical data, a comprehensive clinical and ultrasound examination in the office is usually sufficient to make the correct diagnosis. Occasionally, an examination under anesthesia may be necessary to distinguish retinoblastoma from simulating conditions, such as Coats’ disease, persistent hyperplastic primary vitreous (PHPV), retinal dysplasia, or astrocytic hamartoma. Clinical findings associated with the commonly diagnosed conditions are summarized in the following section (Table 2.2) [3–5]. J. W. Kim (*) Department of Ophthalmology, USC Roski Eye Institute, Keck School of Medicine of USC, Children’s Hospital Los Angeles, Los Angeles, CA, USA e-mail: [email protected] A. D. Singh Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_2

Fig. 2.1  Leukocoria due to cataract induced by a chronic retinal detachment

It is important to carefully and urgently evaluate any child with leukocoria for the possibility of retinoblastoma, although fortunately many children referred for this complaint will have a normal examination (i.e., pseudo-leukocoria). Commonly, it is the parents who first notice the abnormal or asymmetric pupil reflex in a photograph. The flash from a camera typically causes the eye to appear red, since the pupil does not have time to contract and the camera captures a red reflection from the normal retina. Any condition that blocks the camera’s flash from reaching the retina may produce a unilateral whitish pupil reflex (i.e., photoleukocoria) [2]. However, it 11

12

J. W. Kim and A. D. Singh

Table 2.1  Differential leukocoria

diagnosis

of

childhood

1. Tumors  Retinoblastoma  Medulloepithelioma  Leukemia  Combined retinal hamartoma  Astrocytic hamartoma (Bourneville’s tuberous sclerosis) 2. Congenital malformations  Persistent fetal vasculature (PFV)  Posterior coloboma  Retinal fold  Myelinated nerve fibers  Morning glory syndrome  Retinal dysplasia  Norrie’s disease  Incontinentia pigmenti  Cataract 3. Vascular diseases  Retinopathy of prematurity (ROP)  Coats’ disease  Familial exudative vitreoretinopathy (FEVR) 4. Inflammatory diseases  Ocular toxocariasis  Congenital toxoplasmosis  Congenital cytomegalovirus retinitis  Herpes simplex retinitis  Other types of fetal iridochoroiditis  Endophthalmitis 5. Trauma  Intraocular foreign body  Vitreous hemorrhage  Retinal detachment

should be kept in mind that photoleukocoria does not always indicate an underlying pathologic condition. There are case series of patients with documented unilateral leukocoria on photographs who had normal ocular examinations [6]. This phenomenon has been termed pseudo-leukocoria since the examination did not reveal any pathology. In these cases, the child appears to be fixating 15° off axis (inward deviation), which likely resulted in an abnormal light reflex off the optic nerve in that eye (Fig. 2.2). Therefore, photoleu-

Fig. 2.2  Pseudo-leukocoria noticed on a photograph. Notice unilateral occurrence in the eye that appears to be fixating 15° off axis (inward deviation)

Table 2.2  Differential diagnosis of retinoblastoma: demographics and ultrasonographic features Age of Condition presentation Risk factors Retinoblastoma 90% 3 quadrants of seeding were associated with a worse visual outcome, although it should be noted that with the use of intravitreal chemotherapy instead of external beam radiation therapy for seeding [21, 22], the impact of seeding on visual outcomes is less clear.

 llowing Intraocular Grouping A to Change in Case of Disease Progression Currently there is no provision in the grouping schema for the group assignment of an eye to change with persistent or recurrent disease. During the course of treatment, the clinical features can change including the development of seeding in the setting of a recurrence. Consideration is being given to a concept referred to as event-free ocular survival [EFOS]. Such a term would be analogous to the term “event-free survival” [EFS] commonly used in clinical trials to define the time from study entry until an “event” such as disease progression, tumor relapse, second malignancy, death, or last contact occurs. Event-free ocular survival [EFOS] could define the time from study entry until an ocular “event” such as disease progression that cannot be controlled by local consolidation or last visit occurs. Once an EFOS has occurred, a revised group assignment to reflect the current status of the ocular disease might be considered. Further study is underway regarding a system for recurrent disease.

 linical Application of International C Retinoblastoma Classification Staging the patient with retinoblastoma should have a focus on overall survival as well as ocular survival with both staging of the patient and clinical grouping of the eye (Table 3.5).

36

J. L. Berry and A. L. Murphree

Table 3.5  Application of the international retinoblastoma classification

Clinical scenario Previously untreated; no clinical or imaging evidence of extraocular disease; no family history RE Group D, LE Group E Left eye previously enucleated; unilateral sporadic, Group E, left eye; tumor at cut end of nerve but no imaging evidence of tumor mass in the orbit Metastatic Rb to bones, bone marrow but no CNS involvement; unilateral sporadic, Group D right eye enucleated Bilateral sporadic retinoblastoma, right eye Group C, left eye Group E, enucleated; received adjuvant chemotherapy for tumor posterior to lamina cribrosa but not to cut end

International classification Staging Grouping Stage 0 Right eye Group D Left eye Group E Stage II Right eye Left eye Group E (enucleated) Stage IVa Stage I

Right eye Group D Left eye normal Right eye Group C Left eye Group E (enucleated)

 oving Toward a New TNM System M for Retinoblastoma: AJCC 8th Ed In 2017 the 8th edition of the AJCC was published which introduced a comprehensive reclassification of the retinoblastoma grouping system including the addition of the H “heritable trait” for genetic status of the patient [23]. This was based on retrospective evaluation of 1728 eyes diagnosed between 2001 and 2011 from multiple retinoblastoma centers; Kaplan-Meier analyses of the proportion of eyes salvaged without external beam radiation were evaluated by clinical features. The salvage of the eyes based on the TNM grouping was the most predictive of any other clinical systems. Comparison of the 8th Ed. TNM with the IIRC, ICRB, and IRSS clinical staging systems is shown in (Table 3.6). This is currently being actively adopted at many centers worldwide.

Summary This chapter discusses staging and grouping schema for retinoblastoma, with a focus on the International Intraocular Classification of Retinoblastoma and the more recent TNMH sys-

Comments Stage 0 conveys that neither eye has been enucleated. After enucleation, this patient’s disease will be Stage I if there is no microscopic residual tumor. High-risk pathology would not make this Stage II This patient has Stage II retinoblastoma because there is proven microscopic residual disease in the orbit (tumor extended beyond the surgical margin)

This patient has Stage I retinoblastoma. Following enucleation, the pathologic finding of high-risk pathology does not imply residual microscopic orbital tumor

tem. Common and uniform use of both staging and grouping in all patients will give pediatric oncologists and ophthalmologists who treat retinoblastoma a road map for initial therapy and a reliable means to discuss variation in outcomes across centers. Appropriate staging provides clinicians with an assessment of the likely prognosis for salvage of the child and his or her eye[s] before treatment begins. It will allow prediction of treatment morbidity and create a sound environment for successful clinical trials. Uniform staging and grouping allows medical professionals, government officials, and parents from any country to focus on minimizing the loss of life and loss of vision from retinoblastoma. Finally, and perhaps most importantly, the classification will need to evolve as the field does. Just as treatment paradigms created a clinical need to shift away from the Reese-Ellsworth classification, new standards will need to be updated frequently particularly with newer therapies such as intravitreal chemotherapy for seeding. The new TNMH system is an important step in this direction although it now needs prospective validation. Additionally, while a biopsy for retinoblastoma is currently contraindicated, other means of attaining biomarkers may be available in the future that impact diagnosis and prognosis and would be considered in any classification system [24].

37

3  Retinoblastoma: Staging and Grouping Table 3.6  Summary of clinical staging systems

cT1  cT1a

 cT1b cT2  cT2a

AJCC Clinical Staging 8th edition, 2017 [23] Intraretinal tumor(s) with subretinal fluid ≤5 mm from base of any tumor Tumors ≤3 mm and further than 1.5 mm from disc and fovea Tumors >3 mm or closer than 1.5 mm from disc or fovea Intraocular tumor(s) with retinal detachment, vitreous seeding, or subretinal seeding Subretinal fluid >5 mm from the base of any tumor

 cT2b

Vitreous seeding and/or subretinal seeding

cT3  cT3a  cT3b

Advanced intraocular tumor(s) Phthisis or pre-phthisis bulbi Tumor invasion of choroid, pars plana, ciliary body, lens, zonules, iris, or anterior chamber Raised intraocular pressure with neovascularization and/or buphthalmos Hyphema and/or massive vitreous hemorrhage Aseptic orbital cellulitis Extraocular tumor(s) involving orbit, including optic nerve Radiologic evidence of retrobulbar optic nerve involvement or thickening of optic nerve or involvement of orbital tissues Extraocular tumor clinically evident with proptosis and/or an orbital mass Evidence of preauricular, submandibular, and cervical lymph node involvement Clinical signs of distant metastasis Tumor(s) involving any distant site (e.g., bone marrow, liver) on clinical or radiologic tests Tumor involving the CNS on radiologic imaging (not including trilateral retinoblastoma) Hereditary trait Unknown or insufficient evidence of a constitutional RB1 gene mutation Normal RB1 alleles in blood tested with demonstrated high-sensitivity assays Bilateral retinoblastoma, retinoblastoma with an intracranial primitive neuroectodermal tumor (i.e., trilateral retinoblastoma), patient with family history of retinoblastoma, or molecular definition of a constitutional RB1 gene mutation

 cT3c  cT3d  cT3e cT4  cT4a

 cT4b N1 cM1  cM1a  cM1b

H  HX  H0  H1

IIRC Group Murphree, 2005 [5]

ICRB Group Shields, 2006 [19]

A, > 3 mm to A, > 3 mm to fovea or B, 1.5 to fovea or B, 1.5 to 3 mm 3 mm B, ≤3 mm orC, B 3 to 5 mm

C, >5 mm orD, > 1 quadrant C, “local” orD, “diffuse”

IRSS Stage Chantada, 2006 [12]

–

C, orE, tumor >50% of eye volume C, ≤3 mm orD, > 3 mm orE, tumor >50% of eye volume

E E

E E

I or II I or II

E

E

I or II

E E

E E

I or II I or II

I or II

IIIa IIIb

IVa IVb

Used with permission of the American College of Surgeons, Chicago, Illinois. The original and primary source for this information is the AJCC Cancer Staging Manual, Eighth Edition (2017) published by Springer International Publishing.

38

References 1. Fleming ID.  Staging of pediatric cancers: problems in the development of a national system. Semin Surg Oncol. 1992;8:94–7. 2. Berman JJ.  Tumor classification: molecular analysis meets Aristotle. BMC Cancer. 2004;4:10. 3. Reese AB, Ellsworth RM. The evaluation and current concept of retinoblastoma therapy. Trans Am Acad Ophthalmol Otolaryngol. 1963;67:164–72. 4. Ellsworth RM.  The practical management of retinoblastoma. Trans Am Ophthalmol Soc. 1969;67:462–534. 5. Murphree AL. Intraocular retinoblastoma: the case for a new group classification. Ophthalmol Clin N Am. 2005;18:41–53, viii 6. Pratt CB, Fontanesi J, Lu X, et al.  Proposal for a new staging scheme for intraocular and extraocular retinoblastoma based on an analysis of 103 globes. Oncologist. 1997;2:1–5. 7. Grabowski EF, Abramson DH. Intraocular and extraocular retinoblastoma. Hematol Oncol Clin North Am. 1987;1:721–35. 8. Stannard C, Sealy R, Hering E, et al. Postenucleation orbits in retinoblastoma: treatment with 125I brachytherapy. Int J Radiat Oncol Biol Phys. 2002;54:1446–54. 9. Retinoblastoma. In: Greene FL, Page DL, Fleming ID, et al., editors. AJCC cancer staging manual. 6th ed. New York: Springer; 2002. p. 371–6. 10. Wolff JBC, Ellsworth R. Extraocular retinoblastoma. In: Childrens cancer Study Group Protocol CCSG 962; 1978. 11. Kivela T, Kujala E.  Prognostication in eye cancer: the latest tumor, node, metastasis classification and beyond. Eye (Lond). 2013;27:243–52. 12. Chantada G, Doz F, Antoneli CB, et  al. A proposal for an international retinoblastoma staging system. Pediatr Blood Cancer. 2006;47:801–5. 13. Chantada GL, Sampor C, Bosaleh A, et al. Comparison of staging systems for extraocular retinoblastoma: analysis of 533 patients. JAMA Ophthalmol. 2013;131:1127–34. 14. Sastre X, Chantada GL, Doz F, et al., International Retinoblastoma Staging Working Group. Proceedings

J. L. Berry and A. L. Murphree of the consensus meetings from the International Retinoblastoma Staging Working Group on the pathology guidelines for the examination of enucleated eyes and evaluation of prognostic risk factors in retinoblastoma. Arch Pathol Lab Med. 2009;133:1199–202. 15. de Graaf P, Goricke S, Rodjan F, et al., European Retinoblastoma Imaging Collaboration (ERIC). Guidelines for imaging retinoblastoma: imaging principles and MRI standardization. Pediatr Radiol. 2012;42:2–14. 16. Novetsky DE, Abramson DH, Kim JW, et al. Published international classification of retinoblastoma (ICRB) definitions contain inconsistencies--an analysis of impact. Ophthalmic Genet. 2009;30:40–4. 17. Berry JL, Jubran R, Kim JW, et al. Long-­term outcomes of Group D eyes in bilateral retinoblastoma patients treated with chemoreduction and low-dose IMRT salvage. Pediatr Blood Cancer. 2013;60:688–93. 18. Shields CL, Mashayekhi A, Au AK, et al. The International Classification of Retinoblastoma predicts chemoreduction success. Ophthalmology. 2006;113:2276–80. 19. Shields CL, Shields JA. Basic understanding of current classification and management of retinoblastoma. Curr Opin Ophthalmol. 2006;17:228. 20. Berry JL, Jubran R, Wong K, et al. Factors predictive of long-term visual outcomes of Group D eyes treated with chemoreduction and low-dose IMRT salvage: the Children's Hospital Los Angeles experience. Br J Ophthalmol. 2014;98:1061–5. 21. Berry JL, Bechtold M, Shah S, et al.  Not all seeds are created equal: seed classification is predictive of outcomes in retinoblastoma. Ophthalmology. 2017;124:1817–25. 22. Francis JH, Brodie SE, Marr B, et al.  Efficacy and toxicity of intravitreous chemotherapy for retinoblastoma: four-­ year experience. Ophthalmology. 2017;124:488–95. 23. Mallipatna AC. AJCC cancer staging manual. 8th ed. New York: Springer; 2017. 24. Berry JL, Xu L, Murphree AL, et al. Potential of aqueous humor as a surrogate tumor biopsy for retinoblastoma. JAMA Ophthalmol. 2017;135:1221–30.

4

Retinoblastoma: Incidence and Etiologic Factors Manuela Orjuela-Grimm, Nakul Singh, Silvia Bhatt-Carreño, and Arun D. Singh

Introduction Retinoblastoma is the paradigm for the two-hit model of carcinogenesis [1]. From a genetic standpoint, three forms may be considered: familial, sporadic heritable, and nonheritable retinoblastoma. These three forms are thought to account for most instances of retinoblastoma. However, findings on imprinting and mosaicism indicate that our understanding of the genetics of this disease is still evolving and that the genetics are more complex than indicated by three main forms [2–4]. The key to understanding development of retinoblastoma is an understanding of the role of the retinoblastoma protein, pRb, the product of RB1, the retinoblastoma gene, in regulating the cell cycle and its transition from resting to synthesis. Retinoblastoma is one of few tumors for which the molecular origin of the disease is well understood. Epidemiology contributes in understanding what external factors may contrib-

M. Orjuela-Grimm (*) · S. Bhatt-Carreño Department of Epidemiology, Mailman School of Public Health, Columbia University Medical Center, New York, NY, USA e-mail: [email protected] N. Singh School of Medicine, Case Western University, Cleveland, OH, USA A. D. Singh Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_4

ute to the lack of functional pRb, whether because of truncated protein (from functional mutations in RB1 leading to premature stop codons) or total absence of pRb resulting from large chromosomal deletions or from inactivation of pRb through viral complexes that specifically bind pRb or by methylation of the RB1 promoter leading to silencing with lack of expression and gene product.

Familial Retinoblastoma Ten percent of children with retinoblastoma inherit a RB1 gene mutation from a parent. In this circumstance, the condition is referred to as familial retinoblastoma. Every cell in the body of these children contains a RB1 gene mutation, the “first hit” [5]. The mutation to the other copy of the RB1 gene, the “second hit,” occurs in a retinal cell sometime after conception. The inherited gene mutation is highly penetrant, and nearly all, about 95%, of such children develop retinoblastoma.

Sporadic Heritable Retinoblastoma Another 30% of children with retinoblastoma also harbor a RB1 mutation in all of their cells and are at the same risk for developing ­retinoblastoma as children who inherit a muta39

40

M. Orjuela-Grimm et al.

tion. However, these children do not have a parent with the mutation. Rather, their RB1 mutation occurred as a new germ line mutation. Although these children did not inherit the gene from an affected parent, they will be able to pass the mutation onto their children. This is referred to as “sporadic heritable” retinoblastoma.

combined) and unilateral rates as reflecting nonheritable disease. In the only report of international variation in incidence by laterality, the incidence of unilateral disease was observed to vary markedly, much more so than bilateral disease [8].

Nonheritable Retinoblastoma The remaining 60 percent of retinoblastoma patients have a nonheritable disease. Their retinoblastoma develops as the result of mutations or changes in both alleles of RB1 that occur in a single retinal cell sometime after conception and result in a lack of functional pRb.

Because 95% of cases are diagnosed under the age of 5 years, incidence rates are better expressed as “per million children 0–4 years of age,” than as “per million children 0–14 years of age,” as is common for other childhood cancers. In the graphs and discussion that follow, we present rates for children ages 0–4  years of age, whenever the data are available.

Variations in Incidence

Geographic Variation in Incidence

Unilateral and Bilateral Retinoblastoma

Variation in incidence among countries, regions, and ethnic groups or over time can provide clues to etiology. Environmental (defined here as nongenetic) factors are implicated in cancers that show great variation in incidence. The rates of retinoblastoma vary about 50-fold across the continents [9], a degree of variability higher than that for several adult cancers, namely, stomach, colon, cervical, and pancreatic cancer, and lower than that for lung and esophageal cancer, among others [9]. Overall, the rates appear somewhat higher in less industrialized countries than in more industrialized countries. In addition, there are large variations in incidence within some countries. The data suggest variation by economic development with higher rates in poorer regions of countries such as Brazil and Mexico [10]. Clearly, the differences in the incidence rates of retinoblastoma between regions of higher and lower incidence may be due to other factors such as ethnic origin, genetic susceptibilities, and cultural and behavioral practices. Closer examination of the differences in incidence may s­ uggest risk factors for the development of retinoblastoma. Below we examine patterns of incidence using data from the latest compilation organized by the

The vast majority of children with familial or sporadic heritable retinoblastoma develop bilateral disease though about 10–15% present with disease affecting only one eye [6]. Rarely patients with the heritable form develop an intracranial midline primitive neuroectodermal tumor, in addition to the disease in both retinae, leading to what is referred to as trilateral disease. Trilateral disease is estimated to affect 3–5% of those patients with bilateral disease, thus affecting less than 1% of patients with retinoblastoma [7]. All nonheritable retinoblastomas are unilateral. Although incidence rates would be most informative if they were available for the three subtypes of retinoblastoma, incidence rates are generally presented only for retinoblastoma in toto. Rates by laterality are available only for few countries. As explained above, bilateral retinoblastoma includes most instances of familial and sporadic heritable disease, while the vast majority of unilateral disease is nonheritable retinoblastoma. Therefore, bilateral rates can be interpreted as reflecting the incidence of heritable retinoblastoma (familial and sporadic heritable

Expression of Incidence

41

4  Retinoblastoma: Incidence and Etiologic Factors

International Agency for Research on Cancer (IARC) [11]. A total of 153 population-based registries from 62 countries contributed data to the recently published third edition of the global/ international incidence of childhood cancer. These data include new registries not included in the prior versions. Some variations in incidence of retinoblastoma noted in the prior versions now appear somewhat different [12]. Because the overwhelming majority of cases of retinoblastoma are diagnosed in children younger than 5 years, we examined variation in incidence by regions or continents in Figs.  4.1, 4.2, 4.3, 4.4, and 4.5 citing the IARC calculated rates in this age range. However, the N for cases was not listed separately for this age group for any cancer; thus the N we list is the total N for those younger than 15 years [12]. The N thus provides context to the number of cases in the registry and the stability of the calculated rates. Additionally, some countries had multiple registries with national coverage. Some countries had a general registry and a pediatric cancer registry. The pediatric registries generally had more years of informative data. In such instances, the registry with the largest number of cases was chosen

as more representative of the number of cases for each country. We examined incidence by geographic regions. Countries were grouped into geographic regions following United Nations’ definitions, consistent with other reports based on registry data collected by IARC [13, 14]. All the rates listed in the following sections are agestandardized with rates per million for ages 0–4, and the N listed is for ages 0–14 in parenthesis in corresponding tables.

North America The incidence of retinoblastoma in the United States (USA) did not change significantly from 1975 to 1995 [15, 16]. Incidence for the USA and Canada for 2000–2010 is shown in Fig. 4.1 using SEER registry data for the USA. Incidence by race/ ethnicity and region in the SEER registries generally range from 10.8 to 14.6 per million children aged 0–4 per year, with higher rates among “nonHispanic Black” and “Hispanic White” subpopulations. The overall rate for children 0–4 in the USA is 12.9 based on SEER data (Fig. 4.1, Table 4.1). Additionally, there appears to be some regional

North America

Age Standardized Rates per Million 0-4 Years

16 13.8

14 12

13.7

12.9 12

12.2

11.8 10.8

10 8 6 4 2 0 CANADA

USA

USA, Asian Pacific USA, Black Islander

Fig. 4.1 Incidence of retinoblastoma in the North America in children ages 0–4  years. Note that the rates shown are for ages 0 to 4  years. However the N that is

USA, Hispanic White

USA, Native American

USA, NH White

available and published by IARC is the N of cases ages 0 to 14 years. (Based on data from Steliarova-Foucher et al. [67])

M. Orjuela-Grimm et al.

42

variation in the USA with several subpopulations appearing to have a higher incidence in children under 5 years that would be categorized as among the higher incidence rates globally. However, the variation has to be observed with caution as incidence is based on a small number of cases in several registries. These subpopulations reflected registry data collected from 1998 to 2012: the District of Columbia with an N of 12 and a rate of 24.20 per million children; Delaware, in the non-Hispanic White population, with an N of 11 and a rate of Table 4.1  Registries included in Fig.  4.1 with time period and total number of retinoblastoma cases: North America

Age Standardized Rates per Million 0-4 Years

Registry by country Canada, 9 registries USA, SEER 18 USA, API, SEER 18 USA, Black, SEER 18 USA, Hispanic White, SEER 18 USA, Native American, SEER 18 USA, NH White, SEER18

Year N 1992–2013 463 1993–2012 1270 1993–2012 115 1993–2012 189 1993–2012 375 1993–2012 21 1993–2012 545

Within Europe, there is also some variability in incidence [13]. Interestingly the pediatric registries tend to have a higher number of contributing cases and calculate incidence rates that appear slightly higher than those rates derived from general regional cancer registries (Fig. 4.2, Table 4.2). The lowest incidence reported in Europe is in Austria with an incidence of 4.6 (N of 47), while Belgium has the highest incidence for all of

16.2

16 12.6

16 12.2 11.9

12

12.1 11.4

10.2

10

11.3 11.4 10.5

12.9 10.9

8.6 7.2 5.9

6.8

11.2

11

9.8

8.4

8 6

Europe

Europe

18

14

23.50; and Florida in the Hispanic White population, with an N of 76 and an incidence rate of 21.10. Interestingly, recent analyses by the Centers for Disease Control and Prevention examined childhood cancer by state (2003–2014) and noted that incidence of retinoblastoma is highest in the northeastern states, though the rates were only slightly higher than in the south or west, while rates in the Midwest seemed lower than in other areas [17].

8.6

12.1

10.2

9.3 7.8

4.6

4 2 0

E IA M Y S IA IC S D Y D N IA E LY TA IA AL IA IN IA IA Y N D D K NC STR GIU AN AND LAN ON AN WA DE LAN LAN U ARU AR BL AR LAN TIO VAK AIN ITA AL OAT UG EN PA A E E E T U R M M R RT OV S L R L LG PU NG PO RA O KR FR AU BE ER ERL ZE ES ITH NO SW IC IR C O SL E SL U BE BU RE HU L D G H IT P H T W FE C E S E N N IA E CZ SS TH RU

Western Europe

Northern Europe

Fig. 4.2  Incidence of retinoblastoma in Europe in children ages 0–4 years. Note that the rates shown are for ages 0 to 4 years. However the N that is available and published

Eastern Europe

Southern Europe

by IARC is the N of cases ages 0 to 14 years. (Based on data from Steliarova-Foucher et al. [67])

4  Retinoblastoma: Incidence and Etiologic Factors Table 4.2  Registries included in Fig.  4.2 with time period and total number of retinoblastoma cases: Europe Registry by country France, pediatric Austria Belgium Germany, pediatric The Netherlands Switzerland, pediatric Estonia Lithuania Norway Sweden Iceland Ireland UK Belarus, pediatric Bulgaria Czech republic Hungary Poland Russian federation, 2 registries Slovakia Ukraine Italy, 26 registries Malta Croatia Portugal Slovenia Spain, 2 pediatric registries

Year 2000–2012 1990–2012 2004–2013 1996–2012 1993–2013 1990–2013 1990–2012 2000–2012 1990–2013 1990–2011 1990–2014 1994–2012 2000–2011 1990–2014 1990–2013 1990–2012 1991–2014 2001–2013 1998–2015 1990–2009 2002–2012 1992–2013 1994–2013 2001–2014 1991–2012 1991–2012 1991–2013

N 644 47 104 671 261 124 15 13 85 148 4 64 510 150 68 139 129 239 42 58 239 156 8 25 116 22 215

Europe at 16.2 (with an N of 104), followed by Malta at 16.0 (N of 8), Portugal at 12.9 (N of 116), France 12.6 (N of 644), and Switzerland incidence of 12.2 (N of 124). The East German registry shows an incidence of 8.4 (N of 40) based on the six registries in former East Germany for the years 2001–2007. Although the rates are similar to those from the five Western registries in Germany, (9.20), the Western registry reports data from a longer period (1994–2012) with a greater number of cases (N of 147). In comparison, the German pediatric registry (with national data) has an N of 671, reports data from 1996 to 2012, and has the highest rate for Germany at 10.20 per million children for ages 0–4. In the UK, there are three large registries with similar rates: the UK, England, has an incidence of 11.1 (N of 870) in the years 1990 to 2013, the UK-wide registry has an incidence of 11.3 (N of 510) in the years 2000 to 2011(listed in the

43

graphs), and the UK, England, and Wales pediatric registry has an incidence of 11.6 (N of 793) in the years 1991 to 2010. Spain has two big registries, the Spain 11 registries with an incidence of 10.1 (N of 107) in the years 1990 to 2013 and the Spain pediatric registry with an incidence 12.1 (N of 215) in the years 1991 to 2013 (Fig.  4.2, Table  4.2) [9]. Although many of the registries have small numbers of cases, the differences within Europe are intriguing and do not appear to follow an easily discernible pattern.

Central and South America Population-based registries do not exist for all countries in Central and South America, and for some countries, rates are only available within select cities (Fig. 4.3, Table 4.3). However, even with these limitations, there appear to be two groups in Central and South America, those regions with incidence under 12.6 per million per year in children ages 0–4 and those with an incidence greater than 14 per million per year [9]. The data for Brazil is only based on what was submitted to this volume of the IARC compilation. Prior editions reported higher incidence but included additional Brazilian registries, such as Belem. Brazil has a total of 20 registries which collect pediatric cancer cases, some of which have data beginning in 1967 [18]. We note this as it is possible that Brazil may have a different incidence, possibly higher, as published in other reports, but our graphs only reflect the data reported in the IARC volume 3.

Asia Incidence also varies greatly in Asia (Figs.  4.4 and 4.5, Tables 4.4 and 4.5) [9]. The highest rate is found in Western Asia in Jordan at 16.9 (N of 173). The lowest incidence in Asia is also found in Western Asia in Qatar at 0.09 (N of 1). The next lowest is Cyprus at 1.5 (N of 1) followed by Kuwait at 1.6 (N of 14). Israel which is classified as Western Asia by the UN [14] interestingly does not have a difference between Jews and

11.8

12.6

9.2

11.9

15.8

17.7

14

South America

ARGENTINA URUGUAY BRAZIL CHILE COLOMBIA ECUADOR PERU

0

2

4

6

8

10

12

14

16

18

20

7.7

10.5 9.8

Central America

MEXICO HONDURAS COSTA GF, GP, RICA MTQ

11.7

CUBA

9.3

Caribbean

JAMAICA

14.7

USA, Puerto Rico

7.2

Fig. 4.3  Incidence of retinoblastoma in Central and South America and the Caribbean in children aged 0–4 years. Note that the rates shown are for ages 0 to 4 years. However the N that is available and published by IARC is the N of cases between ages 0 and 14 years. (Based on data from Steliarova-Foucher et al. [67])

Age Satndardized Rates per Million 0-4 Years

South,Central America and Caribbean

44 M. Orjuela-Grimm et al.

4  Retinoblastoma: Incidence and Etiologic Factors

45

non-Jews in their registries although the N for the non-Jews is much lower. From 1990 to 2012 in Israel, the registries show an incidence of 9.5 (N of 109) for Jews, 9.6 (N of 42) for nonJews, and 9.5 (N of 151) total for the country. Table 4.3  Registries included in Fig.  4.3 with time period and total number of retinoblastoma cases: Latin America and the Caribbean Registry by country Argentina, pediatric Uruguay Brazil, 5 registries Chile, pediatric Colombia, 4 registries Ecuador, 5 registries Peru, Lima Mexico, Mexico City, pediatric Honduras, Francisco Morazán Costa Rica French Guiana, Guadeloupe, Martinique, 3 registries Jamaica, Kingston and St Andrew Cuba USA, Puerto Rico

Year N 2000–2013 624 1993–2012 69 1995–2012 63 2007–2011 79 1992–2013 94 1993–2013 135 2010–2012 37 1997–2013 46 2002–2012 26 1993–2012 87 1990–2012 10 1982–2012 2000–2012 1992–2012

31 90 44

Turkey had registries in two categories, one for the whole country made of eight registries and one pediatric. The pediatric registry is used in the charts because it has a higher N (171) and an incidence of 8.6 in years 2009 to 2011. However, it is important to note that the other compilation of eight registries for Turkey has an N of 137 and a higher incidence of 10.0 and a longer collection period (1992 to 2012).

Table 4.4  Registries included in Fig.  4.4 with time period and total number of retinoblastoma cases: Southern and Eastern Asia in children Registry by country India, 7 registries Iran, Golestan Pakistan, Lahore Philippines, 2 registries Thailand, 6 registries Vietnam, Ho Chi Minh City China, 6 registries Republic of Korea Japan, 8 registries

Year 1990–2013 2004–2011 2008–2012 1993–2012 1993–2013 1995–2013 1990–2013 1999–2012 1990–2013

N 599 2 81 495 133 125 248 448 252

Asia (South,East) 18 Age Standardized Rates per Million 0-4 Years

16.5 16 14.1

14 11.7

12 10

14.4 11.6

12.6

9.9

9.8

8 6 4 1.8

2 0

INDIA

IRAN

PAKISTAN PHILIPPINES THAILAND VIETNAM+

Southern Asia

South-Eastern Asia

Fig. 4.4  Incidence of retinoblastoma in Southern and Eastern Asia in children ages 0–4  years. Note that the rates shown are for ages 0 to 4 years. However the N that is available and published by IARC is the N of cases between ages 0 and 14  years. + Vietnam has the third

CHINA REPUBLIC OF JAPAN KOREA

Eastern Asia

highest rate, but the data has one of the lowest proportions of microscopically verified cases in Asia, and the proportion of unspecified cases is higher than 20% which is the high cutoff for unspecified neoplasms. (Based on data from Steliarova-Foucher et al. [67])

M. Orjuela-Grimm et al.

46 Western Asia,Northern Africa 25

Age Standardized Rates per Million 0-4 Years

21.8 20 16.9

16.6 15

11.8 9.5

10

8.6

8.2

7.6 5.5 5

3.8

3.7 1.6

1.5

0.9

0 BAHRAIN ISRAEL

KUWAIT LEBANON

SAUDI ARABIA

JORDAN

QATAR

Western Asia

CYPRUS

TURKEY ALGERIA

EGYPT

LIBYA MOROCCO TUNISIA

Northern Africa

Fig. 4.5  Incidence of retinoblastoma in Western Asia and Northern Africa in children ages 0–4 years. Note that the rates shown are for ages 0 to 4 years. However the N that

is available and published by IARC is the N of cases between ages 0 and 14  years. (Based on data from Steliarova-Foucher et al. [67])

Table 4.5  Registries included in Fig.  4.6 with time period and total number of retinoblastoma cases Western Asia and Northern Africa in children

dence of 15.9 (N of 165) [19]. In East Asia, China has the highest rate of 14.4 (N of 248). Vietnam has a high rate at 14.1 (N of 125), but the data has one of the lowest proportions of microscopically verified cases in Asia, and the proportion of unspecified cases is higher than 20% which is the higher cutoff for unspecified neoplasms. Although microscopic verification is not needed for retinoblastoma, in Vietnam given the distribution of types of cases during the periods reflected by the registry (1995–2013), one would expect that most unilateral cases would be diagnosed and confirmed microscopically. Although one recent report from Usmanov et al. suggests that incidence in Asia and Oceania (the Asia-Pacific region) may be expected to decrease in the next 5 years [24], incidences in that report were not calculated for children in a particular age group but rather for the population as a whole, thus potentially reflecting changes in population demographics rather than actual changes in incidence of disease among children potentially at risk for developing retinoblastoma.

Registry by country Bahrain Israel Kuwait Lebanon Saudi Arabia, Riyadh Jordan Qatar Cyprus, Southwest Turkey, pediatric Algeria, 5 registries Egypt, Gharbiah Libya, Benghazi Morocco, 2 registries Tunisia, 2 registries

Year 1998–2012 1990–2012 1994–2012 2008–2010 1994–2012 2000–2012 2002–2014 1998–2012 2009–2011 1996–2014 1999–2010 2003–2008 2005–2012 1993–2007

N 6 151 14 11 211 173 1 1 171 27 27 14 72 55

In Saudi Arabia, years 1994 to 2012, the total incidence at ages 0–4 is 16.6 (N of 211), but when broken down into Saudis and non-Saudis, there is a large difference in the incidence, again noting the big difference in number of cases for each group. Non-Saudis have an incidence of 20.1 (N of 46), and Saudis have a smaller inci-

4  Retinoblastoma: Incidence and Etiologic Factors

Africa In Africa, where there are few population-based registries, incidence is also quite variable. Africa is divided into two regions by the UN, Northern Africa, and sub-Saharan Africa further divided into Eastern Africa, Middle Africa, Southern Africa, and Western Africa [14]. The IARC data contains incidence rates for some but not all regions; in the figures brackets have been added to show the different breakdown of the regions and data. In South Africa, the pediatric registry has an overall incidence of 7.7 (N of 643) during the years 1998 to 2012. However, the incidence and Ns appear to vary by ethnic subgroups, though the majority of cases are in South African Black children and other ethnic subgroups have small numbers of cases. Incidence appears lowest among South African Asians with an incidence of 5.6 (though the N is only 7), while in the South African Blacks, incidence is 7.6 (N of 549) and 8.3 in Colored children (N of 54) and 8.5  in South African White children (N of 33) (Figs.  4.5 and 4.6,

Tables 4.5 and 4.6) [9]. Incidence in sub-­ Saharan Africa is much higher than in Northern Africa which was previously classified as the Middle East but is now included in Africa. However even within the higher rates of sub-­ Saharan Africa, there is variability. The highest rates are in Eastern Africa and generally lower rates in the central and southern regions of the continent. South Africa has a rate of 7.7 (N of 643) for years 1998 to 2012. But it is important to note the breakdown by race where Black has the highest incidence at 7.6 and the majority of Table 4.6  Registries included in Fig.  4.6 with time period and total number of retinoblastoma cases: sub-­ Saharan Africa in children Registry by country Kenya, 2 registries Mauritius Réunion Uganda, Kampala Zimbabwe. Harare Cameroon, Yaoundé Botswana South Africa, pediatric

Year 2000–2012 2001–2013 1990–2011 1996–2013 1995–2013 2004–2006 1999–2007 1998–2012

N 78 10 13 82 87 14 22 643

Africa(Sub-Saharan)

25 Age Standardized Rates per Million 0-4 Years

47

21.4 20 16 14.2

15

11.9 10

8.6

8.9

10.8 7.7

5

0

KENYA

MAURITIUS RÉUNION

UGANDA ZIMBABWE CAMEROON BOTSWANA SOUTH AFRICA

Eastern Africa

Fig. 4.6 Incidence of retinoblastoma in sub-Saharan Africa in children ages 0–4  years. Note that the rates shown are for ages 0 to 4  years. However the N that is

Middle Africa

Southern Africa

available and published by IARC is the N of cases between ages 0 and 14  years. (Based on data from Steliarova-­ Foucher et al. [67])

M. Orjuela-Grimm et al.

48

cases at N of 549 for the same number of years. Mali has an incidence of 58.3 (N of 208) but is not included in the figure due to possible data inconsistencies as that registry is noted to have one of the lowest proportions of microscopically verified cases in Africa, and the proportion of unspecified cases is higher than IARC’s 20% cutoff for proportion of unspecified neoplasms. Similarly to Vietnam noted above, one would expect that during the years reported by the registry (2005–2014), the majority of unilateral cases would have microscopic evaluations. Additionally, the incidence in Mali appears so extreme that including it in the figure would distort the ability to examine variations in incidence between other populations.

Oceania

Gender Variation in Incidence Overall there appears to be no difference in incidence rates by gender for retinoblastoma [12], with similar incidence rates for males and females in most countries of the world [9]. However, 21 countries appear to have varying incidence by gender. Males appear to have a higher incidence than females in India, Australia, China, Jordan, Turkey, Ecuador, Portugal, Colombia, Zimbabwe, Kenya, Morocco, Uruguay, Italy, Croatia, Slovenia, USA Native Americans SEER 18, and Chile (Fig. 4.8, Table  4.8). The inverse appears true in Sweden, Switzerland, Costa Rica, Brazil, and Algeria in which females appear to have higher rates. Although most prior reports did not indicate variation in incidence by gender [16], a report based upon SEER 18 data indicated that incidence during 2000–2009 among boys was significantly higher than among girls (IRR, 1.18) [20].

Fig. 4.7  Incidence of retinoblastoma in Oceania in children ages 0–4 years. Note that the rates shown are for ages 0 to 4 years. However the N that is available and published by IARC is the N of cases between ages 0 and 14 years. (Based on data from Steliarova-Foucher et al. [67])

Age Standardized Rates per Million 0-4 Years

The incidence in Australia is similar to that of the Table 4.7  Registries included in Fig.  4.7 with time United States and Canada at 11.5 (N of 350) period and total number of retinoblastoma cases in Oceania (Fig.  4.7, Table  4.7) [9]. The incidence in New Registry by country Year N Zealand which is 15.5 (N of 96), although higher Australia 1992–2014 350 than that in Australia, appears somewhat lower in French Polynesia, New 1990–2013 10 Maori at 12.8 (N of 13) than in non-Maori at 14.2 Caledonia, 2 registries (N of 44) populations. New Zealand 1993–2012 96

Oceania 18 15.5

16 14 12

11.5

10

9.1

8 6 4 2 0

AUSTRALIA

FRENCH POLYNESIA, NEW CALEDONIA

NEW ZEALAND

4  Retinoblastoma: Incidence and Etiologic Factors

49

Incidence by Gender 14 11.7

12

9.7

10

6

8.2

8

8

7.3 6.2 6.2

4.9 5.2 5.2 4 3.6 3.9

6.2 4

7.6

5.7 5.7

8.3

8.7

6.2

5.8

4.6

4.4

4.4

6.3 6.2

5.6 3.6

3

4.9 3

5.6 4.5

2.7

3.6

5

4.8 3.7

3.3 2.6

1.8

2

1

0 A A L Y A IA SA IA WE YA EN ND ICA AZIL ERI DI ALIA INA DAN EY DOR GA CO UAYITAL ATI EN U ED LA R IN MB AB KEN OC U R R O G G V CH OR URK UA R T A O L T B U R L S A J CR SLO SW ZE T EC OR ST MB UR AU MO CO ZI P IT CO SW

Female

Male

Fig. 4.8  Comparison of age-standardized rates (ASR) by gender for boys and girls in ages 0–14 years in countries with rates differing by gender. Note that the rates shown

are for ages 0 to 14 years. The N that is available and published by IARC is the N of cases between 0 and 14 years. (Based on data from Steliarova-Foucher et al. [67])

Table 4.8  Comparison of age-standardized rates (ASR) by gender for boys and girls in ages 0–14 years in countries with rates differing by gender

Comment on Data Interpretation

Registry by country India, 7 registries Australia China, 6 registries Jordan Turkey, pediatric Ecuador, 5 registries Portugal Colombia, 4 registries Zimbabwe, Harare Kenya, 2 registries Morocco, 2 registries Uruguay Italy, 2 pediatric registries Croatia Slovenia USA, SEER 18 Native American Sweden Switzerland, pediatric Costa Rica Brazil, 5 registries Algeria, 5 registries

Year 1990–2013 1992–2014 1990–2013 2000–2012 2009–2011 1993–2013 1991–2012 1992–2013 1995–2013 2000–2012 2005–2012 1993–2012 1998–2011 2001–2014 1991–2012 1993–2012 1990–2011 1990–2013 1993–2012 1995–2012 1996–2014

N 599 350 248 173 171 135 116 94 87 78 72 69 34 25 22 21 148 124 87 63 27

Conflicting Results The incidence of retinoblastoma has been challenging to estimate globally. Few registries have data on laterality (unilateral vs. bilateral). There is conflicting data regarding relationship between the socioeconomic status and the incidence of retinoblastoma. The SEER data and some global data suggest increased incidence in areas with lower economic indices [8]. However, one recent report examining global variation in incidence of neuroblastoma, Wilms tumor, and retinoblastoma (from previously published IARC compiled registry data [9]) by human development indicators (composite score incorporating life expectancy, education, and standard of living) reported that the incidence of retinoblastoma did not vary by development indices, though data on laterality was not available [21]. Some global differences are particularly intriguing or paradoxical, given

50

expected similarities in ethnicity and presumed shared environmental exposures, for example, that Australia and New Zealand have different rates which are not easily explainable (especially given that rates by ethnic group in New Zealand show higher rates in the non-Maori) (Fig.  4.7, Table 4.7). Overall global variation suggests that environmental factors may play a role, though genetic susceptibility to particular environmental and behavioral risk factors may explain some of the differences.

Quality of Data and Subpopulations The highest incidence of retinoblastoma is noted in Mali (Bamako). However, it is useful to note the quality of the data from standards set by the IARC, the proportion of microscopically verified (MV%) cases and proportion of unspecified cases (NOS%). For MV% Mali has a 70% which is lower than 85%, the lower cut point for good validity of diagnosis. Mali also has an NOS% of 33.6% which is far higher than the 20% upper limit for the percent with unspecified neoplasms. It is also important to note the differences in years the registry has existed, as well as number of cases for each registry when comparing incidence rates for each country. Table 4.9 shows the incidence rates for ages 0–14 and 0–4 as a comparison of age incidence, for the top 14 countries with available data. Listed here are the top registries that have the highest incidence (incidence age 0–14 (N ages 0–14) and incidence ages 0–4). In some registries or countries, there were subpopulations with an incidence that was markedly higher than that of the overall registry or nation. For all of those instances, the number of cases is much smaller in the subpopulation; thus rates are generally less stable. Of note, three city-based registries for countries listed had higher rates than those of their corresponding national registries: for the Quito, Ecuador, registry, the incidence is 9.40 (for ages 0–14) (N  =  67) and 23.80 (for ages 0–4); for the Casablanca, Morocco, registry, the incidence is 9.40 (for ages 0–14) (N  =  65) and 22.00 (for

M. Orjuela-Grimm et al.

ages 0–4), and for Eldoret, Kenya, the incidence was 8.00 (for ages 0–14; N  =  32) and 17.10 (for ages 0–4). Similarly, among countries which had the top incidence rates, there were two ethnic subpopulations whose incidence rate was higher than the national or overall registry rate: for New Zealand, the Asian-Pacific Islander registered an incidence rate of 9.00 for ages 0–14 (N = 13) and 23.40 for ages 0–4, and for Saudi Arabia, Riyahd, the non-Saudi population had an incidence of 8.30 for ages 0–14 (N = 46) and 20.10 for ages 0–4. Additionally, there are regional registries that had rates in the range to be included in the top 14 registries, but for whom the national or overall registry data was not elevated enough to be in the range of the countries with the highest incidence. For the USA, the three US registries discussed in the North America section above had higher incidence than data for the nation as a whole, and, similarly, for Turkey, the rate in Ankara was high with 8.20 (N = 37) and 20.70, while that for Turkey overall was not, and for Italy, incidence in Umbria was much higher than in the rest of Italy with 8.10 (N  =  13) and 20.80 (Table  4.9) [22]. Although the elevated incidence in subpopulations needs to be interpreted with caution, some of the regional and ethnic variability may provide etiologic clues for future study. Some countries whose incidence previously placed them among the highest global rates in the prior IARC Childhood Cancer Incidence volume [9] were not able to provide data in time for the IARC 3rd volume deadlines and were excluded. Among these, a notable absence is Denmark.

 odels Based on Scandinavian Birth M Cohorts and Birth Rate An alternate approach to examining incidence has advocated using modeling for regions that have sparse registry data. Given high birth rate and large populations in Asia and Africa, it is not surprising that the vast majority of retinoblastoma occur in these two continents (57% and 25%, respectively) [23]. The alternate method

51

4  Retinoblastoma: Incidence and Etiologic Factors Table 4.9  Registries with higher incidence retinoblastoma, rates 0–14 and 0–4 Country (registry/ ethnicity) Zimbabwe, Harare (1995–2013) Morocco, 2 registries (2005–2012) Ecuador, 5 registries (1993–2013) Jordon (2000–2012) Kenya, 2 registries (2000–2012) Philippines, 2 registries (1993–2012) Malta (1994–2013) Saudi Arabia, Riyadh (1994–2012) Belgium (2004–2013) Colombia, 4 registries (1992–2013) Uganda, Kampala (1996–2013) New Zealand (1993–2012) Jamaica, Kingston and St Andrew (1982–2012) Peru, Lima (2010–2012)

Incidence and (N) ages 0–14 9.60 (87)

Incidence ages 0–4 21.40

9.20 (72)

21.80

7.20 (135)

17.70

7.10 (173) 7.00 (78)

16.90 16.00

6.90 (495)

16.50

6.90 (8) 6.80 (211)

16.00 16.60

6.60 (104) 6.50 (94)

16.20 15.80

6.40 (82)

14.20

6.30 (96)

15.50

6.20 (31)

14.70

6.00 (37)

14.00

Countries with high incidence rates whose registry data did not meet the threshold for the IARC quality indicators are not included in this table. Additional explanation is found in Sect. 4.5.2 of the text

proposed by Kivela et al. uses the rate of retinoblastoma from two Scandinavian birth cohorts to project the number of cases anticipated in Eastern Asia. The birth cohort based modeling uses data from 465 children in Sweden and Finland who were diagnosed with retinoblastoma between 1958 and 1998 where incidence appeared stable during this period. In their calculation, based on number of cases per live births, the incidence of retinoblastoma was fixed to be uniform across populations, at one in 16,642 live births [16, 23, 24]. In contrast, the cancer registry data presented above and compiled by IARC report population-­ based incidence calculated from cases occurring in children in the ages in which there is risk for developing retinoblastoma. In these registry data, there appears to be variation in incidence.

 tiology: Risk Factors for Sporadic E Heritable Retinoblastoma Sporadic heritable retinoblastoma results from a new germ line mutation that is of paternal origin in over 90% of patients [25, 26]. By virtue of being a new germ line mutation, the mutation occurs before the child’s conception. Based on these two facts, it seems logical that the search for genetic and nongenetic risk factors for sporadic heritable retinoblastoma should focus on parental exposures before the child’s conception [27]. However, our understanding of retinoblastoma genetics and etiology is still evolving, and we should not dismiss the possible effects of maternal exposure. It would be reasonable to hypothesize preconception exposure to mutagens, variants of metabolizing genes that prolong the duration or increase the level of a mutagen in the body, and variants of DNA repair genes that result in less efficient repair of DNA damage as possible risk factors for the second “hit.” Only a few epidemiologic studies have investigated possible risk factors for new germ line mutation. Such studies have been limited in scope, mostly focusing on paternal age. The cohort studies of children of cancer survivors and of atomic bomb survivors have limited power to detect anything but large effects, but newer population level birth cohorts may yet provide relevant clues.

Parental Risk Factors A number of studies have examined paternal age in relation to sporadic heritable retinoblastoma with a wide range of results [28–31]. In the largest, most methodologically sound studies, the observed paternal age difference between those with retinoblastoma and the general population was about 1 year. This is much smaller than the difference of 4–10  years observed in achondroplasia [32, 33] and 2–5 years observed for Alpert syndrome [34, 35], genetic conditions for which a paternal age effect is well established. Increased risk with greater paternal age has been explained by the fact that the stem cells that give rise to

52

sperm are continuously dividing. Thus, the stem cells of an older man are more likely than those of a younger man to have sustained a mutation arising from an error during DNA replication [36]. The number of cell divisions between stem cell and mature sperm is estimated to be 197 at age 20, 427 at age 30, and 772 at age 45 [37]. While the explanation about the increasing number of cell divisions at older ages might be expected to apply to all conditions due to a new germ line mutation, a paternal age effect is observed, for reasons unknown, only in some of these conditions. Overall, the evidence for a paternal age effect on sporadic heritable retinoblastoma is consistent with other findings in other diseases with genetic origins. A recent study based in the Children’s Oncology Group observed an association with paternal diagnostic x-ray exposure prior to the child’s conception and observed an association with maternal exposure as well; both showed increasing risk with increasing dose [38]. The association with paternal x-ray exposure replicates a statistically nonsignificant finding from an earlier, small study [39]. In another study, aspects of maternal and paternal diet and supplement use before the child’s conception were also associated with risk [40]. High cured meat intake of fathers appeared to increase risk, while high intake of dairy products and associated nutrients appeared protective, as did calcium supplements. Maternal use of multivitamins close to the child’s conception also appeared protective. Several findings about father’s occupational exposures have been reported; employment in the metal manufacturing industry [41], exposure to welding fumes [41], exposure to pesticides [42], and exposure to polycyclic aromatic hydrocarbons (PAH) and paints were associated with risk [43]. Exposure to PAH has been further examined in studies examining air pollution in California linking exposure to birth records [44]. Prenatal exposure to traffic-related air pollutants such as benzene and other PAH, as well as some metals (nickel and chromium), has been found to be associated with increased odds of both bilateral and unilateral retinoblastoma [45].

M. Orjuela-Grimm et al.

Because retinoblastoma is a rare disease, it is not possible to study its incidence prospectively. Several of the associations have been observed only once, with some reported from the same datasets. Furthermore, estimates of these associations are with wide confidence intervals and often of marginal statistical significance. Despite this, some clear patterns emerge including an increased risk with increased paternal age and apparent exposure to air pollutants that are known clastogens.

 tiological Factors for Nonheritable E Retinoblastoma Nonheritable retinoblastoma occurs as a result of somatic mutation. The child does not have a germ line RB1 mutation; rather, both copies of the RB1 gene are inactivated in a single developing retinal cell. As the mutations are somatic, they must occur after the child’s conception, either during gestation or early postnatal life. Therefore, examination of potential risk factors should focus on exposures of the mother that would affect the child in utero and the child after birth. The data on such possible risk factors for nonheritable retinoblastoma are very limited. Most of the findings have not yet been replicated and cannot be considered conclusive. Rather, the studies provide preliminary clues to be pursued in future studies.

Environmental Exposure Maternal use of insect or garden sprays during pregnancy, diagnostic x-rays with direct fetal exposure, and father’s employment as a welder, machinist, or related metal worker have been associated with increased risk of nonheritable retinoblastoma [39]. Maternal smoking before and during pregnancy is also weakly associated with increased unilateral retinoblastoma risk [46]. Prenatal and perinatal exposure to PAH and traffic exposure-related metals have been associated with increased risk of sporadic RB [47].

4  Retinoblastoma: Incidence and Etiologic Factors

 aternal Diet and/or Vitamin Intake M During Pregnancy The limited evidence suggests a role for diet and/ or use of multivitamin supplements during pregnancy. In a case-control study in central Mexico, lower intake of vegetables and fruits during pregnancy was associated with an increased risk of retinoblastoma in the child [48]. Another study found that multivitamin use in the first trimester appeared to decrease the risk of (nonheritable) retinoblastoma in the child [39]. These findings suggest that gestational intake of one or more nutrients may influence risk. Folate and lutein/ zeaxanthin have been suggested as possibly protective as they are necessary for DNA methylation, synthesis, and/or retinal function [48]. Additional findings suggest that risk may be associated with intake of folate and with variants in metabolism of ingested folate and folic acid during key periods of development [49, 50].

In Vitro Fertilization (IVF) A study in the Netherlands estimated that children born after in vitro fertilization (IVF) had a five- to sevenfold increased risk of retinoblastoma [51]; however, results were not reported by form of retinoblastoma. In a population-based study extending the period of observation from the initial study by Moll et al., the incidence of retinoblastoma was not increased suggesting possible variations in effect with changing techniques in assisted reproduction [52]. Thus far, studies done in birth cohorts of children born after IVF in the UK, Denmark, France, and Australia have not reported increased incidence of retinoblastoma [53–56].

Inactivation of pRb Through Viral Complexes Some viral proteins bind to and inactivate pRb; thus, it is hypothesized that these viruses may contribute to the development of retinoblastoma. One such viral protein is the human papillomavi-

53

rus (HPV) protein, E7. In support of the viral hypothesis, DNA sequences from oncogenic HPV subtypes were detected in approximately onethird of retinoblastoma tumors studied in central Mexico [57]. In southern Brazil and northern Mexico, oncogenic HPV sequences were seen in similar proportions of retinoblastoma tumors [58, 59]. The oncogenic HPV subtypes found, 16, 18, 31, 33, 35, and 51, are causally associated with cervical cancer. Detection of HPV sequences in Central and South American tumor samples is particularly intriguing given the finding that the use of barrier methods of contraception around the time of conception was associated with lower risk of having a child with retinoblastoma [39].

Future Directions The field of epidemiology is increasingly incorporating precision-based elements that offer improved clinical subclassifications and molecularly defined subgroups. By incorporating these epidemiologic studies can inform our understanding of incidence as well as of disease progression. Investigators have long examined the incidence of retinoblastoma and believed that differences in stage at diagnosis were due to delays in diagnosis [60, 61]. Investigators in Argentina had examined the contributions of factors such as insurance and parental patterns of seeking care as contributing factors [62]. Other work had examined incidence comparing stage at presentation before and after publicity campaigns to raise awareness about retinoblastoma concluding that increased awareness shortened time to diagnosis leading to less advanced disease (in Honduras) [63]. However, this work had examined retinoblastoma without accounting for tumor laterality. Once tumor laterality was included, a Mexico-­based study found that extent of disease at presentation varied significantly with delays in obtaining medical attention in bilateral disease, but not in unilateral disease, even after accounting for maternal education [64]. Instead maternal education appeared to be a significant (independent) predictor of degree of invasive unilateral disease, with an inverse association (more advanced disease with less advanced mater-

54

nal education) and predicted survival in patients with unilateral disease [64]. Overall these suggest biologic differences in disease at time of development and a potential environmental impact during tumor progression. Additionally, epidemiologic studies suggest that disease is heterogeneous within lateralities. Lessons learned from advances in molecular biology have allowed examination using RNA sequencing which has revealed patterns of chromosomal loss and gain [65] and patterns of microRNA expression in subgroups of retinoblastoma tumors [66].

Summary Our knowledge of the role of the underlying risks and origin of the events leading to lack of functional pRb remains limited. Environmental exposures may contribute more to carcinogenesis when they occur during early eye development. The international variation in incidence suggests that nongenetic risk factors for the development of retinoblastoma may exist. The findings of the few studies that have investigated possible risk factors provide clues for further research. Based on our molecular understanding of the disease, we can begin to identify the critical time period (before vs. after conception) and the family member in which the critical event occurred for sporadic heritable and nonheritable retinoblastoma. Epidemiologic studies are now increasingly accounting for molecular differences in order to better identify the critical time period of exposure in the individuals at risk. Such studies will improve our understanding of disease development, potential risk factors, and potential strategies for prevention or for more effective treatment of retinoblastoma.

References 1. Knudson AG. Mutation and cancer: statistical study of retinoblastoma. PNAS. 1971;68:820–3. 2. Rushlow D, Piovesan B, Zhang K, et al. Detection of mosaic RB1 mutations in families with retinoblastoma. Hum Mutat. 2009;30(5):842–51.

M. Orjuela-Grimm et al. 3. Kanber D, Berulava T, Ammerpohl O, et al.  The human retinoblastoma gene is imprinted. PLoS Genet. 2009;5(12):e1000790. 4. Price EA, Price K, Kolkiewicz K, et al.  Spectrum of RB1 mutations identified in 403 retinoblastoma patients. J Med Genet. 2014;51(3):208–14. 5. Friend SH, Bernards R, Rogelj S, et al.  A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature. 1986;323:643–6. 6. Richter S, Vandezande K, Chen N, et al. Sensitive and efficient detection of RB1 gene mutations enhances care for families with retinoblastoma. Am J Hum Genet. 2003;72(2):253–69. 7. de Jong MC, Kors WA, de Graaf P, et al. The incidence of trilateral retinoblastoma: a systematic review and meta-analysis. Am J Ophthalmol. 2015;160(6):1116– 26.e5. 8. Stiller CA, Parkin DM. Geographic and ethnic variations in the incidence of childhood cancer. Br Med Bull. 1996;52(4):682–703. 9. Parkin DM, Kramarova E, Draper GJ, et al., editors. International incidence of childhood cancer. Lyon: International Agency for Research on Cancer; 1998. 10. Bravo-Ortiz J, Mendoza-Sanchez H, Fajardo Guttierrez A.  Algunas caracteristicas epidemiologicas del retinoblastoma en ninos residentes del Distrito Federal. Bol Med Hosp Infant Mex. 1996;53:234–9. 11. Steliarova-Foucher E, Colombet M, Ries LAG, et al. Indicators of data quality. Lyon: International Agency for Research on Cancer; 2017. [cited 2018 August 20]. Available from: http://iicc.iarc.fr/results/. 12. Steliarova-Foucher E, Colombet M, Ries L, et al., IICC-3contributors. International incidence of childhood cancer, 2001-10: a population-based registry study. Lancet Oncol. 2017;18(6):719–31. 13. Steliarova-Foucher E, Fidler MM, Colombet M, et al., ACCIS Contributors. Changing geographical patterns and trends in cancer incidence in children and adolescents in Europe, 1991–2010 (Automated Childhood Cancer Information System): a population-based study. Lancet Oncol. 2018;19(9):1159–69. 14. Methodology: standard country or area codes for statistical use (M49): United Nations. Available from: https://unstats.un.org/unsd/methodology/m49/. 15. Ries LAG, Smith MA, Gurney JG, et al., editors. Cancer incidence and survival among children and adolescents: United States SEER Program 1975– 1995, NIH Pub. No. 99-4649 ed. Bethesda: National Cancer Institute; 1999. 16. Broaddus E, Topham A, Singh AD. Incidence of retinoblastoma in the USA: 1975-2004. Br J Ophthalmol. 2009;93(1):21–3. 17. Siegel DA, Li J, Henley SJ, et al. Geographic variation in pediatric cancer incidence—United States, 2003– 2014. Morb Mortal Wkly Rep. 2018;67(25):707. 18. de Camargo B, de Oliveira Santos M, Rebelo MS, et al. Cancer incidence among children and adolescents in Brazil: first report of 14 population-based cancer registries. Int J Cancer. 2010;126(3):715–20.

4  Retinoblastoma: Incidence and Etiologic Factors 19. IARC-WHO.  International incidence of childhood cancer, vol. 3. Lyon: IARC-WHO; 2012. [cited 2012]. Available from: http://iicc.iarc.fr/about/index. php. 20. Wong JR, Tucker MA, Kleinerman RA, et al. Retinoblastoma incidence patterns in the US surveillance, epidemiology, and end results program. JAMA Ophthalmol. 2014;132(4):478–83. 21. Kamihara J, Ma C, Fuentes Alabi SL, et al. Socioeconomic status and global variations in the incidence of neuroblastoma: call for support of population-­ based cancer registries in low-­ middle-­ income countries. Pediatr Blood Cancer. 2017;64(2):321–3. 22. Martinez DE, Slack J, Beyerlein K, et al. The Migrant Border Crossing Study: a methodological overview of research along the Sonora-Arizona border. Popul Stud (Camb). 2017;71(2):249–64. 23. Kivela T. The epidemiological challenge of the most frequent eye cancer: retinoblastoma, an issue of birth and death. Br J Ophthalmol. 2009;93(9):1129–31. 24. Seregard S, Lundell G, Svedberg H, et al. Incidence of retinoblastoma from 1958 to 1998 in Northern Europe: advantages of birth cohort analysis. Ophthalmology. 2004;111(6):1228–32. 25. Kato MV, Ishizaki K, Shimizu T, et al. Parental origin of germ-line and somatic mutations in the retinoblastoma gene. Hum Genet. 1994;94(1):31–8. 26. Dryja TP, Mukai S, Petersen R, et al. Parental origin of mutations of the retinoblastoma gene. Nature. 1989;339(6225):556–8. 27. Allen JW, Ehling UH, Moore MM, et al.  Germ line specific factors in chemical mutagenesis. Mutat Res. 1995;330(1–2):219–31. 28. Czeizel A, Gardonyi J.  Retinoblastoma in Hungary, 1960–1968. Humangenetik. 1974;22(2):153–8. 29. Matsunaga E, Minoda K, Sasaki MS.  Parental age and seasonal variation in the births of children with sporadic retinoblastoma: a mutation-epidemiologic study. Hum Genet. 1990;84(2):155–8. 30. Moll AC, Imhof SM, Kuik DJ, et al. High parental age is associated with sporadic hereditary retinoblastoma: the Dutch retinoblastoma register 1862–1994. Hum Genet. 1996;98(1):109–12. 31. Pellie C, Briard ML, Feingold J, et al.  Parental age in retinoblastoma. Humangenetik. 1973;20(1): 59–62. 32. Morch ET.  Chondrodystrophic dwarfs in Denmark. Copenhagen: Ejnar Munksgaard; 1941. 33. Orioli IM, Castilla EE, Scarano G, et al.  Effect of paternal age in achondroplasia, thanatophoric dysplasia, and osteogenesis imperfecta. Am J Med Genet. 1995;59:209–17. 34. Blank CE.  Apert’s syndrome (a type of acrocephalosyndactyly)-observations on a British series of thirty-nine cases. Ann Hum Genet. 1960;24:151–64. 35. Moloney DM, Slaney SF, Oldridge M, et al. Exclusive paternal origin of new mutations in Apert syndrome. Nat Genet. 1996;13:48–53.

55 36. Vogel F, Rathenberg R. Spontaneous mutation in man. Adv Hum Genet. 1975;5:223–318. 37. Woodall AA, Ames BN.  Nutritional prevention of DNA damage to sperm and consequent risk reduction in birth defects and cancer in offspring. In: Bendich A, Deckelbaum RJ, editors. Preventive nutrition the comprehensive guide for health professionals. Totowa: Humana Press; 1997. p. 373–85. 38. Bunin GR, Felice MA, Davidson W, et al.  Medical radiation exposure and risk of retinoblastoma resulting from new germline RB1 mutation. Int J Cancer. 2011;128(10):2393–404. 39. Bunin GR, Meadows AT, Emanuel BS, et al. Pre- and postconception factors associated with sporadic heritable and nonheritable retinoblastoma. Cancer Res. 1989;49(20):5730–5. 40. Bunin GR, Tseng M, Li Y, et al. Parental diet and risk of retinoblastoma resulting from new germline RB1 mutation. Environ Mol Mutagen. 2012;53(6):451–61. 41. Bunin GR, Petrakova A, Meadows AT, et al. Occupations of parents of children with retinoblastoma: a report from the Children’s Cancer Study Group. Cancer Res. 1990;50(22):7129–33. 42. Abdolahi A, van Wijngaarden E, McClean MD, et al. A case-­control study of paternal occupational exposures and the risk of childhood sporadic bilateral retinoblastoma. Occup Environ Med. 2013;70(6):372–9. 43. Omidakhsh N, Ganguly A, Bunin GR, et al. Residential pesticide exposures in pregnancy and the risk of sporadic retinoblastoma: a report from the Children’s Oncology Group. Am J Ophthalmol. 2017;176:166–73. 44. Scheurer ME, Lupo PJ, Schuz J, et al. An overview of disparities in childhood cancer: report on the Inaugural Symposium on Childhood Cancer Health Disparities, Houston, Texas, 2016. Pediatr Hematol Oncol. 2018;35(2):95–110. 45. Ghosh JK, Heck JE, Cockburn M, et al.  Prenatal exposure to traffic-related air pollution and risk of early childhood cancers. Am J Epidemiol. 2013;178(8):1233–9. 46. Azary S, Ganguly A, Bunin GR, et al. Sporadic retinoblastoma and parental smoking and alcohol consumption before and after conception: a report from the Children’s Oncology Group. PLoS One. 2016;11(3): e0151728. 47. Heck JE, Park AS, Qiu J, et al.  Retinoblastoma and ambient exposure to air toxics in the perinatal period. J Expo Sci Environ Epidemiol. 2015;25(2):182–6. 48. Orjuela MA, Titievsky L, Liu X, et al. Fruit and vegetable intake during pregnancy and risk for development of sporadic retinoblastoma. Cancer Epidemiol Biomark Prev. 2005;14(6):1433–40. 49. Orjuela MA, Cabrera-Munoz L, Paul L, et al. Risk of retinoblastoma is associated with a maternal polymorphism in dihydrofolatereductase (DHFR) and prenatal folic acid intake. Cancer. 2012;118(23):5912–9. 50. de Lima EL, da Silva VC, da Silva HD, et al. MTR polymorphic variant A2756G and retinoblastoma risk in Brazilian children. Pediatr Blood Cancer. 2010;54(7):904–8.

56 51. Moll AC, Imhof SM, Cruysberg JR, et al. Incidence of retinoblastoma in children born after in-vitro fertilization. Lancet. 2003;361:309–10. 52. Marees T, Dommering CJ, Imhof SM, et al. Incidence of retinoblastoma in Dutch children conceived by IVF: an expanded study. Hum Reprod. 2009;24(12): 3220–4. 53. Bradbury BD, Jick H.  In vitro fertilization and childhood retinoblastoma. Br J Clin Pharmacol. 2004;58(2):209–11. 54. Bruinsma F, Venn A, Lancaster P, et al.  Incidence of cancer in children born after in-vitro fertilization. Hum Reprod. 2000;15(3):604–7. 55. Lidegaard O, Pinborg A, Andersen AN.  Imprinting diseases and IVF: Danish National IVF cohort study. Hum Reprod. 2005;20(4):950–4. 56. Foix-L'Helias L, Aerts I, Marchand L, et al. Are children born after infertility treatment at increased risk of retinoblastoma? Hum Reprod. 2012;27(7):2186–92. 57. Orjuela M, Ponce Castaneda V, Ridaura C, et al. Presence of human papilloma virus in tumor tissue from children with retinoblastoma: an alternative mechanism for tumor development. Clin Cancer Res. 2000;6:4010–6. 58. Montoya-Fuentes H, de la Paz Ramirez-Munoz M, Villar-Calvo V, et al. Identification of DNA sequences and viral proteins of 6 human papillomavirus types in retinoblastoma tissue. Anticancer Res. 2003;23(3C):2853–62. 59. Palazzi MA, Yunes JA, Cardinalli IA, et al. Detection of oncogenic human papillomavirus in sporadic retinoblastoma. Acta Ophthalmol Scand. 2003;81(4):396–8.

M. Orjuela-Grimm et al. 60. Wallach MBA, Munier F, Houghton S, et al. Shorter time to diagnosis and improved stage at presentation in Swiss patients with retinoblastoma treated from 1963 to 2004. Pediactrics. 2006;118:e1493–8. 61. Bai S, Ren R, Li B, et al. Delay in the diagno sis of retinoblastoma in China. Acta Ophthalmol. 2011;89(1):e72–4. 62. Chantada G, Fandino A, Manzitti J, et al. Late diagnosis of r­ etinoblastoma in a developing country. Arch Dis Child. 1999;80(2):171–4. 63. Leander C, Fu LC, Pena A, et al. Impact of an education program on late diagnosis of retinoblastoma in Honduras. Pediatr Blood Cancer. 2007;49(6):817–9. 64. Ramirez-Ortiz MA, Ponce-Castaneda MV, Cabrera-­ Munoz ML, et al.  Diagnostic delay and sociodemographic predictors of stage at diagnosis and mortality in unilateral and bilateral retinoblastoma. Cancer Epidemiol Biomark Prev. 2014;23(5):784–92. 65. Garcia-Chequer AJ, Mendez-Tenorio A, Olguin-­Ruiz G, et al. Overview of recurrent chromosomal losses in retinoblastoma detected by low coverage next generation sequencing. Cancer Genet. 2016;209(3):57–69. 66. Castro-Magdonel BE, Orjuela M, Camacho J, et al. miRNome landscape analysis reveals a 30 miRNA core in retinoblastoma. BMC Cancer. 2017; 17(1):458. 67. Steliarova-Foucher E, Colombet M, Ries LAG, et al., editors. International incidence of childhood cancer, vol. III (electronic version). Lyon: International Agency for Research on Cancer; 2017. Available from: http://iicc.iarc.fr/results/. Accessed 20 Aug 2018.

5

Retinoblastoma: An International Perspective Guillermo L. Chantada and Carlos A. Leal

Introduction Retinoblastoma represents a challenge in developing countries. While more than 90% of affected children survive in affluent societies, fewer children living in developing nations outlive this disease [1]. In this chapter we review some aspects of retinoblastoma regarding the incidence, delayed diagnosis, and challenges of the treatment in the developing countries.

Incidence It has been suggested that the incidence of nonheritable retinoblastoma may be higher in some developing countries, especially among the poorer populations. Increased incidence of retinoblastoma has been reported in tropical Brazil, south of Mexico, indigenous populations in Alaska, and some African countries [2–4]. A study reporting population-based data from a national registry in Argentina showed a comparable incidence to Western Europe and the United States of America (USA) [5]. However, there were some variations G. L. Chantada Hemato-oncology Department, Hospital JP Garrahan, Buenos Aires, Argentina C. A. Leal (*) Department of Oncology, Instituto Nacional de Pediatria, Mexico City, Mexico e-mail: [email protected] © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_5

within the country since there was a trend for higher incidence and lower survival in the areas with lower socioeconomic indicators. It was recently reported that the population from Hispanic origin in the USA had a higher incidence of retinoblastoma [6]. Also, these vulnerable populations have a poorer outcome in the USA [7]. Other reports from national or regional registries showed some variation in the incidence of retinoblastoma worldwide, but these differences were higher in other tumors such as neuroblastoma [8]. As reliable data on cancer incidence in many developing countries are usually lacking, these findings should be confirmed in larger, properly designed, population-­based studies. There is no sound explanation for this reported increased incidence. These authors suggest that variation in the incidence may be due to environmental factors. Results trying to link the human papilloma virus (HPV) to the pathogenesis of retinoblastoma led to controversial results [9– 11]. Low intake of fruits and vegetables during pregnancy also correlates with a higher risk of having a child with sporadic nonheritable retinoblastoma in Mexico [10].

Clinical Features  resenting Signs of Retinoblastoma P in Developing Countries Presenting signs of retinoblastoma vary depending where in the world the affected 57

58

G. L. Chantada and C. A. Leal

child lives (Fig. 5.1). Strabismus, a presenting sign in 20% of children in the USA, is not recognized as a presenting sign in Central Africa [12]. Proptosis due to orbital extension of retinoblastoma, which is rarely a presenting sign of retinoblastoma in the USA [13], is one of the commonest presenting signs in lowerincome countries [14, 15]. In middle-income countries, leukocoria is the most common presenting sign [16, 17]. In that setting, overt extraocular disease is rela-

tively uncommon, but patients still present with advanced intraocular disease as evidenced by choroidal or optic nerve invasion.

 xtraocular Retinoblastoma at E Presentation There is evidence that retinoblastoma presents more frequently with massive extraocular dissemination in developing countries (Fig. 5.2) [1]. It is important to recognize that these children usually present with severe malnutrition leading to cachexia and severe orbital pain in extreme cases, so treatment should include prompt supportive care. Delayed diagnosis is implicated as a major factor leading to extraocular dissemination and subsequent metastasis.

100 80 60 40

Delayed Diagnosis

20

Delayed diagnosis is a complex phenomenon in which patient- and physician-related factors and socioeconomic factors play a role.

0

SA

U

ge

,(A n)

so

a)

ad

s)

nt

ha

p)

ue

rig

(C

am

a,

od

al

un

br

in

(R

nt

il,

az

Ar

Br

)

bo

m

ai

G y,( ke

,(K go

on

r Tu

C

Proptosis

Strabismus

Leukocoria

Patient-related factors include symptoms of retinoblastoma in young children who are unable to express visual disturbances together with the lack

Fig. 5.1  Comparison of presenting signs of patients with retinoblastoma in different countries

a

Patient-Related Factors

b

Fig. 5.2  Patient with bilateral retinoblastoma and overt extraocular extension OS (a) which is confirmed by computed tomography (b)

59

5  Retinoblastoma: An International Perspective

of awareness of the general population that ocular abnormalities such as strabismus and leukocoria may be signs of cancer.

Physician-Related Factors Invariably, parents or other family members are the first to notice the visual abnormality. Pediatricians are frequently the first physicians to evaluate the affected child. It is rare for the pediatrician to detect leukocoria because of limited ophthalmic examination with undilated pupil in routine examinations. Therefore, they rarely recognize the significance of the parents’ complaints. As a result, many patients are not diagnosed or referred to an ophthalmologist on the first visit to the pediatrician. In a large cohort from Mexico, the majority of first contact physicians lacked basic information about retinoblastoma [18] (Fig. 5.3). A recent report from a referral center in

24% 32%

Brazil also showed that medical delay was responsible for up to 70% of the cases [19]. All these factors add critical weeks or months to the delay in diagnosis of retinoblastoma (Fig. 5.4). In Africa, most children present to the clinics and the nurses are the first to examine the child. If there is suspicion of abnormal examination or non-response to initial treatment, the child is referred then to the general doctor. Physician’s delay in the recognition of the symptoms was found to be the main reason in failing to recognize retinoblastoma which had the longest delay to diagnosis – 5 months – in a study in South Africa [20].

Socioeconomic Factors Socioeconomic factors such as parental education, lack of health insurance, and living in villages remote from large cities and human development indexes are significant risk factors for systemic dissemination of disease and ultimately survival [1]. Patient-related factors are not always associated with the low level of education of the parents or the socioeconomic conditions [20].

Survival

6%

8%

10%

20% Retinoblastoma

Infection

Retinoblastoma(no explanation to the parents)

Strabismus

No diagnosis,referred to specialist

It’s nothing

Fig. 5.3  Action taken by the pediatrician or ophthalmologist after the first consultation of a patient with retinoblastoma in Mexico

 urvival with Retinoblastoma Is S Lower in the Developing Countries Survival rates lower than 50% have been reported in lower-income countries [1]. There is a higher prevalence of extraocular disease in these settings justifying the poorer results, along with limited treatment options and poor compliance [21]. Since more than 80% of the world’s children live in developing countries, globally, there may be more children dying of retinoblastoma than surviving. In middle-income countries, significant advances have been made in the past decades, and survival rates greater than 80% are obtained in many countries.

60

G. L. Chantada and C. A. Leal Referral

First symptom

First consultation

First diagnosis Early stages

Time line (months)

1

2

3

4

5

6

7

8

9

10

11

12

13

14

Advanced stages First symptom

First consultation

First diagnosis

Referral

Fig. 5.4  Schematic representation of the diagnostic pathway of children with retinoblastoma in Mexico according to the disease extension

Steps to Improve Survival An improvement in the survival of patients with retinoblastoma in developing countries should not depend only on better treatment for extraocular disease. Rather, early detection and diagnosis with consequent reduction in systemic dissemination is expected to improve overall survival rate. A coordinated multistep approach involving public awareness, professional education, screening, but ultimately socioeconomic development is necessary. To be effective, resources must also aim to decrease the probability of treatment refusal.

 ublic Awareness Programs P In order to address this public health problem, some developing countries have embarked on public awareness programs about the signs and symptoms associated with retinoblastoma. One of the earliest and most important programs of this kind was developed in Brazil. To increase awareness of leukocoria as a presenting sign, the program targeted the general population through TV advertising and billboards. Other groups in Central America distributed pamphlets with information at vaccination centers and pediatricians’ offices [22]. The impact of these programs in the outcome of retinoblastoma is difficult to

estimate. Awareness campaigns should target populations where the leading problem is mortality because of metastatic retinoblastoma at presentation. Their role for the detection of early intraocular disease is unknown.

 rofessional Education Programs P An educational program to increase awareness of retinoblastoma among primary care physicians, especially those working in rural areas, has been established in some countries. In addition more detailed information about retinoblastoma has been inserted into the medical school curriculum. Screening for Retinoblastoma Retinoblastoma may be an ideal candidate for screening. Since children with a family history for retinoblastoma have been screened for many years by dilated examination under anesthesia, the natural history of the intraocular disease is well known; however in middle-income countries, most children with a family history of retinoblastoma are not screened [23]. Additionally, retinoblastoma presents in a narrow age range, constituting a well-defined target population to be screened. Because retinoblastoma occurs at an age when routine visits to the pediatrician are more common, these practitioners should proba-

5  Retinoblastoma: An International Perspective

bly be involved in the screening. Recently, an application for the use in cell phones was developed for detection of leukocoria [24]. This app is under evaluation in some programs as a tool for early diagnosis. However, the perfect test for screening and a proper program are still to be developed.

Minimizing Treatment Refusal Families refuse or withdraw treatment in as many as 30% of children diagnosed with intraocular retinoblastoma in many parts of the developing world [25]. Refusal of enucleation is the major cause of treatment withdrawal attests to many cultural and religious barriers to effective treatment of retinoblastoma that exist among indigenous populations in the developing world. Socioeconomic factors also play a large role, especially in health systems where medical care is not free of charge for the families [26]. Because families frequently must travel long distances to receive medical care for retinoblastoma, many choose, following diagnosis and treatment recommendations, to return home where the child dies. The lack of financial resources to support the family during a stay in the referral center as would be required for an extended course of chemoreduction is a common cause of treatment refusal. Also, there are other family members at home who must be cared for. Therefore, treatment programs must take all of these factors into consideration. Refusal of enucleation is more critical in children with bilateral retinoblastoma if the single remaining eye should be enucleated rendering the child blind. In many middle-income countries, this is the most common situation for refusal of enucleation affecting about 2% of the cases [27], but in lower-middle or lower-income countries, families often refuse enucleation even in unilateral cases. Compliance for follow-up when conservative therapy has been done is critical since failure to detect intraocular relapse may result in extraocular dissemination and death. Measures to reduce treatment withdrawal, adapting protocols to this situation, and early detection of familial cases are probably the most cost-­effective measures that can be

61

taken in many developing countries where treatment programs are well established.

Socioeconomic Development Socioeconomic development leading to the increased availability of high-quality health care may be the only sustainable way to reduce late diagnosis and ultimately the death rate from retinoblastoma.

Treatment Challenges in Developing Countries Because extraocular retinoblastoma is a rare event in developed countries, therefore there is only limited data on the treatment. Only a few large prospective trials of the treatment of systemic retinoblastoma from developed countries have been reported [28], but in the past years, cooperative groups in developing countries have been created and reported important information [29, 30]. Treatment of retinoblastoma in developing countries poses many challenges to the treating physicians.

 he Challenge of Conservative T Therapy Conservative therapy is seldom an option in unilateral disease in most developing countries and these patients are usually treated with enucleation of the affected eye, followed by adjuvant therapy if pathology risk factors are present. However, in middle-income countries, systemic chemoreduction followed by local consolidation for conservative therapy of bilateral disease has been implemented with success [31, 32]. More recently, intra-arterial chemotherapy has been successfully implemented in middle-income countries [27, 33]. It is possible that chemoreduction either intravenously or intra-arterially allowed for an increased eye preservation rate and the need for external beam radiotherapy has been dramatically reduced with intra-arterial chemotherapy [27]. No

62

increased mortality was reported with the use of intra-arterial chemotherapy in programs with clear selection criteria and adequate follow-up in Latin America [27, 34]. However, use of expensive equipment, frequent visits to the hospital, and the need for strict follow-­up are some of the factors that limit use of such treatments in the developing countries. Toxic mortality caused by systemic chemotherapy complications was reported, even from centers with relatively adequate resources [32]. In all these programs, an unanticipated problem associated with the introduction of the chemoreduction program was the dramatic increase in patient burden on the medical system. Because of all of these difficulties, local resources should be carefully evaluated before starting an eye conservative program in developing countries, and its ultimate benefit compared to the use of external beam radiotherapy has not been established in that setting.

The Challenge of Adjuvant Therapy Patients presenting with advanced disease involving the optic nerve, choroid, or sclera are more frequent in developing countries. Identification of such patients is critical because the use of adjuvant therapy is needed to improve their survival rate. However, the correct identification of such factors needs a specialized pathologist capable of analyzing comprehensively the eyeball following international standards. Consensus guidelines for the handling of enucleated eyes in order to identify and report uniformly pathology risk factors were recently published [35]. There is some controversy on which patients need adjuvant therapy after enucleation. It is undeniable that children in whom the tumor was not completely removed after enucleation, such as those with tumor beyond the resection margin of the optic nerve or those with trans-scleral extension, need adjuvant therapy. The need for adjuvant therapy for those children presenting with pathology risk factors in enucleated eyes that underwent complete resection of the tumor is more controversial. The use of adjuvant therapy for children with massive choroidal invasion or those

G. L. Chantada and C. A. Leal

with postlaminar optic nerve involvement or intrascleral invasion may improve survival results. However, in children with isolated choroidal invasion in whom the relapse rate is relatively low (about 4%), each center must balance between the risk of toxic mortality and the intention to reduce extraocular relapse by the use of adjuvant therapy. A recent prospective study showed that the relapse rate was 0% in children treated only with enucleation [34].

 he Challenge of Treatment of Overt T Extraocular Disease In most lower-income countries, children with retinoblastoma present with extensive dissemination to the orbit, usually in conjunction with metastatic dissemination to the CNS or to the bone marrow or bones. These severely affected children are not curable with current standard chemotherapy, but its use with an intent of life prolongation may be considered since retinoblastoma is a highly chemosensitive tumor. Excellent response to chemotherapy is seen in the overwhelming majority of the cases with low- to moderate-intensity chemotherapy and radiotherapy. In these settings it is important to discriminate between children with only orbital dissemination and those with metastatic disease performing extensive staging procedures because the former may survive with conventional therapy. Children with extensive orbital disease should not be treated with initial surgery, which would involve orbital exenteration (a mutilating and disfiguring procedure) since the tumor mass usually shrinks after a few cycles of chemotherapy allowing for a more conservative approach [36]. Encouraging survival results have been reported from Latin America with the implementation of high-dose chemotherapy and stem cell rescue for the treatment of metastatic retinoblastoma not involving the central nervous system [37]. However, those patients with central nervous system invasion are still not curable even with intensive treatment, so most centers in less developed countries offer palliative, life-­ prolonging therapy in these cases.

63

5  Retinoblastoma: An International Perspective

 evelopments that Provide Hope D for the Future Creation of Cooperative Groups Cooperative groups for the treatment of childhood cancer have been established in developing countries because of limited financial support and infrastructure. Recently, the Children’s Oncology Group in North America has launched clinical trial protocols that provide the framework for international applications [38]. In addition, cooperative groups for the treatment of retinoblastoma have been created in Mexico, Brazil, South America (GALOP), India, and Central America. The GALOP group reported that the overall survival of children with unilateral nonmetastatic retinoblastoma was 98 in their region [34]. The Central American AHOPCA group performed two prospective protocols for the management of retinoblastoma including a special branch for the management of families with compliance problems including a brief period of pre-enucleation chemotherapy [29]. The International Society of Pediatric Oncology (SIOP) developed a consensus guideline for graduated intensity treatment of retinoblastoma [39]. These developments should provide evidence-­ based treatment guidelines that will benefit children from developing countries.

International Collaborative Efforts Collaborative efforts between retinoblastoma centers in the northern and southern hemispheres have proved successful in improving outcomes in pediatric oncology [22]. The transfer of knowledge and resources is the main purpose or aim of these programs. The first program of this kind included cooperation between New  York City institutions, sponsored by the Fund for Ophthalmic Knowledge and Buenos Aires, Argentina [40]. This cooperation included donations of teaching material, participation in common research studies, and financial support for laboratory research. The International Network for Cancer Treatment and Research (www.inctr.

org) created a retinoblastoma group involving researchers from many different countries. The Global Pediatric Medicine Department of the St Jude Children Research Center (www.stjude.org) supports treatment of retinoblastoma for Central America based upon Internet transmission of digital images and online discussions, as well as an active teaching program [29]. Other programs include cooperation between national groups (Children’s Oncology Group and India) and hospitals (Children’s Hospital, Los Angeles and Mexico City; Institut Curie, Paris and North Africa [41], Canada, and Kenya). A web resource for locating retinoblastoma treatment centers has been launched. http://map.1rbw.org/

Summary Retinoblastoma presents unique challenges to treating physicians in developing countries. The burden of caring for 80% of the world’s retinoblastoma cases falls to individuals and national health-care systems with limited resources where caring for children with extraocular disease is relatively common. Retinoblastoma specialists from developing countries have taken the lead in creating a new International Staging system for extraocular retinoblastoma. Understanding the cause(s) of nonheritable or environmental retinoblastoma will likely take place in countries o­ utside of North America and Europe. The need for the cost-containment will lead to more effective and less expensive approaches. Initiatives that lead to early diagnosis and improve the quality of medical care of retinoblastoma patients in developing countries will be a valuable contribution to the rest of the world.

References 1. Canturk S, Qaddoumi I, Khetan V, et  al. Survival of retinoblastoma in less-developed countries impact of socioeconomic and health-related indicators. Br J Ophthalmol. 2010;94:1432–6. 2. Amozorrutia-Alegria V, Bravo-Ortiz JC, Vazquez-­ Viveros J, et al. Epidemiological characteristics of retinoblastoma in children attending the Mexican Social Security Institute in Mexico City, 1990–94. Paediatr Perinat Epidemiol. 2002;16:370–4.

64 3. Lanier AP, Holck P, Ehrsam Day G, et al. Childhood cancer among Alaska natives. Pediatrics. 2003;112:e396. 4. BenEzra D, Chirambo MC.  Incidence of retinoblastoma in Malawi. J Pediatr Ophthalmol. 1976;13:340–3. 5. Moreno F, Sinaki B, Fandino A, et al. A population-­ based study of retinoblastoma incidence and survival in Argentine children. Pediatr Blood Cancer. 2014;61:1610–5. 6. Friedrich P, Itriago E, Rodriguez-Galindo C, et al. Racial and ethnic disparities in the incidence of pediatric extracranial embryonal tumors. J Natl Cancer Inst. 2017;109(10). 7. Truong B, Green AL, Friedrich P, et al. Ethnic, racial, and socioeconomic disparities in retinoblastoma. JAMA Pediatr. 2015;169:1096–104. 8. Steliarova-Foucher E, Colombet M, Ries LAG, et al. International incidence of childhood cancer, 2001-­ 10: a population-based registry study. Lancet Oncol. 2017;18:719–31. 9. Orjuela M, Castaneda VP, Ridaura C, et al. Presence of human papilloma virus in tumor tissue from children with retinoblastoma: an alternative mechanism for tumor development. Clin Cancer Res. 2000;6:4010–6. 10. Orjuela MA, Titievsky L, Liu X, et al. Fruit and vegetable intake during pregnancy and risk for development of sporadic retinoblastoma. Cancer Epidemiol Biomark Prev. 2005;14:1433–40. 11. Anand B, Ramesh C, Appaji L, et  al. Prevalence of high-risk human papillomavirus genotypes in retinoblastoma. Br J Ophthalmol. 2011;95:1014–8. 12. Abdu L, Malami S.  Clinicopathological pattern and management of retinoblastoma in Kano, Nigeria. Ann Afr Med. 2011;10:214–9. 13. Abramson DH, Frank CM, Susman M, et al. Presenting signs of retinoblastoma. J Pediatr. 1998;132:505–8. 14. Menon BS, Alagaratnam J, Juraida E, et  al. Late presentation of retinoblastoma in Malaysia. Pediatr Blood Cancer. 2009;52:215–7. 15. Bekibele CO, Ayede AI, Asaolu OO, et al. Retinoblastoma: the challenges of management in Ibadan, Nigeria. J Pediatr Hematol Oncol. 2009;31:552–5. 16. Navo E, Teplisky D, Albero R, et al. Clinical presentation of retinoblastoma in a middle-income country. J Pediatr Hematol Oncol. 2012;34:e97–101. 17. Zhao J, Li S, Shi J, et al.  Clinical presentation and group classification of newly diagnosed intraocular retinoblastoma in China. Br J Ophthalmol. 2011;95:1372–5. 18. Leal-Leal CA, Dilliz-Nava H, Flores-Rojo M, et al. First contact physicians and retinoblastoma in Mexico. Pediatr Blood Cancer. 2011;57:1109–12. 19. Mattosinho CCS, Grigorovski N, Lucena E, et  al. Prediagnostic intervals in retinoblastoma: experience at an Oncology Center in Brazil. J Glob Oncol. 2017;3:323–30.

G. L. Chantada and C. A. Leal 20. Stefan DC, Siemonsma F. Delay and causes of delay in the diagnosis of childhood cancer in Africa. Pediatr Blood Cancer. 2011;56:80–5. 21. Pant G, Verma N, Kumar A, et al. Outcome of extraocular retinoblastoma in a resource limited center from low middle income country. Pediatr Hematol Oncol. 2017;34:419–24. 22. Leander C, Fu LC, Pena A, et al. Impact of an education program on late diagnosis of retinoblastoma in Honduras. Pediatr Blood Cancer. 2007;49:817–9. 23. Chantada GL, Dunkel IJ, Qaddoumi I, et al. Familial retinoblastoma in developing countries. Pediatr Blood Cancer. 2009;53:338–42. 24. Abdolvahabi A, Taylor BW, Holden RL, et  al. Colorimetric and longitudinal analysis of leukocoria in recreational photographs of children with retinoblastoma. PLoS One. 2013;8:e76677. 25. Sitorus RS, Moll AC, Suhardjono S, et al. The effect of therapy refusal against medical advice in retinoblastoma patients in a setting where treatment delays are common. Ophthalmic Genet. 2009;30:31–6. 26. Bonilla M, Rossell N, Salaverria C, et al. Prevalence and predictors of abandonment of therapy among children with cancer in El Salvador. Int J Cancer. 2009;125:2144–6. 27. Funes S, Sampor C, Villasante F, et  al. Feasibility and results of an intraarterial chemotherapy program for the conservative treatment of retinoblastoma in Argentina. Pediatr Blood Cancer. 2018;65:e27086. 28. Aerts I, Sastre-Garau X, Savignoni A, et al. Results of a multicenter prospective study on the postoperative treatment of unilateral retinoblastoma after primary enucleation. J Clin Oncol. 2013;31:1458–63. 29. Luna-Fineman S, Barnoya M, Bonilla M, et  al. Retinoblastoma in Central America: report from the Central American Association of Pediatric Hematology Oncology (AHOPCA). Pediatr Blood Cancer. 2012;58:545–50. 30. Leal-Leal C, Flores-Rojo M, Medina-Sanson A, et al. A multicentre report from the Mexican Retinoblastoma Group. Br J Ophthalmol. 2004;88:1074–7. 31. Antoneli CB, Ribeiro KC, Steinhorst F, et  al. Treatment of retinoblastoma patients with chemoreduction plus local therapy: experience of the AC Camargo Hospital, Brazil. J Pediatr Hematol Oncol. 2006;28:342–5. 32. Naseripour M, Nazari H, Bakhtiari P, et  al. Retinoblastoma in Iran: outcomes in terms of patients’ survival and globe survival. Br J Ophthalmol. 2009;93:28–32. 33. Grigorovski N, Lucena E, Mattosinho C, et  al. Use of intra-arterial chemotherapy for retinoblastoma: results of a survey. Int J Ophthalmol. 2014;7:726–30. 34. Perez V, Sampor C, Rey G, et al. Treatment of nonmetastatic unilateral retinoblastoma in children. JAMA Ophthalmol. 2018;136:747. 35. Sastre X, Chantada GL, Doz F, et  al. Proceedings of the consensus meetings from the International

5  Retinoblastoma: An International Perspective Retinoblastoma Staging Working Group on the pathology guidelines for the examination of enucleated eyes and evaluation of prognostic risk factors in retinoblastoma. Arch Pathol Lab Med. 2009;133: 1199–202. 36. Ali MJ, Reddy VA, Honavar SG, et al. Orbital retinoblastoma: where do we go from here? J Cancer Res Ther. 2011;7:11–4. 37. Palma J, Sasso DF, Dufort G, et al. Successful treatment of metastatic retinoblastoma with high-dose chemotherapy and autologous stem cell rescue in South America. Bone Marrow Transplant. 2012;47: 522–7. 38. Dunkel IJ, Krailo MD, Chantada GL, et al. Intensive multi-modality therapy for extra-ocular retinoblas-

65 toma (RB): A Children’s Oncology Group (COG) trial (ARET0321). J Clin Oncol. 2017;35:10506. 39. Chantada G, Luna-Fineman S, Sitorus RS, et al. SIOPPODC recommendations for graduated-­ intensity treatment of retinoblastoma in developing countries. Pediatr Blood Cancer. 2013;60:719–27. 40. Chantada GL, Dunkel IJ, Schaiquevich PS, et  al. Twenty-year collaboration between North American and South American retinoblastoma programs. J Glob Oncol. 2016;2:347–52. 41. Traore F, Sylla F, Togo B, et al. Treatment of retinoblastoma in Sub-Saharan Africa: experience of the paediatric oncology unit at Gabriel Toure Teaching Hospital and the Institute of African Tropical Ophthalmology, Bamako, Mali. Pediatr Blood Cancer. 2018;65:e27101.

6

Retinoblastoma Tumorigenesis Rachel C. Brennan and Michael A. Dyer

Introduction Tumorigenesis is a multistep process that involves sequential genetic alterations, only a small number of which may be relevant to malignant transformation [1]. Despite its relative rarity, retinoblastoma has been at the heart of many of the landmark discoveries that have advanced our understanding of the cellular events in tumorigenesis over the past several decades. The human retina is differentiated before completion of gestation, but if genetic or epigenetic aberrations allow persistent cell division, hyperplasia of immature cells may result in tumor formation through the early years of childhood. By studying the inheritance pattern of retinoblastoma, Knudson proposed a “two-hit” model to explain how a mutant “tumor suppressor” gene could be inherited as a dominant trait in which inactivation of the second, normal allele occurred in a susceptible somatic tissue such as the developing retina [2]. The Knudson hypothesis was confirmed by the cloning of the RB1 gene from retinoblastomas in 1986 by a team headed by Weinberg and Dryja [3]. As predicted by Knudson, one copy of R. C. Brennan (*) Department of Oncology, St. Jude Children’s Research Hospital, Memphis, TN, USA e-mail: [email protected] M. A. Dyer Department of Developmental Neurobiology, St. Jude Children’s Research Hospital, Memphis, TN, USA © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_6

the RB1 gene, located on chromosome 13q14, is mutated in the germline of susceptible individuals, whereas both copies of the gene are disrupted in the retinal tumors. Surprisingly, RB1 mutations subsequently were found in many other tumors unrelated to retinoblastoma, such as lung and breast cancers [4, 5], and the Rb protein is inactivated in the vast majority of all human cancers [6], indicating that the RB1 gene is broadly important as a tumor suppressor. While RB1 mutation leads to chromosomal instability (CIN) in cultured cells, recent evidence indicates the complexity of retinoblastoma tumorigenesis and importance of epigenetic mechanisms in regulation of gene expression. Comprehensive analyses and utilization of preclinical models of disease will further elucidate mechanisms of tumor progression and identify potential novel therapeutic targets for this rare disease.

Retinoblastoma Cell of Origin The cell of origin of retinoblastoma has been the subject of intense debate for many years. This concept is important for understanding why selected cells are susceptible to transformation when Rb is lost, how the initiating genetic mutation leads to clonal tumor expansion, and which cell types should be targeted for targeted ­molecular therapy [7, 8]. If retinoblastomas arise 67

68

R. C. Brennan and M. A. Dyer

from a specific cell type during a restricted period of retinal development, then the regulatory pathways that this specific process may provide highly directed targets for molecular therapy. For example, a small-molecule inhibitor of the Hedgehog pathway recently was shown to prevent medulloblastoma progression in a mouse model [9]. Genetic models of retinoblastoma have provided insights into the cell of origin of retinoblastoma. There are at least four possible cells of origin for retinoblastoma: a retinal stem cell, a retinal progenitor cell, a newly postmitotic cell committed or biased toward a particular retinal fate, or a differentiated neuron or glial cell (Fig. 6.1) [10]. Studies have suggested that there is no retinal stem cell in either the human or mouse neural retina [11, 12]. It is unlikely that a fully differentiated retinal neuron or glial cell gives rise to retinoblastoma, since the susceptibility to retinoblastoma is generally limited to a small window of time in embryonic development and early infancy prior to cell cycle exit and terminal differentiation in the developing retina [13]. Therefore the most likely candidates for the

Progenitor

b

a

Adult retina

cell of origin are a retinal progenitor cell or a newly postmitotic cell in the developing retina [14]. However, these results must be interpreted with caution in light of the fact that the genetic manipulations in these mice result in widespread disruption of normal retinal lamination, perturbing the normal position of retinal progenitor cells and newly postmitotic cells within the developing retina. One approach that is often used to identify cancer cell of origin is to analyze differentiation markers. The main assumption of this approach is that the normal cell type that expresses a given protein may be the cell of origin for a cancer in which that protein is expressed. Retinoblastomas have been shown to express photoreceptor-­ specific genes, which initially suggested this cell type as the cell of origin [15]. However, further analysis has shown that human retinoblastoma samples express a variety of other cell-specific markers [13]. Indeed, a more recent comparison of gene expression array data of human and mouse retinoblastomas suggests that retinoblastomas co-express multiple differentiation p­ athways that are normally

Horizontal

Cone

Fig. 6.1  Retinoblastoma arises during retinal development. (a) The mature retina is made up of seven major classes of retinal cell types (rods, cones, ganglion cells, bipolar cells, horizontal neurons, amacrine cells, and Müller glia). (b) During development, multipotent retinal progenitor cells produce each of the retinal cell types in an evolutionary conserved birth order. Retinal birth order is overlapping with horizontal cells, cones, and ganglion

Ganglion

Amacrine

Bipolar

Rod

Muller

cells born early during development and rods, bipolars, and Müller glia born at the end of retinogenesis. Retinoblastoma arises in the developing retina as progenitor cells produce retinal neurons. It is not known which cell type gives rise to retinoblastoma, but the tumors have a hybrid differentiation signature of progenitor cells, rods, cones, amacrine, and horizontal neurons

6  Retinoblastoma Tumorigenesis

incompatible during retinogenesis [16]. These indeterminate results reflect the difficulty in the differentiation marker approach to cell of origin studies; tumor cells that express different markers could have arisen from different cell types, or they may simply have arisen from the same multipotent progenitor cell at a different point in maturation. Further, gene expression changes in retinoblastoma, which is a developmental regulator, may reflect a nonspecific, deregulated developmental program initiated by the loss of Rb. While the focus remains on retinal progenitor cells and newly postmitotic cells as the possible retinoblastoma cell of origin, additional studies are required to definitively determine which cell type(s) require Rb to avoid cell cycle deregulation and malignant transformation. Unfortunately, genetically engineered mouse models (GEMMs) are not ideal for these studies because multiple Rb gene family members must be inactivated in the murine retina for tumors to form [10, 16]. Also, the tumors that do form in GEMMs have distinct gene expression and epigenomic profiles from human tumors suggesting that their cellular origins are distinct [17]. The field of stem cell research may offer an alternative path for investigating the developmental processes leading to retinoblastoma tumor formation. Somatic cells, such as fibroblasts from a skin biopsy or peripheral mononuclear blood cells, may be reprogrammed by the ectopic expression of transcription factors (Oct4, Sox2, Klf4, c-Myc) to generate induced pluripotent stem cells [18]. Human induced pluripotent stem cells (iPSCs) provide a unique opportunity to study fundamental questions regarding human retinal development and follow the lineage-­ specific cell production in vivo. The retina is an optimal system to investigate iPSCs further, as markers to distinguish each stage of retinogenesis are well-known. In theory, iPSCs from patients with germline RB1 mutations could be used to produce three-dimensional retinal organoids in culture for modeling human retinoblastoma tumor initiation in the laboratory. Moreover, patients with germline RB1 mutations are at increased risk for developing second-

69

ary malignancies [19], and differentiation of iPSCs from this population into other cell types could impact our understanding of how second malignancies develop.

Events in Retinoblastoma Progression Retinoblastoma Genomics Retinoblastoma presents in two distinct clinical forms. The first is heritable disease, which most commonly presents as bilateral and/or multifocal disease (one-third of cases), though a small percentage (~10%) of patients with unilateral disease are affected, and is characterized by the presence of germline mutations of the RB1 gene. Multifocal retinoblastoma may be inherited from an affected survivor (25%) or be the result of a new germline mutation (75%). The second is nonheritable disease, characterized by unilateral or unifocal disease (two-thirds of cases). Patients with a germline mutation can also be mosaic, with low-level mosaicism now identifiable through whole-genome sequencing. It is not known if these patients share the same increased likelihood of developing a second cancer later in life but bears close monitoring as clinical genomic efforts are launched for this and other pediatric malignancies. RB1 inactivation through point mutations (nonsense, frameshift, splicing, missense, or promoter mutations) are most often the first hit, while the second “mutation” is more likely to be chromosomal in nature (i.e., loss of heterozygosity). Nonsense and frameshift are the most common germline and somatic mutations, though no mutational hotspots have been identified [20]. There is no clear clinical correlation between the type of mutation and the severity of disease [21]. While the initiating genetic event in most retinoblastoma tumors  – biallelic inactivation of the RB1 gene – is well established, much less is known about subsequent genetic events that contribute to retinoblastoma progression. There are recurrent somatic mutations in BCOR, but it is not known how those lesions contribute to ­ retinoblastoma

70

R. C. Brennan and M. A. Dyer

progression [22]. Also, MYCN amplification is found in retinoblastoma and many tumors express high levels of MYCN irrespective of the gene copy number, but it has not yet been demonstrated that MYCN promotes tumorigenesis in the context of biallelic inactivation of RB1 [23].

Recurrent Chromosomal Abnormalities in Retinoblastoma RB1 plays an important role in regulating proper cohesion and segregation of chromosomes in human retinoblastoma [22]. In many cancers with suppressed RB pathways, an increase in mitotic defects, aneuploidy, and subsequent chromosomal instability (CIN) supports tumor progression [24–27]. Recent whole-genome sequencing of retinoblastomas showed that some retinoblas-

tomas have very few mutations or chromosomal alterations other than RB1 and BCOR mutations (Figs. 6.2 and 6.3) [22]. Moreover, the passage of orthotopic xenografts in the eyes of immunocompromised mice retains stable genomes [22]. Taken together, these data suggest that genomic instability and the subsequent genetic lesions that may result from such instability are not required for retinoblastoma progression [22]. The few regions of the genome that are recurrently gained or lost in retinoblastoma involve known oncogenes and tumor suppressor genes, and changes in copy number correlate with gene expression. Studies initially performed in retinoblastoma cell lines to identify these regions have now been expanded to validate key aberrations and identify candidate genes in those regions as potential drivers of retinoblastoma progression [28, 29].

a

b

Fig. 6.2  Whole-genome sequencing of retinoblastoma. (a) CIRCOS plots of the whole-genome sequence data for representative retinoblastoma primary tumors (SJRB001 and SJRB002) and an orthotopic xenograft derived from

one of those tumors (SJRB001X). (b) Beyond RB1, the only recurrently mutated gene was BCOR, an epigenetic regulator

71

6  Retinoblastoma Tumorigenesis

a

b

c

d

Fig. 6.3  Integrated epigenetic analysis of retinoblastoma. (a) Plot of the epigenetically deregulated genes in retinoblastoma. These data are integrated from ChIP-on-­ chip, gene expression, and DNA methylation results. The spleen tyrosine kinase gene (SYK) was epigenetically

upregulated in human retinoblastoma. (b) Validation of increase mRNA expression for SYK and validation of upregulation of SYK protein in retinoblastoma (c). Immunohistochemistry of 82 human retinoblastomas (d) showed upregulation of SYK in 100% of tumors

Chromosome 2p (MYCN amplification) MYCN amplification is present in many retinoblastoma tumors, and additional tumors have high levels of MYCN expression but irrespective of their gene copy number [23].

which places a role in intracellular transport and cell division, was noted in retinoblastoma patients with an older age at diagnosis [35]. The role of MDM4 has been described above.

Isochromosome 6p (E2F3 and DEK) Isochromosome 6p gain is found in 45% of retinoblastomas and is the most common change observed by comparative genomic hybridization (CGH, 54%) [28, 30, 31]. The minimal region of gains was narrowed to band p22, with DEK and E2F3, as potential targets. Both genes are known to be overexpressed in human malignancies. E2F3 amplification [32] may disrupt the p53 response that is normally triggered by loss of Rb. Chromosome 1q (KIF14 and MDM4) The genes KIF14 and MDM4 have been validated in the minimal common region gained (1q31–1q32), with two CGH studies showing gains in 1q were associated with advanced tumors in older children [33, 34]. Furthermore, overexpression of KIF14,

Chromosome 16 (CDH11) One-third of retinoblastoma cases showed loss of all or part (16q) of chromosome 16 in CGH studies, identifying CDH11 as the candidate tumor suppressor gene in the region [28]. This gene encodes cadherin-11, which mediates calcium-­ dependent cell-cell adhesion. Loss of CDH11 and changes in cell adhesion has been implicated in invasion of the optic nerve in retinoblastoma patients [36].

 scaping Death and Promoting E Clonal Expansion: p53 Pathway in Retinoblastoma Human retinoblastomas typically have a very high rate of apoptosis, suggesting that the apop-

72

totic response is still intact but that proliferation is simply outstripping apoptosis [20]. A major unexplained question is why loss of Rb does not trigger an overwhelming apoptotic response that eliminates nascent retinoblastoma cells before clonal expansion can occur. In most cancers, there are mutations in the p53 tumor suppressor or other members of the p53 pathway that explain the acquired resistance to apoptosis [6]. Further, in mouse models of retinoblastoma, tumor development is greatly enhanced when p53 is inactivated [37]. However, there is no evidence that p53 is mutated frequently in human retinoblastomas [38]. Further, p53 can be activated in retinoblastomas, suggesting that the protein is functional [39]. Therefore, retinoblastoma may represent a unique exception in which the p53 pathway is intact. There are several potential explanations for how retinoblastoma tumor cells with functional p53 protein can circumvent apoptosis: (1) p53 is intact but functionally inactivated through another mechanism; (2) there may be genetic events in other apoptosis-related pathways, such as the Bcl2 pathway, to promote survival; or (3) the retinoblastoma cell of origin may be naturally resistant to apoptosis [40]. Other members of the p53 pathway can be disrupted in cancer, leading to functional inhibition of p53. One example is increased expression of p53 antagonists such as MDM2 and MDM4 [41]. MDM4 inhibits cell cycle arrest and apoptosis by binding to the transcriptional activation domain of the p53 tumor suppressor protein and inhibiting its activity. Increased expression may be achieved through somatic gene amplification, germline polymorphisms at the MDM2 or MDM4 locus, or splice variants of MDM4 messenger RNAs (mRNAs) producing stable forms of MDM4 that subsequently suppress p53 [42]. Analysis of human retinoblastoma revealed that MDM4 was amplified in 65% of the tumors, negatively correlating with p53 levels [43]. Furthermore, gene expression array analysis of 52 human retinoblastoma tumors showed that MDM4 was expressed at high levels irrespective of the MDM4 copy number [16], suggesting that MDM4 expression may be elevated through

R. C. Brennan and M. A. Dyer

mechanisms unrelated to gene copy number, such as the presence of small nucleotide polymorphisms (SNPs). A genotype study of MDM4 polymorphisms found a higher frequency of SNP rs116197192G, though no correlative study with gene and protein expression was performed [44]. Another study found no correlation between MDM4 SNP7 (rs1563828) and gene expression, but high levels of MDM4 protein expression were present in all samples with SNP34091 A/A allele (rs4245739) [45]. This SNP was identified in a study by Wynendaele and colleagues, who found this SNP creates a putative target site for miR-191 with the C-allele [46]. The SNP34091-A allele is not efficiently recognized by miR-191, and this in turn leads to increased MDM4 protein expression and increased risk of high-grade carcinoma [46]. Further analysis to determine the role, if any, of the ANP34091 A/A allele and tumor progression in retinoblastoma patients will require analysis in a larger cohort.

Retinoblastoma Epigenomics While inactivation of the RB1 gene in retinoblastoma may result in defects in sister chromatid cohesion [36], this does not necessarily lead to genomic instability. Instead, it has been shown that inactivation of the RB1 gene leads to massive epigenetic deregulation of known oncogenes and tumor suppressors, and this may contribute to the rapid progression of the disease following the second mutational event in RB1 [22]. Two epigenetic mechanisms with importance in RB tumorigenesis are microRNAs (miRNAs) and DNA methylation.

miRNAs miRNAs function in transcriptional and posttranscriptional regulation of gene expression, and several miRNAs are potential candidate c­ omponents of key oncogenic and tumor suppressor networks in retinoblastoma [29]. The most widely studied tumor suppressor miRNAs are members of the let-7 family, which is involved in repressing members of the Ras family, HMGA2, and c-Myc oncogenes. miR-34a is thought to play a role in

6  Retinoblastoma Tumorigenesis

the p53 pathway, functioning as a tumor suppressor in the normal retina. The miR-­17~92 cluster (OncomiR-1) is a well-­characterized family of miRNA genes that promote proliferation, inhibit differentiation, increase angiogenesis, and sustain cell survival [47]. Studies have shown overexpression of members of this cluster in primary retinoblastoma tumors and cell lines, as well as increased survival of retinoblastoma cell lines treated with miRNA inhibitors. From this cluster, miR-17 and miR-20 have been identified as putative therapeutic targets for prevention and/or treatment of retinoblastoma [48].

DNA Methylation RB1 promoter methylation, first discovered as methylation of a CpG island (CpG 106) overlapping the RB1 promoter and exon 1  in 1989 [49], was later confirmed and correlated with decreased RB1 gene expression [50, 51], providing the first evidence of an epigenetic component to retinoblastoma tumorigenesis. Several subsequent studies have identified methylation status of other known tumor suppressor genes in retinoblastoma. Evidence of hypermethylation of RASSF1A (RAS-associated domain family 1A), O [6]-methylguanine-­DNA (MGMT), and p16INK4A in retinoblastoma tumors is confirmed, with potential clinical implications for MGMT (associated with advanced-stage [52]) and p16INK4A (a potential heritable susceptibility marker [53]). Moving beyond a restricted view of hypermethylation in tumor suppressor genes, a genomewide promoter methylation analysis comparing 19 primary retinoblastoma tumor samples with six normal fetal retinae was performed. This analysis identified 118 genes with differential expression and correlative DNA methylation profiles: 35 genes with promoter hypermethylation and gene expression downregulation, and 83 genes with promoter hypomethylation and gene overexpression [22]. Furthermore, an integrative epigenetic analysis utilizing chromatin immunoprecipitation-on-chip, DNA methylation analysis, and gene expression arrays identified 60 genes with differential expression correlating with histone modifications and DNA methylation

73

profiles. Among the 10 known cancer genes, 3 were upregulated (TFF1, SYK, and MCM5), with downregulation of ASLC1, DTNND1, SOX2, ADAMTS18, GLI3, PCDH11X, and DKK1. Previously of unknown importance in retinoblastoma tumorigenesis, upregulation of the tyrosine kinase SYK was shown to be required for tumor survival, with no function of this kinase in the developing visual system and no recurrent genetic lesions in SYK to suggest it is a driver of tumorigenesis. Furthermore, preclinical modeling of SYK inhibition resulted in reduced tumor growth both in vitro and in vivo. However, additional preclinical testing with pharmacokinetic analysis revealed the chosen SYK antagonist (R406) failed to reach a tumoricidal concentration in the vitreous, regardless of route of administration, thus limiting the viability of this drug candidate in future clinical trials [54].

Uncommon RB Genesis 13q Syndrome RB1 gene is located on chromosome 13q14.2. Patients with deletion involving the entire region of 13q14.2 are considered as a separate group with 13q deletion syndrome. These patients present with a broad spectrum of clinical features, including mild to moderate developmental delays, dysmorphic features (microcephaly, typical facies, malformations of the brain, genitourinary and gastrointestinal tract), and growth retardation, that vary based on the location and extent of the genetic deletion. All patients with 13q deletion should be routinely screened for retinoblastoma, though disease is not fully penetrant in this population. A correlation of genotype and phenotype in 13q deletion retinoblastoma patients showed milder phenotypic expression of ­retinoblastoma with larger deletions (>1 Mb) that contained the MED4 gene [55]. In addition, during systemic therapy for retinoblastoma, patients with 13q deletion syndrome had more prolonged neutropenia resulting in delayed chemotherapy and more frequent dose reductions, though overall outcome was not impacted compared with non-13q deletion counterparts [56].

R. C. Brennan and M. A. Dyer

74

 YCN Amplification (Chromosome 2p) M The concept of retinoblastoma tumor formation driven by a single oncogenic event rather than biallelic inactivation of RB1 was initially proposed by Gallie et  al. [57]. Their report that showed MYCN amplification without any RB1 mutations (RB1+/+MYCNA) in a small subset of retinoblastoma tumors (~1%) was confirmed with whole-genome sequencing analysis in a separate cohort of 46 primary retinoblastoma tumors [23]. The few cases that have been identified seem to have distinct clinical phenotypes (younger age at diagnosis) compared with other non-familial RB1 driven retinoblastoma [57], but further evaluation is needed to determine how and if genotype affects phenotype and prognosis for this rare cohort. Chromothripsis A chromosomal, regional, and focal CNVs and LOH analysis of 94 human retinoblastoma samples identified a somewhat higher rate of chromosomal, regional, and focal lesions in 11% of tumors [23]. These ten tumors with no identifiable RB1 germline mutation were further characterized by IHC and FISH analysis. An acute genomic event, chromosomal “shattering” (called chromothripsis), was identified in 30% of the samples, suggesting this mechanism may initiate retinoblastoma by inactivating the RB1 gene in these tumors. This is important because chromothripsis at the RB1 locus results in inactivation of the gene, but is not detected by conventional methods of RB1 gene analysis. Specifically, exon sequence analysis, promoter methylation analysis, analysis of LOH, and copy number changes would all appear to be wild type in a tumor where RB1 is inactivated by chromothripsis. All the exons of RB1 would be present, the promoter would be hypomethylated, both copies of RB1 would be present, and any copy number changes would be minimal. Thus, the tumor would appear to be wild type for RB1, but it would actually have a gene inactivation. Whole-genome sequencing combined with fluorescence in situ hybridization is currently the only way to identify retinoblastoma samples with chromothripsis.

Preclinical Models of Retinoblastoma The first genetically engineered animal model of spontaneous retinoblastoma was a transgenic mouse model in which the oncogenic T antigen from the SV40 virus was expressed in the retina (Chap. 7) [58]. T antigen inhibits the Rb protein, providing an explanation for the retinal tumors, but it also inhibits Rb family members p107 and p130, as well as the p53 tumor suppressor and many other proteins. Therefore, this model was not ideal for studying the molecular genetics of retinoblastoma as it occurs in humans. Intriguingly, when another mouse transgenic model was developed in which Rb was inhibited by E7, a viral oncoprotein encoded by human papillomavirus that does not inhibit p53, retinoblastomas did not develop unless the mice were bred into a p53-null background. In an attempt to reconcile these observations, some investigators postulated that p53 or another anti-apoptotic gene must be mutated in human retinoblastomas. In the search for a more accurate genetic model of retinoblastoma, several groups generated mice in which one copy of the RB1 gene was nonfunctional, thereby replicating the situation of patients with heritable retinoblastoma [59–61]. Surprisingly, however, these mice developed pituitary tumors, but none developed retinoblastoma. The first clue to solving this apparent inconsistency between human and mouse retinoblastoma was provided by workers in the Berns lab who showed that deletion of RB1 in the mouse retina leads to retinal tumors if the Rb family member p107 was also deleted [62]. Subsequent work confirmed these findings and showed that loss of RB1 in the mouse (but not humans) is compensated by upregulation of p107 [63], thus explaining the apparent contradiction between mouse and human susceptibility to retinoblastoma. These findings led to the generation of the first knockout mouse model of retinoblastoma [10], which was confirmed and extended by two other groups [40, 64]. These new, more accurate genetic models of retinoblastoma, one modeling the important epigenetic contribution of MDMX

75

6  Retinoblastoma Tumorigenesis

overexpression seen in human retinoblastoma tumors and the other mimicking p53 inactivation, are yielding important new insights into retinoblastoma biology and are proving to be instrumental in the development of novel treatments for patients with this disease. As a complement to genetic mouse models, orthotopic xenografts provide a resource to directly study human retinoblastoma tumors in a preclinical setting. Human tumors are provided at the time of enucleation (removal of the eye), as patients with retinoblastoma do not undergo biopsy of the primary tumor site to confirm the diagnosis due to the risk of tracking tumor cells outside the globe. Most often, this surgery occurs in patients with unilateral retinoblastoma where the contralateral eye is unaffected by disease. This population is most likely to harbor the sporadic form of retinoblastoma with no germline RB1 mutation. In addition, normal developmental pathways may be altered in tumor cells as a consequence of tumorigenesis [5]. In patients with bilateral disease, an eye may be removed due to progression of disease, which often occurs only after therapy has been started, exposing the tumor cells to the selective pressure of toxic systemic or locally delivered agents. Therefore, despite an impressive tumor bank and library of orthotopic xenografts, our understanding of retinoblastoma tumorigenesis is limited by the rarity of retinoblastoma cases and tumor tissue available for examination. Furthermore, the genetic and orthotopic xenografts do not provide an adequate model for the precise sequence and timing of cellular and molecular events that occur during retinoblastoma tumorigenesis. As noted previously, the field of stem cell research (induced pleuripotent stem cells) may offer an alternative path for investigating the developmental processes leading to retinoblastoma tumor formation.

Clinical Implications While retinoblastoma and the cellular events leading to tumor formation in the retina have served as an important model for cancer biology,

these advances have had not had the expected impact on the clinical management of retinoblastoma. This deficiency is due in part to preclinical models of retinoblastoma that did not accurately recapitulate all aspects of the human disease, such as vitreous and subretinal seeding, which are the most common causes of treatment failure in humans but do not occur in currently available genetically engineered mouse models. Newer animal models have provided novel insights into retinoblastoma biology and will continue to be of use as more detailed studies of the effect of therapy on second primary tumors, a major concern in human retinoblastoma, are undertaken [65– 67]. The development of human orthotopic xenografts of retinoblastoma provides yet another useful model for testing novel therapeutics [22]. These are particularly useful when incorporated into a comprehensive preclinical testing paradigm that recapitulates many of the clinical and therapeutic approaches used to treat children with retinoblastoma [68]. However, translation of discoveries from bench to bedside has been limited. Access to drug development for this rare disease has slowed the progress of inhibitors of the MDM2/MDM4 pathway until recently. The preclinical studies of R406, an SYK inhibitor, identified problems achieving adequate drug concentrations within the eye, regardless of delivery method, before this drug ever reached (and failed in) a pediatric trial [54]. Targeting the SYK/MCL1 pathway with other therapeutic agents, such as BCL2 antagonists, or targeting the epigenetic machinery with HDAC inhibitors, may still prove viable. These novel molecular targeted therapeutics combined with the unique opportunities for local drug delivery provide promising new avenues for continued preclinical and clinical research in the coming years.

References 1. Hahn WC, Weinberg RA. Modelling the molecular circuitry of cancer. Nat Rev Cancer. 2002;2(5):331–41. 2. Knudson AG Jr. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A. 1971;68(4):820–3.

76 3. Friend SH, Bernards R, Rogelj S, et  al. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature. 1986;323(6089):643–6. 4. Harbour JW, Lai SL, Whang-Peng J, et  al. Abnormalities in structure and expression of the human retinoblastoma gene in SCLC.  Science. 1988;241(4863):353–7. 5. Lee EY, To H, Shew JY, et al. Inactivation of the retinoblastoma susceptibility gene in human breast cancers. Science. 1988;241(4862):218–21. 6. Sherr CJ, McCormick F. The RB and p53 pathways in cancer. Cancer Cell. 2002;2(2):103–12. 7. Lapidot T, Sirard C, Vormoor J, et al. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature. 1994;367(6464):645–8. 8. Pardal R, Clarke MF, Morrison SJ. Applying the principles of stem-cell biology to cancer. Nat Rev Cancer. 2003;3(12):895–902. 9. Romer JT, Kimura H, Magdaleno S, et al. Suppression of the Shh pathway using a small molecule inhibitor eliminates medulloblastoma in Ptc1(+/−)p53(−/−) mice. Cancer Cell. 2004;6(3):229–40. 10. Zhang J, Schweers B, Dyer MA.  The first knockout mouse model of retinoblastoma. Cell Cycle. 2004;3(7):952–9. 11. Coles BL, Angenieux B, Inoue T, et al. Facile isolation and the characterization of human retinal stem cells. Proc Natl Acad Sci U S A. 2004;101(44):15772–7. 12. Tropepe V, Coles BL, Chiasson BJ, et  al. Retinal stem cells in the adult mammalian eye. Science. 2000;287(5460):2032–6. 13. DiCiommo D, Gallie BL, Bremner R. Retinoblastoma: the disease, gene and protein provide critical leads to understand cancer. Semin Cancer Biol. 2000;10(4):255–69. 14. Dyer MA, Bremner R. The search for the retinoblastoma cell of origin. Nat Rev Cancer. 2005;5(2):91–101. 15. Nork TM, Schwartz TL, Doshi HM, et al. Retinoblastoma. Cell of origin. Arch Ophthalmol. 1995;113(6):791–802. 16. McEvoy J, Flores-Otero J, Zhang J, et  al. Coexpression of normally incompatible developmental pathways in retinoblastoma genesis. Cancer Cell. 2011;20(2):260–75. 17. Aldiri I, Xu B, Wang L, et  al. The dynamic epigenetic landscape of the retina during development, reprogramming, and tumorigenesis. Neuron. 2017;94(3):550–68.e10. 18. Pankratz MT, Li XJ, Lavaute TM, et  al. Directed neural differentiation of human embryonic stem cells via an obligated primitive anterior stage. Stem Cells. 2007;25(6):1511–20. 19. Kleinerman RA, Tucker MA, Tarone RE, et al. Risk of new cancers after radiotherapy in long-term survivors of retinoblastoma: an extended follow-up. J Clin Oncol. 2005;23(10):2272–9. 20. Brantley MA Jr, Harbour JW.  The molecular biology of retinoblastoma. Ocul Immunol Inflamm. 2001;9(1):1–8.

R. C. Brennan and M. A. Dyer 21. Ali MJ, Parsam VL, Honavar SG, et  al. RB1 gene mutations in retinoblastoma and its clinical correlation. Saudi J Ophthalmol. 2010;24(4):119–23. 22. Zhang J, Benavente CA, McEvoy J, et  al. A novel retinoblastoma therapy from genomic and epigenetic analyses. Nature. 2012;481(7381):329–34. 23. McEvoy J, Nagahawatte P, Finkelstein D, et al. RB1 gene inactivation by chromothripsis in human retinoblastoma. Oncotarget. 2014;5(2):438–50. 24. Manning AL, Longworth MS, Dyson NJ. Loss of pRB causes centromere dysfunction and chromosomal instability. Genes Dev. 2010;24(13):1364–76. 25. Hernando E, Nahle Z, Juan G, et  al. Rb inactivation promotes genomic instability by uncoupling cell cycle progression from mitotic control. Nature. 2004;430(7001):797–802. 26. Amato A, Lentini L, Schillaci T, et al. RNAi mediated acute depletion of retinoblastoma protein (pRb) promotes aneuploidy in human primary cells via micronuclei formation. BMC Cell Biol. 2009;10:79. 27. Iovino F, Lentini L, Amato A, et al.  RB acute loss induces centrosome amplification and aneuploidy in murine primary fibroblasts. Mol Cancer. 2006;5:38. 28. Corson TW, Gallie BL.  One hit, two hits, three hits, more? Genomic changes in the development of retinoblastoma. Genes Chromosomes Cancer. 2007;46(7):617–34. 29. Benavente CA, McEvoy JD, Finkelstein D, et  al. Cross-species genomic and epigenomic landscape of retinoblastoma. Oncotarget. 2013;4(6):844–59. 30. Squire J, Phillips RA, Boyce S, et al. Isochromosome 6p, a unique chromosomal abnormality in retinoblastoma: verification by standard staining techniques, new densitometric methods, and somatic cell hybridization. Hum Genet. 1984;66(1):46–53. 31. Potluri VR, Helson L, Ellsworth RM, et  al. Chromosomal abnormalities in human retinoblastoma. A review. Cancer. 1986;58(3):663–71. 32. Grasemann C, Gratias S, Stephan H, et al. Gains and overexpression identify DEK and E2F3 as targets of chromosome 6p gains in retinoblastoma. Oncogene. 2005;24(42):6441–9. 33. Lillington DM, Kingston JE, Coen PG, et  al. Comparative genomic hybridization of 49 primary retinoblastoma tumors identifies chromosomal regions associated with histopathology, progression, and patient outcome. Genes Chromosomes Cancer. 2003;36(2):121–8. 34. Herzog S, Lohmann DR, Buiting K, et  al. Marked differences in unilateral isolated retinoblastomas from young and older children studied by comparative genomic hybridization. Hum Genet. 2001;108(2):98–104. 35. Madhavan J, Coral K, Mallikarjuna K, et  al. High expression of KIF14  in retinoblastoma: association with older age at diagnosis. Invest Ophthalmol Vis Sci. 2007;48(11):4901–6. 36. Laurie N, Mohan A, McEvoy J, et al. Changes in retinoblastoma cell adhesion associated with optic nerve invasion. Mol Cell Biol. 2009;29(23):6268–82.

6  Retinoblastoma Tumorigenesis 37. Howes KA, Ransom N, Papermaster DS, et  al. Apoptosis or retinoblastoma: alternative fates of photoreceptors expressing the HPV-16 E7 gene in the presence or absence of p53. Genes Dev. 1994;8(11):1300–10. 38. Kato MV, Shimizu T, Ishizaki K, et  al. Loss of heterozygosity on chromosome 17 and mutation of the p53 gene in retinoblastoma. Cancer Lett. 1996;106(1):75–82. 39. Nork TM, Poulsen GL, Millecchia LL, et  al. p53 regulates apoptosis in human retinoblastoma. Arch Ophthalmol. 1997;115(2):213–9. 40. Chen D, Livne-bar I, Vanderluit JL, et al. Cell-specific effects of RB or RB/p107 loss on retinal development implicate an intrinsically death-resistant cell-of-origin in retinoblastoma. Cancer Cell. 2004;5(6):539–51. 41. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat Med. 2004;10(8):789–99. 42. Bond GL, Hu W, Bond EE, et al. A single nucleotide polymorphism in the MDM2 promoter attenuates the p53 tumor suppressor pathway and accelerates tumor formation in humans. Cell. 2004;119(5):591–602. 43. Laurie NA, Donovan SL, Shih CS, et al. Inactivation of the p53 pathway in retinoblastoma. Nature. 2006;444(7115):61–6. 44. de Oliveira Reis AH, de Carvalho IN, de Sousa Damasceno PB, et al. Influence of MDM2 and MDM4 on development and survival in hereditary retinoblastoma. Pediatr Blood Cancer. 2012;59(1):39–43. 45. McEvoy J, Ulyanov A, Brennan R, et  al. Analysis of MDM2 and MDM4 single nucleotide polymorphisms, mRNA splicing and protein expression in retinoblastoma. PLoS One. 2012;7(8):e42739. 46. Wynendaele J, Bohnke A, Leucci E, et  al. An illegitimate microRNA target site within the 3' UTR of MDM4 affects ovarian cancer progression and chemosensitivity. Cancer Res. 2010;70(23):9641–9. 47. Olive V, Jiang I, He L. mir-17-92, a cluster of miRNAs in the midst of the cancer network. Int J Biochem Cell Biol. 2010;42(8):1348–54. 48. Nittner D, Lambertz I, Clermont F, et  al. Synthetic lethality between Rb, p53 and Dicer or miR-17-92 in retinal progenitors suppresses retinoblastoma formation. Nat Cell Biol. 2012;14(9):958–65. 49. Greger V, Passarge E, Hopping W, et  al. Epigenetic changes may contribute to the formation and spontaneous regression of retinoblastoma. Hum Genet. 1989;83(2):155–8. 50. Sakai T, Toguchida J, Ohtani N, et al. Allele-specific hypermethylation of the retinoblastoma tumor-­ suppressor gene. Am J Hum Genet. 1991;48(5):880–8. 51. Ohtani-Fujita N, Fujita T, Aoike A, et al. CpG methylation inactivates the promoter activity of the human retinoblastoma tumor-suppressor gene. Oncogene. 1993;8(4):1063–7. 52. Choy KW, Lee TC, Cheung KF, et  al. Clinical implications of promoter hypermethylation in RASSF1A and MGMT in retinoblastoma. Neoplasia. 2005;7(3):200–6.

77 53. Indovina P, Acquaviva A, De Falco G, et  al. Downregulation and aberrant promoter methylation of p16INK4A: a possible novel heritable susceptibility marker to retinoblastoma. J Cell Physiol. 2010;223(1):143–50. 54. Pritchard EM, Stewart E, Zhu F, et al. Pharmacokinetics and efficacy of the spleen tyrosine kinase inhibitor r406 after ocular delivery for retinoblastoma. Pharm Res. 2014;31(11):3060–72. 55. Mitter D, Ullmann R, Muradyan A, et al. Genotype-­ phenotype correlations in patients with retinoblastoma and interstitial 13q deletions. Eur J Hum Genet. 2011;19(9):947–58. 56. Brennan RC, Qaddoumi I, Billups CA, et al. Patients with retinoblastoma and chromosome 13q deletions have increased chemotherapy-related toxicities. Pediatr Blood Cancer. 2016;63(11):1954–8. 57. Rushlow DE, Mol BM, Kennett JY, et  al. Characterisation of retinoblastomas without RB1 mutations: genomic, gene expression, and clinical studies. Lancet Oncol. 2013;14(4):327–34. 58. Windle JJ, Albert DM, O'Brien JM, et  al. Retinoblastoma in transgenic mice. Nature. 1990;343(6259):665–9. 59. Jacks T, Fazeli A, Schmitt EM, et  al. Effects of an Rb mutation in the mouse. Nature. 1992;359(6393):295–300. 60. Lees E, Faha B, Dulic V, et  al. Cyclin E/cdk2 and cyclin A/cdk2 kinases associate with p107 and E2F in a temporally distinct manner. Genes Dev. 1992;6(10):1874–85. 61. Clarke AR, Maandag ER, van Roon M, et  al. Requirement for a functional Rb-1 gene in murine development. Nature. 1992;359(6393):328–30. 62. Robanus-Maandag E, Dekker M, van der Valk M, et  al. p107 is a suppressor of retinoblastoma development in pRb-deficient mice. Genes Dev. 1998;12(11):1599–609. 63. Zhang J, Gray J, Wu L, et al. Rb regulates proliferation and rod photoreceptor development in the mouse retina. Nat Genet. 2004;36(4):351–60. 64. MacPherson D, Sage J, Kim T, et al. Cell type-specific effects of Rb deletion in the murine retina. Genes Dev. 2004;18(14):1681–94. 65. Abramson DH, Melson MR, Dunkel IJ, et al.  Third (fourth and fifth) nonocular tumors in survivors of retinoblastoma. Ophthalmology. 2001;108(10):1868–76. 66. Eng C, Li FP, Abramson DH, et  al. Mortality from second tumors among long-term survivors of retinoblastoma. J Natl Cancer Inst. 1993;85(14):1121–8. 67. Fletcher O, Easton D, Anderson K, et  al. Lifetime risks of common cancers among retinoblastoma survivors. J Natl Cancer Inst. 2004;96(5):357–63. 68. Brennan RC, Federico S, Bradley C, et al. Targeting the p53 pathway in retinoblastoma with subconjunctival Nutlin-3a. Cancer Res. 2011;71(12): 4205–13.

7

Animal Models in Retinoblastoma Research Thomas A. Mendel and Anthony B. Daniels

Background The biology of retinoblastoma is covered more extensively elsewhere in this text and is only outlined here for the benefit of the subsequent discussion of animal models. Briefly, the retinoblastoma protein (pRb) functions by suppression of the G1 to S cell cycle transition. It is a highly conserved pathway seen in animals and plants [1]. The retinoblastoma gene (RB1) codes for pRb, which ordinarily blocks cell cycle progression by inhibiting transcription factor E2F and thereby preventing transcription of multiple gene products important for S-phase transition. When a cell readies for division, pRb becomes hyperphosphorylated and thereby inactivated by cyclin E, removing its latent inhibition of E2F.  Without inhibition of E2F, numerous cellular division and DNA synthesis-related genes with E2F-responsive promoters are transcribed (Fig. 7.1). Additionally, p107 and p130, known as “pocket proteins,” function similarly to and in parallel with pRb in their T. A. Mendel Department of Ophthalmology and Visual Sciences, Vanderbilt Eye Institute, Nashville, TN, USA A. B. Daniels (*) Department of Ophthalmology and Visual Sciences, Vanderbilt Eye Institute, Nashville, TN, USA Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN, USA e-mail: [email protected] © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_7

inhibition of E2F chromatin complexes, preventing transcription. This understanding of the pRb pathway and cellular function has allowed multiple genetic models of pRb to be developed, which recapitulate portions  – although not all  – of the human retinoblastoma tumor phenotype in preclinical animal models.

Models of Retinoblastoma As with any disease, particularly one underpinned by mutation, genetically engineered model species have proven helpful in our understanding of retinoblastoma. In humans, retinoblastoma is generally thought of as having a single susceptibility gene, mutation of which is responsible for the vast majority of human retinoblastoma. In fact, the RB1 gene is actually named the “retinoblastoma susceptibility gene.” It is therefore surprising that knocking out RB1 in mice does not result in the human phenotype. RB1 knockout mice do not form ocular tumors. This is likely due to species-specific characteristics, different tissue restriction of gene expression, or variable escape pathways. Instead, mouse models have been created by knocking out different or additional genes, with mixed results. Mouse models have been generated that more closely recapitulate the human phenotype of retinal tumors via (sometimes elaborate) genetic designs, combining RB1 knockout with loss of other tumor suppressors, such as p53 or p107. These models, approached in various ways, are discussed below. 79

80 Fig. 7.1  Model of retinoblastoma homeostasis, mutation, and knockout. (a) Normal homeostasis. Rb prevents E2F from binding to promoters. Cyclin E phosphorylates Rb, freeing E2F to promote DNA synthesis-­ related targets. p53 can check transition from G1 to S if not bound by MDM2. (b) If Rb mutated, E2F proceeds with synthesis-related promotion unchecked. MDM2 is upregulated in patients, inhibiting p53. (c) In the knockout model, p107, Rb, and p53 are all mutated, allowing for E2F to proceed unchecked. (d) In the viral mutation model, T antigen blocks E2F binding with Rb and also inactivates p53, allowing for unchecked proliferation

T. A. Mendel and A. B. Daniels a. Normal homeostasis Transcription of E2F target genes blocked by Rb

Phosphorylation of Rb dissociates it from E2F, allowing for E2F-mediated transcription of target genes.

MDM2 inhibits p53 by binding to it and targeting it for ubiquitination p53

p107 (p107 mimics Rb and binds E2F)

Cyclin E

Cyclin E

Rb

p

p53

E2F Rb

p107

If unchecked growth noted, p53 arrests cell at G1/S transition or targets cell for apoptosis

E2F

MDM2

transcription of E2F target genes

Regulated transcription with transition from G1 to S phase of cell cycle

G1/S Transition Cell Cycle G1

S

b. Retinoblastoma mutation MDM2 upregulated, which inhibits p53, leading to unchecked proliferation.

Rb mutated. p107 not sufficient to inhibit E2F in humans. p107 does inhibit E2F in mice. p107

p53

Rb

MDM2

E2F

MDM2

p53

MDM2

Rb

p107

MDM2

E2F

transcription of E2F target genes

PROLIFERATION

G1/S Transition Cell Cycle G1

S

c. Rb1/p53/p107 knockout model Rb and p107 Knock out: Unregulated E2F promotion of transcription targets

p107

Cyclin E

(can mimic Rb end bind E2F)

Unchecked growth would be arrested by p53, but it too is knocked out. p53 does not check proliferation.

p107

Cyclin E

Rb

P

p53

E2F Rb

E2F

transcription of E2F target genes

PROLIFERATION

G1/S Transition Cell Cycle G1

S

d. LHbTAG model T-antigen prevents E2F from forming complexes with Rb and p107 p107

T-antigen inactivates p53 E2F unbound by Rb or p107

E2F Rb

p53

MDM2

E2F

transcription of E2F target genes

PROLIFERATION

G1/S Transition Cell Cycle G1

S

7  Animal Models in Retinoblastoma Research

Early Adenoviral Models Early attempts to produce intraocular retinoblastoma tumors in animal models proved challenging, necessitating adenoviral infection of embryonic rats. A previously used adenovirus was noted to generate medulloblastomas in the central nervous system of hamsters [2]. It was postulated that this approach could be used to generate tumors elsewhere in the central nervous system, including in the retina. SpragueDawley rats were injected monocularly with adenovirus Ad12, generating intraocular tumors that resembled medulloblastomas. Similarly, in nonhuman primates, intravitreal injection of adenovirus Ad12 generated intraocular tumors [3]. However, “take” rates were low, with only 10% of the injected primate eyes developing tumors. In addition, the tumor that developed was closely related to medulloepitheliomas rather than retinoblastoma. Intravitreal injection in newborn rat eyes is technically challenging. Baboons offered larger eyes that were easier to inject; however, increased cost, slow generation time for breeding, and more complicated procedural protocols limited their use. For all these reasons, these early models were not widely adopted, fell out of use, and were eventually replaced by xenograft and genetic mouse models of retinoblastoma.

Genetic Models of Retinoblastoma As mentioned above, knocking out RB1 alone does not generate the human retinoblastoma phenotype in mice. Despite significant prior advances in our understanding of RB1 and its importance as a tumor suppressor gene in humans, there was no suitable knockout mouse that recapitulated the ocular phenotype for laboratory experimentation. Moreover, attempts at heterozygous or homozygous knockout in murine models proved unsuccessful at producing ocular tumors or were lethal when the knockout was delivered in a germline fashion to mouse embryos. Astonishingly, RB1 heterozygous knockout mice developed pituitary adenomas and thyroid carcinomas, but no ocular

81

retinoblastoma tumors [4], distinctly different from patients with (heterozygous) germline/ familial disease. To overcome the hurdle of either producing no phenotype or a lethal one, mosaic animal models were generated by early embryonic transfection, with some of the cells in the organism containing the germline mutations but some cells remaining wild type. However, even this approach failed to produce intraocular retinoblastoma tumors that closely resembled the human phenotype [5].

 b1/p53/p107 Knockout Mouse Model R Zhang et  al. generated the first non-xenograft murine model of retinoblastoma by combining RB1 knockout with knockout of another tumor suppressor, p53, and confining those mutations to the eye [6, 7]. Rather than taking the approach of creating a genetic mosaic through transfection of viral DNA into embryonic cells, the researchers utilized the Cre-lox genetic splice system. In this system, researchers insert DNA that codes for Cre, a protein capable of excising DNA elsewhere in the genome, and drive the transcription and translation of that Cre with a tissue-specific promoter. Once transcribed and translated, the Cre protein is capable of scanning DNA for lox sites. These lox sites are often engineered to flank a gene that is to be excised. Alternatively, if the goal is to express a newly inserted gene, a stop codon upstream of the gene of interest may be “loxed,” such that excision of this stop codon leads to expression of the inserted gene. In this way, researchers can “turn on” the Cre in a specific tissue to excise a native gene to create a tissue-­specific knockout or excise a stop codon thus permitting a newly inserted gene to be transcribed. Researchers can also transfect fluorescent proteins that indicate when a cell has undergone a genetic excision event, to facilitate cellular tracking and to identify tissues expressing the genetic event of interest. Functionally, this creates an elegant method by which ­researchers can excise critical genes such as RB1 or p53 and restrict this gene deletion to the retina, by choosing a promoter known only to function in the retina. The gene to be excised continues to function normally in the rest of the

82

body, allowing for proper development and avoiding embryonic lethality. Zhang et al. generated mice that drive Cre in post-mitotic retinal bipolar cells with heterozygous knockout of both p53 and RB1 after multiple generations of inbreeding for each trait individually. The remaining p53 and RB1 alleles were bounded by lox sites. As mentioned above, the pocket protein p107 functions in parallel with pRb to inactivate E2F in the binding pocket of E2F. It was thought that RB1 deletion alone did not cause retinal tumors in the murine model, as the mouse was able to make better use of parallel, redundant signaling with p107 to compensate for loss of RB1. Thus, p107 was heterozygously knocked out as well. In total, when Cre began expression, the remaining functional alleles of p53 and RB1 were spliced out of the genome specifically in the mouse retinal bipolar cells, based on use of a cell type-specific promoter. The retinal progenitor cells therefore lacked p53, pRb, and p107. These cells failed to exit the cell cycle, thus forming bilateral ocular tumors. Histologically, these tumor cells invaded the deep retina by 6 weeks of age, and the entire eye was filled with tumor at 14 weeks. Rosettes were formed, noted to be a hallmark of the human retinoblastoma phenotype. Occurring commonly in retinoblastoma, rosettes are radially oriented circular clusters of cells with an exterior and an inner lumen that contains cytoplasmic extensions from the surrounding cells [8]. This is thought to represent an immature attempt by the cells to establish polarity as found in the external limiting membrane of the retina, resulting in their radial aggregation. Flexner-Wintersteiner rosettes do not demonstrate fiber-rich neuropil in the lumen, while the more cytologically differentiated Homer Wright rosettes do. These features stand out on histological examination, and their presence in tissue from a mouse tumor supports the validity of that model of retinoblastoma. Because this model requires deletion of three tumor suppressor genes to generate a phenotype that occurs with only a single mutation in humans, it is difficult to study conclusively the effect of RB1 mutation on tumorigenesis in this model.

T. A. Mendel and A. B. Daniels

However, while the p53 gene is not directly mutated or otherwise lost during the development of retinoblastoma tumors in humans, the p53 pathway is posttranscriptionally suppressed through MDM2/MDMX.  In this way, the p53 knockout component of this model actually does recapitulate the signaling pathways seen in patient tumors. Practically speaking, the RB1/ p53/p107 knockout mouse model has become a consistent workhorse in the field of retinoblastoma research over the years. These mice have been used to study various aspects of disease, including retinal cell biology [9], response to chemotherapy [10], and tumor epigenetics [11].

LHβTAG Mouse Model Another important tumor model was created by O’Brien and colleagues. Interestingly, they did not initially set out to create an ocular tumor model but were rather attempting to generate a pituitary tumor model. They noted that mice developed ocular tumors when simian virus T-antigen (SV40-TAG), driven by the human luteinizing hormone β-subunit promoter, was introduced into fertilized single-cell oocytes. Rather than pituitary tumors, the model produced “trilateral” tumors including bilateral ocular tumors arising by 1–2 months as well as pineal gland tumors. The ocular tumors were typically bilateral and multifocal within each eye, similar to patients with germline disease. The histology of the murine tumors resembled that of human retinoblastoma, with rosettes and a high nuclear to cytoplasmic ratio [12]. Additionally, these tumors demonstrated the ability to invade the CNS directly by invasion of the optic nerve into the brain, similar to the optic nerve invasion seen with advanced human disease [13, 14]. In addition, approximately 15% of mice developed pineal tumors, akin to the pinealoblastomas that occur in 3% of patients with germline retinoblastoma. Despite the apparent successful recapitulation of the human disease phenotype in this model, the tumors that develop in these retinas do not express opsin, a hallmark of human retinoblastoma [15, 16]. Subsequent analysis indicated that the tumors arose from Mueller cells [17],

7  Animal Models in Retinoblastoma Research

which do not normally express opsin. Analysis of patient tumor tissues indicated that opsin was expressed in >95% of cells in human disease [18]. The LHβTAG mouse model also only simulated the clinical phenotype of germline RB1 mutations, generating multifocal tumors. A solitary retinal tumor is more commonly seen in human cases of sporadic retinoblastoma, as opposed to germline mutation cases. Various iterations of this model followed, in an attempt to recapitulate more faithfully these human disease features. In an effort to recapitulate the solitary tumor phenotype most commonly seen in humans, subsequent experiments drove SV40TAG by different retina-­specific promoters, producing unifocal ocular tumors that involve the whole photoreceptor layer and demonstrate rosettes along with pineal gland tumors [19–21]. Researchers also attempted to target the rod opsin promoter, thinking that this would more align with human disease. Unexpectedly, when applying the same SV40-TAG to the rod opsin promoter, retinal tumors failed to develop in this model. Instead, extensive photoreceptor loss developed beginning at postnatal day 1–5 [22]. Despite these limitations, the LHβTAG mouse model, with its reproducible bilateral multifocal retinal tumors, has proven to be an invaluable tool for retinoblastoma research. It has been used to study tumor response to radiation therapy [23], combination radiotherapy and hyperthermia [24], as well as local and systemic carboplatin therapy [25–27].

 ell of Origin and Validity of Mouse C Models While both of these murine models recapitulate various aspects of intraocular retinoblastoma tumor phenotype, the tumors in each model arise for a different cell of origin. While tumors that arise from a different cell of origin than human disease may “look” like human retinoblastoma, the validity of deductions about tumor biology made from its use could be questioned. Thus, one would think that the tumor that arises from the same cell of origin as human retinoblastoma would be the model that would gain primacy. However, the true cell of origin of human retino-

83

blastoma remains elusive as well, with much controversy and various lines of contradictory evidence. For decades, retinoblastoma researchers have sought an answer to one of the most fundamental questions of any solid tumor: what is the cell of origin of retinoblastoma? This question is particularly challenging due to the complex architecture of the retina with an array of cell types. Considering the different paths that researchers have taken to generate useful retinoblastoma models, it is reasonable to inquire which model – if any – most faithfully recapitulates not only the clinical phenotype but also the cell biology of human retinoblastoma. Recent work by Xu et al. indicates that the likely cell of origin might be the cone precursor cell. Xu and colleagues tested human embryonic retinal cells, finding that the cone precursor cells in particular are exquisitely sensitive to loss of pRb, while the remainder of the cell types that were tested went on to develop normally even after deletion of RB1 [18, 28]. Their work was largely carried out in vitro with dissociated retinal cells, with subsequent xenografts into the subretinal space of athymic mice, which corroborated their results. Despite clinical evidence that tumors in human patients arise first from the inner nuclear layer  – rather than the outer nuclear layer where cone cells would be expected to arise – the data by Xu et al. is compelling and useful in the ongoing debate about the ultimate cell of origin [29]. The RB1/p53/p107 knockout model generates unchecked cellular proliferation in bipolar cells, with nuclei found in the inner nuclear layer, consistent with in vivo optical coherence tomography imaging of human tumors that show early growth of nascent tumors in the inner nuclear layer [30]. The LHβTAG mouse model was found most closely to resemble differentiated Mueller cells [17]. Neither appears to arise from cone precursor cells. Conversely, tumors arising from cone precursor cells would not be expected to generate tumors in the inner nuclear layer and are thus inconsistent with the in  vivo OCT imaging findings. The debate continues, and as long as it does, no mouse model can claim primacy based on faithful recapitulation of the human cell of origin.

84

T. A. Mendel and A. B. Daniels

Table 7.1  Comparison of the two most commonly used genetic mouse models of retinoblastoma Genetic mouse model Rb1/p53/p107 knockout mouse model

Method of Age of generation tumor onset Histology 6 weeks Tumors Serial Cre/ formed in lox knockout deep retina and + rosettes inbreeding

LHβTAGmouse model

SV40 – T antigen

1–2 months Rosettes, high N:C ratio

Considering these two genetic models, the researcher must weigh the advantages and disadvantages of each in experimental design. Both produce ocular tumors reliably at 1–2 months of age with histologic demonstration of rosettes. The RB1/p53/p107 knockout mouse isolates tumors to the eye; however, it does so only with additional mutation of p53 and p107 knockout, above and beyond that of human retinoblastoma. The LHβTAG model impacts p53 signaling primarily through its T antigen but also impacts E2F/ pRb signaling, although the mechanism is not yet fully elucidated. The particular pathways targeted in each genetic model are shown in Fig. 7.1. The particular differences between the two genetic models, and the pathways targeted in each genetic model, are outlined in Table  7.1. Unfortunately, neither model appears to arise from cone precursor cells, which demonstrate a compelling case for being the putative oncogenic cells of origin in human disease.

Tumorigenic cell Advantages Disadvantages Undifferentiated Ocular specific Requires bipolar cells tumors co-knockout of p53 and p107 in mice – not required in humans Impacts p53 Can generate Likely signaling and E2F trilateral proliferating signaling. Rb tumors, like Mueller cells intact human germline phenotype

injected into the eye (orthotopic xenografts) or elsewhere (heterotopic xenografts). Of course, a truly orthotopic xenograft would require intraretinal implantation [31], which is difficult and often approximated with subretinal injection, leading to subsequent invasion of adjacent retina [32]. Intravitreal injection is truly orthotopic in the sense that the cells are in the same location as the vitreous seeds seen with advanced intraocular retinoblastoma. Several reports have documented that intravitreally injected cell lines faithfully recapitulate the architecture of vitreous seeds seen in patients [31, 32]. Xenograft-based animal models are widely used to test proposed chemotherapeutic agents, where correctly predicting the response of human retinoblastoma tumors is critical.

Location of Xenograft Xenografting allows for a more physiologically appropriate three-dimensional environment to study cell biology, rather than studying the cells in two-dimensional culture in  vitro. The deciXenograft Models of Retinoblastoma sion as to whether to use orthotopic xenoengraftment (with cells injected into the retina To render moot the issue of cell of origin and to or the vitreous) or heterotopic xeno-engraftment ensure that the tumor developing in the mouse (usually the subcutaneous space of the flank) is faithfully recapitulates human disease, human typically made for experimental feasibility contumors may be directly implanted into non-­ siderations. Orthotopic xenografts more accuimmunocompetent animals. This can be done rately simulate the native environment of the with validated human retinoblastoma cell lines tumor, but the injections required are techniderived from human tumors or with xenografts cally challenging, less reproducible, and the derived from recently harvested primary human size of the tumor is ultimately limited by the tissue from retinoblastoma tumors (patient-­ small size of the eyes of small animals. Tests of derived xenografts, PDXs). Tumor cells can be extraocular extension and tumor metastasis are

7  Animal Models in Retinoblastoma Research

more faithfully simulated with orthotopic transplantation. In addition, assessment of local therapies, such as intravitreal, periocular, and intra-arterial chemotherapy, requires orthotopic xenografts. Alternatively, heterotopic xenografts are technically easier to perform, allow tumors to grow to a size more similar to that seen in humans, and allow for tumor expansion into multiple animals. While heterotopic xenografts are sometimes adequate for early-stage assessments of systemic therapies (such as intravenous chemotherapy), investigators need to remember that ocular drug penetration is much lower than the flank, and achieving adequate tumor drug levels is often a limiting factor of systemic chemotherapies. Thus, a drug might prove promising at the levels achieved within a heterotopic xenograft, only to fail to show efficacy against a tumor within the eye.

Choice of Xenografted Tissue To study retinoblastoma in the laboratory, researchers typically select either immortalized cell lines derived from past retinoblastoma patient tumors or new PDX lines. One type of immortalized cell, Y79 cells, is widely used in the literature for in vitro and in vivo experiments. Originally derived from the primary tumor of a 2.5-year-old Caucasian female, these cells are immortalized and readily available commercially. In practice, these tumors are invasive and simulate aggressive disease in a clinical setting. WERI cells, another commonly used human retinoblastoma cell line, have also been applied orthotopically and heterotopically in xenografts [33]. WERI cells (and their subsequent derivatives) were originally derived from a 1-year-old Caucasian girl in 1974. By comparison, WERI cells form more solitary and discrete tumors than their more invasive Y79 counterparts. Lastly, PDX tumors better recapitulate the intra-tumoral heterogeneity seen within human tumors, rather than the more homogenous cell lines. Of course PDX tumors derived from different patients also demonstrate significant inter-tumoral variability, decreasing the degree of reproducibility of any particular experiment. On the other hand, this increased intra-tumoral heterogeneity and inter-­

85

tumoral variability allows researchers to study multiple patients’ tumors for response to chemotherapy. Importantly, only the tumor cells themselves derive from humans. The vasculature and any immune infiltrate (see below) derive from the host species.

 reventing Rejection of the Xenograft P by the Immune System Regardless of whether the xenograft is created orthotopically or heterotopically, and regardless of whether one injects a human-derived cell line or a PDX, the native immune system must be silenced. Without such immunosuppression, the epitopes presented by the transplanted cells from another species are sufficient to elicit a non-self-­ recognition of the host immune system and trigger a robust immune response that results in rejection of the grafted tumor cells. Xenografts will simply not “take” in the setting of an intact recipient immune system. Immunocompromised and Immunologically Naïve Hosts Disruption of the immune system can be achieved through genetic manipulation to create immunocompromised hosts, through exogenous pharmacologic immunosuppression or through introduction of the xenograft into a system that has not yet matured immunologically at the time of xenotransplantation. One popular genetic background is severe combined immunodeficiency syndrome (SCID), such as the widely used nonobese diabetic (NOD)/SCID mouse [34]. NOD/SCID mice demonstrate impaired T and B cells with a variation adding a mutation to the interleukin 2 receptor gamma chain resulting in the NOD/SCID/Gamma (NSG) that also removes natural killer cells. Athymic nude mice do not develop a thymus and thus do not generate T cells. Rag2 mice are homozygous knockouts for recombination activating gene 2 (Rag2), which is required for the molecular epitope recombination that occurs during B- and T-cell maturation in lymphoid tissues, rendering them devoid of those lineages. In contrast, one can take advantage of the species-specific fact that newborn rats do not yet have a mature immune sys-

86

T. A. Mendel and A. B. Daniels

tem at the time of birth, and thus human tissue Mechanistically, cyclosporine works to prevent can be injected. The immune system then devel- the proliferation of T cells, primarily by inhibitops to recognize the human antigens as “self” ing the calcineurin phosphatase pathway and [35]. In all these models, researchers need to con- inactivating DNA transcription factors that assist sider how the immune system normally modu- in T-cell proliferation. By using cyclosporine, lates disease development, aggressiveness, and investigators remove the effective T-cell response metastasis and be cognizant that a non-­ from the host, allowing for viable xenografts. immunocompetent xenograft system may there- While cyclosporine has been used to allow human fore not always exactly mimic human disease retinoblastoma cells to grow in rats [43], cyclobehavior. sporine immunosuppression is most often used to Xenografting human tumor cells into immu- allow xenografts to grow in larger species, where nocompromised hosts has been performed for a immunocompromised models do not exist as wide variety of investigations, including the alternatives. study of nanoparticle treatment [36], oncolytic This approach is most commonly used in rabviral therapy [37], and basic investigation of the bits to develop human retinoblastoma xenografts. cell biology of retinoblastoma [38]. Included Using daily administration of intramuscular among the heterotopic xenografts in the literature cyclosporine, a xenograft model of retinoblasare athymic nude mice receiving Y79 cells [39] toma was developed by Kang and Grossniklaus and retinoblastoma cells injected into the anterior in New Zealand white rabbits, by subretinal injecchamber of SCID mice [40]. Y79 cells have also tion of WERI-RB1 cells [44]. This model was been injected into the vitreous of adult transgenic subsequently modified by Daniels et al. to develop Rag2 knockout immunocompromised mice, intraretinal tumors, subretinal seeds, and vitreous resulting in tumor invasion of the retina, brain, seeds, thus recapitulating many of the features of and subarachnoid space [41]. Heterotopic trans- advanced intraocular retinoblastoma tumors seen plantation of patient tumor tissue into the inter- in patients. This rabbit model has allowed sevscapular fat pad of Swiss nude mice has been eral techniques to be studied that would not have used for testing of bevacizumab and carboplatin. been possible in smaller rodent models, which The results were promising when combining the relied on genetically immunocompromised systwo agents, even with low-dose carboplatin, in tems. For example, Daniels et  al. were able to three different validated patient cell lines of reti- develop a rabbit model of intra-­arterial chemonoblastoma [42]. therapy, taking advantage of the relatively larger vascular caliber of rabbits (­compared to mice). Pharmacological Exogenous Combining this technique with their xenograft Immunosuppression to Facilitate model, they demonstrated that intra-arterial melXeno-engraftment phalan could induce widespread apoptosis of For certain types of experiments, larger animals human WERI-Rb1 cell lines xenografted intravitare often needed. For example, studies of intra-­ really to create vitreous seeds. This represented arterial chemotherapy, plaque brachytherapy, or the first ever assessment of the efficacy of intraphotodynamic therapy cannot be carried out eas- arterial chemotherapy in an animal model and ily on small mouse eyes. While there are excel- opened the way for intra-arterial drug discovery lent immunocompromised or immune-naïve research [31]. The possibility of treating retinoxenograft models of retinoblastoma available in blastoma xenografts with photodynamic therapy mice and rats, immunocompromised versions of was explored by Kim et al. [45]. In this case, piglarger species are not readily available. Thus, mented rabbits were necessary, but once again xeno-engraftment is made feasible in these larger cyclosporine was used as an immunosuppressant species by administration of a pharmacological to facilitate xeno-engraftment. immunosuppressant. The most common agent Disadvantages of a cyclosporine-immunosupused for this purpose is cyclosporine. pressed rabbit model include the cost and regu-

7  Animal Models in Retinoblastoma Research

87

latory considerations associated with the use of rabbits, as well as the toll (both on the animals and on laboratory personnel) of 7 days per week cyclosporine injections. As described above, immunosuppression may alter tumor biology and certainly affects interactions between the tumor and the host’s immune system, which may impact studies of inflammatory response to treatment. In addition, there is evidence that cyclosporine may itself have an antineoplastic effect on retinoblastoma. Cyclosporine has been used in some intravenous chemotherapeutic regimens for retinoblastoma [46]. Therefore, it remains unclear how cyclosporine use for immunosuppression may impact interpretation of chemotherapy efficacy.

 reclinical Animal Models to Assess P Pharmacokinetics and Toxicity Animal models assist in preclinical pharmacokinetic, drug distribution, and safety studies. Assessment of ocular toxicity of pharmacologic agents is no different in a fundamental sense for retinoblastoma research compared to other diseases, except that many more routes of drug delivery are routinely employed for the treatment

of retinoblastoma. Intravenous, periocular, intravitreal, and intra-arterial routes of administration have all been studied in animal models. The specific safety parameters measured and tests that are employed are often dictated by the delivery route and agent being studied. The various routes of chemotherapy delivery used in the treatment of retinoblastoma are covered at length in other chapters throughout this book, but a brief consideration of their differences, particularly as they impact preclinical assessment of toxicity in animal models, is provided as a reference in Table 7.2. In general, pharmacokinetic studies and assessments of drug safety do not require tumor to be present in the model; thus the use of immunocompromised or immunosuppressed animals is not a limiting factor. Instead, researchers seek to maximize the similarity between the ocular anatomy of humans and the animal species being used. The particular similarities that are most prized, and which are therefore most heavily weighted in selecting a particular animal species for study, are variable and highly dependent on the route of administration or the therapy being studied. For example, for intravitreal injection, the overall size of the vitreous cavity, as well as the size of the vitreous cavity relative to the lens,

Table 7.2  Comparison of animal models and their utility for various types of preclinical studies Type of model  Xenograft  Genetic Immune status  Xenograft    Immunocompromised    Immune naive    Immunosuppressed  Genetic Types of studies possible  Pharmacokinetics  Toxicity  Efficacy against tumor Routes  Intravenous  Periocular  Intravitreal  Intra-arterial

Mouse

Rat

Rabbit

Pig

NHP

+ +

+ −

+ −

− −

− −

+ − + Normal

− + + N/A

− − + N/A

N/A N/A N/A N/A

N/A N/A N/A N/A

+ + +

+ + +

+ + +

+ + −

+ + −

+ +/− +/− −

+ +/− +/− −

+ + + +

+ + + +

+ + + +

88

may be important [47]. For assessments of retinal function, the presence of a macula might be considered [48]. For periocular injections, the thickness of the sclera or the volume of the orbit might play a role in species selection [49]. For intra-­arterial chemotherapy delivery, the size of the animal vasculature relative to human babies must be considered, and smaller vasculature is more technically challenging (and ultimately excludes the use of very small animals) [31]. The nature of the retinal vasculature must also be considered, as certain experimental animals are merangiotic, with the retinal vessels lying in front of the internal limiting membrane (ILM), compared to the sub-ILM location of the retinal vessels in humans. Lastly, certain therapies such as laser require the use of a species or strain with a pigmented fundus [45].

Models of Intravitreal Chemotherapy The treatment of vitreous seeds has proven particularly challenging because the inherently avascular nature of the vitreous and the seeds does not allow high chemotherapy levels to be achieved. To combat this, chemotherapy can be injected directly into the vitreous to bathe the retinoblastoma vitreous seeds in high concentrations of drug. However, intravitreal injections of chemotherapy have been shown to be associated with retinopathy, particularly near the injection site where the concentration is initially highest prior to diffusion. In selecting an animal model to study the safety and pharmacokinetic profiles of intravitreal injections, the size of the animal’s eyes, and in particular of the vitreous relative to other ocular structures such as the lens, is a prime consideration. Mice are already widely used in the study of cell biology of retinoblastoma, are relatively inexpensive, easy to handle, and require approximately 1 month per generation in the laboratory for breeding purposes. Thus applying intravitreal chemotherapeutic to models already in use for the study of retinoblastoma pathophysiology is useful preclinically. Murine models of intravitreal injection have been used in retinoblastoma

T. A. Mendel and A. B. Daniels

research for applications ranging from pharmacokinetics of tyrosine kinase inhibitors [50] to pharmacodynamics of bevacizumab [51] and combination anti-VEGF/carboplatin [52]. The murine vitreous does not occupy significant space within the eye, making pars plana intravitreal injection technically challenging with little space between the lens and retina. Also, the injected agent often effluxes from the injection site, calling into question actual dose delivery. Thus, the small murine vitreous is not an excellent choice to recapitulate the human vitreous accurately. To study intravitreal application of therapy for the treatment of retinoblastoma, larger mammal systems such as rabbits have been employed, as they feature a higher proportion of vitreous as total volume in the eye when compared to mice. Therefore, rabbits better recapitulate the relative geometry of the human vitreous. In addition, since the injection is directly into the vitreous and bypasses the retinal vasculature completely, the merangiotic nature of the rabbit retina (described above) is not an important issue. New Zealand white rabbits have been widely used to study pharmacokinetics and toxicity of intravitreal injections, both for retinoblastoma and for other conditions [53]. Electroretinography (ERG) is often used as a readout of retinal function before and after injection of the drug of interest. Buitrago and colleagues utilized ERG to evaluate for retinal functional toxicity in rabbits during a pharmacokinetic study of intravitreal injection of topotecan (0.5–5.0  μg), finding no deleterious effects [54]. Francis and colleagues made use of ERG in a similar fashion, testing for retinal toxicity over time in rabbits receiving serial intravitreal injections of melphalan. In that study, they found that 30  μg serial injections of melphalan resulted in a decreased amplitude, suggestive of retinal toxicity [55]. Intravitreal digoxin was likewise tested in rabbits in order to ascertain not only vitreal peak concentrations over time but also the effect on ERG parameters as a measure of retinal function [56]. One disadvantage of New Zealand white rabbits is that they are an albino species.

7  Animal Models in Retinoblastoma Research

In humans, intravitreal melphalan (the most commonly used drug clinically) causes a pigmentary, “salt and pepper” retinopathy, which cannot be observed in an animal eye devoid of pigment. While a pigmented version of the New Zealand rabbit exists, it is not widely used or available. In these cases, the assumption is that the pigmentary changes are a consequence of the toxicity, and not causative, and ERG parameters and histopathologic findings, rather than the clinical evidence of pigmentary retinopathy, are measured as endpoints. To avoid this issue with pigmentation, Dutch-Belted rabbits have been used by investigators interested in testing the toxicity of combination intravitreal carboplatin and etoposide [47].

Models of Periocular Chemotherapy Periocular chemotherapy injections have been used in an effort to provide sustained high local concentrations of drug without subjecting the patient to high systemic levels. Different animal systems have been used to test periocular application of retinoblastoma chemotherapy, including mouse, rabbit, and pig [57]. One hurdle that provided significant challenge to investigators remains poor drug delivery to the intraocular tumor despite close geographic proximity. Animal studies in rabbits have demonstrated that the robust choroidal blood flow and scleral integrity carry drug away through the vasculature, without allowing significant penetration into the deeper retina [58]. The scleral shell, lying between the drug depot and the retina, demands the drug pass through the episclera, sclera, choroid, Bruch’s membrane, and RPE prior to gaining access to the retina. Thus, steep concentration gradients are found with the lowest level in the vitreous [59]. In fact, pharmacokinetic studies in rabbits suggest that drug injected periocularly ultimately gains access to the retina and vitreous through venous return from the orbit and systemic recirculation, rather than through direct scleral transmigration [49]. This reduces the model to a depot system for subsequent intravenous delivery, replete with the same drawbacks

89

of first pass metabolism and high systemic toxicity. Additionally, periocular administration risks local toxicity. Carboplatin in particular has been studied with periocular administration with little success [60]. To assess local toxicity, New Zealand white rabbits were subjected to subconjunctival injections of carboplatin with eyes analyzed at different time points over the following 24  hours [58]. The authors selected rabbits following prior human trials with the goal of recapitulating human anatomy and pharmacokinetics. Ultimately, intra-arterial chemotherapy and direct intravitreal injection have now replaced periocular injections as a means to achieve high intraocular drug concentrations. These are discussed below.

Models of Intravenous Chemotherapy The study of the pharmacokinetics and toxicity of intravenous chemotherapies for the treatment of retinoblastoma is similar to other systemic chemotherapies for other diseases. Because there are no limitations related to the technique for drug infusion, the size of the eye or vasculature is less critical than for intravitreal or intra-arterial chemotherapy models, respectively. In addition, since pharmacokinetic studies and toxicity studies do not seek to assess directly the effect of chemotherapy on tumors, investigators are not limited to particular animal models based on the nature or source of the tumors they develop. In fact, most pharmacokinetic and toxicity studies performed in animals for retinoblastoma research have been performed in non-tumor-bearing animals. Studies of pharmacokinetics and toxicity of chemotherapeutics for retinoblastoma often include an experimental arm for intravenous administration. While intravenous administration poses more systemic toxicity risk to the organism, its ease of application merits investigation, often probing the role of the blood-retinal barrier. For example, nutlin-3a has been explored as a potential therapeutic for retino-

90

blastoma, because it inhibits MDM2 and therefore restores p53 function, which is usually downregulated in retinoblastoma. Zhang and colleagues tested intravenous administration of nutlin-3a alongside oral delivery [61]. They found significantly higher sustained tissue concentrations when nutlin-3a was administered orally compared to intravenously. In another study, Kim and colleagues tested porosity of the blood-retinal barrier with different sizes of gold nanoparticles, finding that 20 nm particles could deposit in the various layers of the retina when administered intravenously, whereas 100  nm particles could not [62]. This finding may prove useful for further intravenous applications to assess ability of drug to exit the bloodstream through an intact blood-retinal barrier. In another study of that aimed to assess the role of the blood-retinal barrier following intravenous delivery, Pascual-Pasto and colleagues administered intravenous topotecan to New Zealand white rabbits and athymic nude mice [63]. They found that co-administration of pantoprazole, theorized to inhibit efflux transporters that contribute to blood-retinal barrier integrity, leads to higher concentrations of chemotherapeutic drug in the retina. While not always specifically testing efficacy on tumors, these studies still serve an important role for our understanding of how intravenous administration can deliver drug effectively to intraocular tumors. There are only a few considerations unique to retinoblastoma research for studies of intravenous agents. For pharmacokinetic studies, one must recognize that intravenous chemotherapies tend to have low ocular penetration. Thus, the intraocular concentrations of the study drug (within the retina and/or the vitreous) must be directly measured, unlike with other cancers that do not arise in the eye where blood levels alone are often used. For toxicity studies related to intravenous chemotherapy, one must consider how the chemotherapyinduced death of an intraocular tumor might contribute to overall ocular toxicity. While this may be different from the ocular toxicity seen in non-tumor-bearing animal eyes, in practice,

T. A. Mendel and A. B. Daniels

most assessments of ocular toxicity in retinoblastoma animal models are performed in nontumor-bearing animals.

Models of Intra-arterial Chemotherapy Endovascular microcatheter-based delivery of chemotherapy into the ophthalmic artery (intra-­ arterial chemotherapy, or IAC) is technically challenging to learn to perform even in the relatively large vasculature of humans. In addition, the size of the human (or animal) eye relative to the overall size of the animal is widely variable, both between species and at different ages. In selecting an animal model to study intra-arterial drug delivery, the vasculature is therefore of paramount importance. Very small laboratory rodents, such as mice or rats, would be impossible to catheterize, as their lumens are much smaller than even the smallest commercially available microcatheters.

Porcine Models The first animal model used to study pharmacokinetics of intra-arterial chemotherapy was the pig. The pig eye is comparable in vascular supply and anatomy to that of humans. Landrace pigs were chosen as a standard porcine model, having been used previously for similar pharmacokinetic studies. In addition, the large size of a 70 kg pig made the technique easier to perform. Researchers measured intravitreal concentrations of both melphalan and topotecan following infusion of each drug into the ophthalmic artery [57, 64]. The concentration of melphalan was noted to result in a threefold vitreous/ plasma concentration ratio. However, the great costs of obtaining and housing large pigs ultimately limits the ability to perform more involved or longer-term experiments. In addition, there is currently no porcine model of retinoblastoma, so experiments can only be limited to the study of pharmacokinetics or short-­term toxicity. Experiments related to the in vivo efficacy of chemotherapeutic agents cannot be pursued in pigs.

7  Animal Models in Retinoblastoma Research

 onhuman Primate (NHP) Models N The rhesus macaque was chosen as a model to study the toxicity of intra-arterial melphalan, having been validated as an anticancer drug pharmacokinetic system previously. The weight and size of an adult rhesus macaque are similar to that of a 2-year-old human child. While testing toxicity melphalan and carboplatin, investigators also capitalized on the larger size of the eye – when compared to other model systems  – to validate in  vivo imaging and testing with photography and fluorescein angiography [65]. However, vascular complications were more severe and more frequent in the NHP model than is seen clinically in patients with retinoblastoma. For example, a high rate of vascular thrombosis and embolus formation were observed. Part of this may have been technical, related to direct catheterization of the ophthalmic artery and perhaps to the failure to anticoagulate the macaques periprocedurally with heparin (as is routine in clinical practice). Additional considerations are that adult macaques were used to simulate the vasculature of human babies. While the size of the vessels might be similar, adult macaques have more aged vessels, including the presence of atherosclerosis. The presence of atherosclerosis increases the risks and morbidity associated with percutaneous intravascular procedures. The high rate of embolism seen in this model might be related to this fact, and the aged vessels of NHPs might therefore not predict the risk to young children. Rabbit Models Daniels et al. [31] described the first small animal model of intra-arterial chemotherapy. Despite the small vasculature, the procedure could be performed easily and reproducibly in rabbits. The overall size and weight of the rabbit better correlates to the size of a human baby, as compared to 70  kg Landrace pigs. In addition, despite the smaller doses infused in the rabbit model, concentrations in retina and vitreous were high. In fact, they were much higher than in the pig model, despite the fact that a much larger overall dose of drug was injected intra-arterially into the large pigs. In addition, the severe vascu-

91

lar toxicity and occlusions that were seen in the aged rhesus macaque model were not seen with the rabbit model. Using a battery of structural and functional assessments of the retina and retinal vasculature, Daniels et  al. have shown that dose-dependent retinal and retinal vascular toxicity can be recapitulated and measured in their rabbit model. The toxicities seen with both melphalan and carboplatin can likewise be recapitulated. Importantly, the toxicity seen in the rabbit model is related to the specific drug used and is dose-dependent, whereas the intra-arterial procedure and technique they use does not itself cause ocular toxicity or vascular damage. Further, of all the animal models of intra-arterial chemotherapy, the rabbit model is the only one in which assessments of efficacy can be performed, as rabbits are the only species in which a retinoblastoma tumor xenograft model exists [31, 44] (see below section on Testing Therapeutic Efficacy).

 nimal Models to Test Therapeutic A Efficacy Against Retinoblastoma Ultimately, the goal of drug and therapeutic discovery for retinoblastoma lies beyond pharmacokinetics and toxicity studies. With the multitude of available retinoblastoma animal models, the question is how best to use these models to assess the efficacy of novel therapeutics in  vivo. The selection of a particular animal model is guided by the particular therapy being studied and is often highly dependent on the route of administration. Species, tumor model, specific cell or tumor line, and route of administration are all factors to be considered. The various animal models, and the ways in which they are being used in therapeutic discovery, are discussed below.

Intravenous Models Considerations Intravenous chemotherapy was the mainstay of retinoblastoma care beginning in the mid-1990s

92

and remains the primary mode of therapy used in many retinoblastoma treatment centers around the world. Because intravenous chemotherapy mimics other routes of systemic administration, studies of drugs intended for systemic or intravenous use are occasionally administered to the animals via other routes besides into the vein, including oral and intraperitoneal injection. Since drugs administered systemically reach all tissues, the specific anatomy of the eye, orbit, and ocular vasculature are less critical, compared to regional routes of administration discussed below. Thus, most systemic drugs are tested for efficacy in mouse models, because mice have the widest variety of options for retinoblastoma models available and are relatively inexpensive to purchase, breed, and house and because most labs are familiar with their use. The only limitation is the technical difficulty of orthotopic injection of tumor cells into the eye, which is a difficulty with all mouse orthotopic xenograft models, regardless of the drug being studied. Heterotopic xenografts avoid these technical issues and often are sufficient to assess the effect of a drug on human-­ derived retinoblastoma cells. Therefore, many experiments with drug efficacy often begin with the study of heterotopic grafts. However, a limiting factor of many intravenous chemotherapies is the drug level that can be achieved in the eye. Conversely, if a small fraction of the systemic concentration reaches the eye, then in order to achieve adequate drug levels in the eyes, a very high and toxic dose may need to be given intravenously. While early pharmacokinetic studies can provide guidance, pharmacokinetic experiments in non-tumor-bearing eyes may be insufficient as the presence of tumor in the retina itself may impact the amount of drug that can penetrate into the eye. Thus, orthotopic tumor models are ultimately needed to confirm not just that the drug works but that it works at the levels that can be attained in the eye. There are other considerations for systemic medications or biologics that are specific to the class of drug itself. An additional issue arises with systemic drugs or biologics whose mechanism is immune modulation, such as immune checkpoint inhibitors. As mentioned above,

T. A. Mendel and A. B. Daniels

xenograft models require suppression or compromise of the immune system to facilitate engraftment of human tissues, making them poor models to study the effect of the immune system or drugs that alter it. In addition, a lot of these biologics may be specific to the human immune system or to human tumor antigens (e.g., IMCgp100, which recruits T cells directly to human melanoma cells). Thus far, however, immunomodulatory agents or immune checkpoint inhibitors have not played nearly as major a role in the care of retinoblastoma patients as they have for patients with other tumors, such as melanoma. In contrast, when considering murine genetic models of retinoblastoma, one needs to be cognizant of the fact that protein tyrosine kinase inhibitors, or other molecularly targeted inhibitors of tumor pathways, may not necessarily bind with as high an affinity (or at all) to the mouse version of the protein. In such circumstances, human xenograft tumors may be preferable. In addition, because of intra-tumoral heterogeneity seen clinically within patient tumors and also variation in tumor susceptibility to drug seen between patients, patient-derived xenografts may provide better “real-world” guidance on drug efficacy, particularly if several xenografts derived from the tumors of different patients are each studied. In the study by Pascual-Pasto et al. of pantoprazole to inhibit efflux transporters in the ­blood-­retinal barrier (see intravenous pharmacokinetics and toxicity section above), investigators also measured the response of intraocular PDX xenografts to intravenous topotecan in athymic nude mice [63]. They found that intraocular topotecan levels increase with co-administration of pantoprazole. They also show that the growth of orthotopic xenografts can be inhibited in  vivo, quantifying efficacy as time until tumor invasion into the anterior chamber or proptosis. They compared efficacy of the various regimens, demonstrating that mouse eyes survived longer with combination therapy compared to topotecan alone or no treatment. While clinical response to intravenous therapy is often assessed in human patients, preclinical animal models can serve a vital role in suggesting not only therapeutic ocular concentrations but also tumor response.

7  Animal Models in Retinoblastoma Research

Combination Periocular/Systemic Therapy

93

xenograft models with vitreous seed formation have been described, as discussed above [31, 44]. While, other larger animals may mimic the Multiple different modalities and routes of che- human eye even better than rabbits, there are no motherapy administration can be tested in combi- animal models of retinoblastoma in any animals nation in animal models to assess for synergy of larger than the rabbit. efficacy against retinoblastoma tumors as well as Investigators have successfully tested the efficumulative ocular and systemic toxicity. Nemeth cacy of intravitreal chemotherapy against orthoand colleagues tested the impact of combined topically grafted PDX retinoblastoma cells in periocular and systemic carboplatin and topote- adult NOD/SCID immunocompromised mice can in an orthotopic xenograft model [10]. In that [67]. Cassoux and colleagues generated their study, they determined that subconjunctival car- orthotopic xenografts as previously demonstrated boplatin in combination with systemic (intraperi- by Aerts et al. [68], first growing PDX in a subtoneal) topotecan was more effective and less scapular location and then harvesting the tumors, toxic than subconjunctival topotecan and intra- creating a cell suspension, and injecting a 2 μL peritoneal carboplatin in orthotopic xenografts suspension of 20,000 PDX cells subretinally with generated by injecting Y79 cells into 2-week-old 33 gauge needles. Tumors were then observed to immunologically naïve rats. Specifically, subcon- grow over 4–6  weeks in all transplanted eyes. junctival deposition resulted in higher vitreous-­ Subsequent weekly intravitreal injections of melto-­plasma concentrations of both drugs. Tumor phalan or carboplatin were reportedly able to burden was reduced most markedly in the sub- clear 33% and 83% of the tumor bulk. In this conjunctival carboplatin and intraperitoneal way, researchers were able to test the efficacy of topotecan group. The rats that received subcon- intravitreal chemotherapy in a model that simujunctival topotecan and intraperitoneal carbopla- lates human intraretinal retinoblastoma tumors. tin experienced significant morbidity, all dying One limitation is that intravitreal chemotherapy within 5 days of drug administration. Additional is usually used to treat residual seeds following considerations about the role of venous recircula- systemic or intra-arterial chemoreduction of the tion following periocular/subconjunctival admin- retinal tumors, whereas in this case, the intravitistration have been discussed above. real melphalan or carboplatin were being used as primary therapy for the retinal tumors themselves. Intravitreal Models While angiogenesis has become a critical target for many other eye diseases, this approach is Considerations not currently in use for the treatment of retinoDifferent mouse models have been used to study blastoma. Angiogenesis is an attractive target for the in  vivo efficacy of various intravitreally retinoblastoma, as retinoblastoma tumor cells are injected agents on retinoblastoma tumors. These extremely oxygen sensitive. Tumors generate include genetic models of retinoblastoma such as their own vascular supply, and tumor cells that the LHβTAG mouse model, to xenograft models are not near tumor vessels undergo necrosis. such as nude mice into which Y79 human retino- Histologically, one sees sheathes of live tumor blastoma cell lines have been injected intravit- cells surrounding intra-tumor vessels, with interreally to directly recapitulate vitreous seeds [66]. vening areas of necrosis away from the vessels. However, because the architecture and size of the Bajenaru and colleagues injected anecortave acemouse vitreous cavity relative to other ocular tate, a steroid with anti-angiogenic properties structures (such as the lens) are unlike that of under study for macular generation, into the vithumans, rabbit models may represent a better reous of LHβTAG mice. They investigated the posalternative in which to assess the efficacy of sible role of inhibition of matrix metalloproteinases intravitreally injected chemotherapeutics. Rabbit 2 and 9, impacting the angiogenesis of retino-

94

blastoma, by injected anecortave acetate subconjunctivally and then measuring gelatinase activity at 1, 2, and 7 days thereafter [69]. The investigators found that gelatinases are inhibited by anecortave acetate, suggesting that it may slow the angiogenic component of retinoblastoma, but did not study tumor response specifically.

Intra-arterial Model Considerations Intra-arterial chemotherapy has become an increasingly common modality for treatment of children with retinoblastoma. Several of the drugs currently in use were first tried through off-­ label use in children with retinoblastoma, and toxicities and efficacy were described based on clinical experience. New animal models of intra-­ arterial chemotherapy now allow investigators to assess the toxicity and efficacy of potential antineoplastic agents in vivo in animals prior to their use in babies. Thus, in the future, drug discovery will increasingly rely on animal models of intra-­ arterial chemotherapy. Compared to intravenous, intravitreal, and periocular delivery, intra-arterial chemotherapy is the modality that is most limited by the relatively few species in which it has been performed, due to the need for relatively large vasculature through which to navigate the microcatheter. Animal models of intra-arterial chemotherapy only exist in pigs, in nonhuman primates, and in rabbits [31, 49]. While pharmacokinetics and toxicity can be studied in each model, to assess the efficacy of a new agent, the same species must also have a tumor model. Among these species, there are no genetic models of retinoblastoma. A xenograft model of retinoblastoma was recently described in rabbits [31, 44]. Thus, future preclinical drug discovery for intra-arterial chemotherapy will rely on the rabbit model described by Daniels et al., as only that intra-­arterial model allows the efficacy of chemotherapy agents to be tested against human retinoblastoma xenografts. Daniels et al. performed the first ever demonstration in an animal model of the in vivo efficacy of intra-arterial chemotherapy to treat intraocular

T. A. Mendel and A. B. Daniels

retinoblastoma, showing that intra-arterial melphalan could induce widespread apoptosis of vitreous seeds at doses that are commonly used clinically. They subsequently showed that neither the same doses of melphalan given intravenously nor standard intravenous carboplatin/etoposide/ vincristine could induce similar death of vitreous seeds in this model.

References 1. de Jager SM, Murray JA.  Retinoblastoma proteins in plants. Plant Mol Biol. 1999;41(3):295–9. Epub 1999/12/22. 2. Kobayashi S, Mukai N.  Retinoblastoma-like tumors induced in rats by human adenovirus. Investig Ophthalmol. 1973;12(11):853–6. Epub 1973/11/01. 3. Mukai N, Kalter SS, Cummins LB, et al. Retinal tumor induced in the baboon by human adenovirus 12. Science. 1980;210(4473):1023–5. Epub 1980/11/28. 4. Hu N, Gutsmann A, Herbert DC, et al. Heterozygous Rb-1 delta 20/+mice are predisposed to tumors of the pituitary gland with a nearly complete penetrance. Oncogene. 1994;9(4):1021–7. Epub 1994/04/01. 5. Williams BO, Schmitt EM, Remington L, et al. Extensive contribution of Rb-deficient cells to adult chimeric mice with limited histopathological consequences. EMBO J. 1994;13(18):4251–9. Epub 1994/09/15. PubMed PMID: 7925270; PMCID: PMC395352. 6. Zhang J, Schweers B, Dyer MA.  The first knockout mouse model of retinoblastoma. Cell cycle (Georgetown, Tex). 2004;3(7):952–9. Epub 2004/06/11. 7. Jacks T, Fazeli A, Schmitt EM, et al. Effects of an Rb mutation in the mouse. Nature. 1992;359(6393):295–300. Epub 1992/09/24. https:// doi.org/10.1038/359295a0. 8. Mendoza PR, Grossniklaus HE. The biology of retinoblastoma. Prog Mol Biol Transl Sci. 2015;134:503– 16. Epub 2015/08/28. https://doi.org/10.1016/ bs.pmbts.2015.06.012. 9. Lambertz I, Nittner D, Mestdagh P, et al. Monoallelic but not biallelic loss of Dicer1 promotes tumorigenesis in  vivo. Cell death and differentiation. 2010;17(4):633–641. Epub 2009/12/19. https:// doi.org/10.1038/cdd.2009.202. PubMed PMID: 20019750; PMCID: PMC2892162. 10. Nemeth KM, Federico S, Carcaboso AM, et al. Subconjunctival carboplatin and systemic topotecan treatment in preclinical models of retinoblastoma. Cancer. 2011;117(2):421–434. Epub 2010/09/08. https://doi.org/10.1002/cncr.25574. PubMed PMID: 20818652; PMCID: PMC3000447. 11. Benavente CA, Finkelstein D, Johnson DA, et al. Chromatin remodelers HELLS and UHRF1 mediate the epigenetic deregulation of genes that drive

7  Animal Models in Retinoblastoma Research retinoblastoma tumor progression. Oncotarget. 2014;5(20):9594–9608. Epub 2014/10/23. https:// doi.org/10.18632/oncotarget.2468. PubMed PMID: 25338120; PMCID: PMC4259422. 12. Kivela T, Virtanen I, Marcus DM, et al. Neuronal and glial properties of a murine transgenic retinoblastoma model. Am J Pathol. 1991;138(5):1135–48. Epub 1991/05/01. PubMed PMID: 1708946; PMCID: PMC1886007. 13. O’Brien JM, Marcus DM, Niffenegger AS, et al. Trilateral retinoblastoma in transgenic mice. Trans Am Ophthalmol Soc. 1989;87:301–22; discussion 22-6. Epub 1989/01/01. PubMed PMID: 2576479; PMCID: PMC1298548. 14. O’Brien JM, Marcus DM, Bernards R, et al. A transgenic mouse model for trilateral retinoblastoma. Arch Ophthalmol. 1990;108(8):1145–51. Epub 1990/08/01. PubMed PMID: 1696469. 15. Bogenmann E, Lochrie MA, Simon MI.  Cone cell-­ specific genes expressed in retinoblastoma. Science. 1988;240(4848):76–8. Epub 1988/04/01. 16. Vrabec T, Arbizo V, Adamus G, et al.  Rod cell-specific antigens in retinoblastoma. Arch Ophthalmol. 1989;107(7):1061–3. Epub 1989/07/01. 17. Pajovic S, Corson TW, Spencer C, et al. The TAg-RB murine retinoblastoma cell of origin has immunohistochemical features of differentiated Muller glia with progenitor properties. Invest Ophthalmol Vis Sci. 2011;52(10):7618–7624. Epub 2011/08/25. https:// doi.org/10.1167/iovs.11-7989. PubMed PMID: 21862643; PMCID: PMC3183982. 18. Xu XL, Fang Y, Lee TC, et al. Retinoblastoma has properties of a cone precursor tumor and depends upon cone-­specific MDM2 signaling. Cell. 2009;137(6):1018–1031. Epub 2009/06/16. https:// doi.org/10.1016/j.cell.2009.03.051. PubMed PMID: 19524506; PMCID: PMC5659855. 19. Marcus DM, Lasudry JG, Carpenter JL, et al. Trilateral tumors in four different lines of transgenic mice expressing SV40 T-antigen. Invest Ophthalmol Vis Sci. 1996;37(2):392–6. Epub 1996/02/01. 20. Al-Ubaidi MR, Font RL, Quiambao AB, et al. Bilateral retinal and brain tumors in transgenic mice expressing simian virus 40 large T antigen under control of the human interphotoreceptor retinoid-binding protein promoter. J Cell Biol. 1992;119(6):1681–7. Epub 1992/12/01. PubMed PMID: 1334963; PMCID: PMC2289740. 21. Howes KA, Lasudry JG, Albert DM, et al. Photoreceptor cell tumors in transgenic mice. Invest Ophthalmol Vis Sci. 1994;35(2):342–51. Epub 1994/02/01. PubMed PMID: 8112979. 22. Al-Ubaidi MR, Hollyfield JG, Overbeek PA, et al. Photoreceptor degeneration induced by the expression of simian virus 40 large tumor antigen in the retina of transgenic mice. Proc Natl Acad Sci U S A. 1992;89(4):1194–8. Epub 1992/02/15. PubMed PMID: 1311085; PMCID: PMC48415. 23. Sobrin L, Hayden BC, Murray TG, et al. External beam radiation “salvage” therapy in transgenic murine reti-

95 noblastoma. Arch Ophthalmol. 2004;122(2):251–7. Epub 2004/02/11. https://doi.org/10.1001/archopht. 122.2.251. 24. Murray TG, O’Brien JM, Steeves RA, et al. Radiation therapy and ferromagnetic hyperthermia in the treatment of murine transgenic retinoblastoma. Arch Ophthalmol. 1996;114(11):1376–81. Epub 1996/11/01. 25. Harbour JW, Murray TG, Hamasaki D, et al.  Local carboplatin therapy in transgenic murine retinoblastoma. Invest Ophthalmol Vis Sci. 1996;37(9):1892–8. Epub 1996/08/01. 26. Hayden BH, Murray TG, Scott IU, et al. Subconjunctival carboplatin in retinoblastoma: impact of tumor burden and dose schedule. Arch Ophthalmol. 2000;118(11):1549–54. Epub 2000/11/14. 27. Murray TG, Cicciarelli N, O’Brien JM, et al. Subconjunctival carboplatin therapy and cryotherapy in the treatment of transgenic murine retinoblastoma. Arch Ophthalmol. 1997;115(10):1286–90. Epub 1997/10/24. 28. Xu XL, Singh HP, Wang L, et al. Rb suppresses human cone-precursor-derived retinoblastoma tumours. Nature. 2014;514(7522):385–388. Epub 2014/09/26. https://doi.org/10.1038/nature13813. PubMed PMID: 25252974; PMCID: PMC4232224. 29. Bremner R, Sage J. The origin of human retinoblastoma. Nature. 2014;514:312. https://doi.org/10.1038/ nature13748. 30. Gallie BL, Campbell C, Devlin H, et al. Developmental basis of retinal-specific induction of cancer by RB mutation. Cancer Res. 1999;59(7 Suppl):1731s–5s. Epub 1999/04/10. 31. Daniels AB, Froehler MT, Pierce JM, et al. Pharmacokinetics, tissue localization, toxicity, and treatment efficacy in the first small animal (Rabbit) model of intra-arterial chemotherapy for retinoblastoma. Invest Ophthalmol Vis Sci. 2018;59(1):446– 454. Epub 2018/01/26. https://doi.org/10.1167/ iovs.17-22302. PubMed PMID: 29368001; PMCID: PMC5783625. 32. Kang SJ, Grossniklaus HE.  Rabbit model of retinoblastoma. J Biomed Biotechnol. 2011;2011:5. https:// doi.org/10.1155/2011/394730. 33. Brodowska K, Theodoropoulou S, Meyer Zu Horste M, et al.  Effects of metformin on retinoblastoma growth in  vitro and in  vivo. Int J Oncol. 2014;45(6):2311–2324. Epub 2014/09/13. https:// doi.org/10.3892/ijo.2014.2650. PubMed PMID: 25215935; PMCID: PMC4215581. 34. Zhang B, Li Y, Zhong X, et al. Establishment of retinoblastoma model in NOD-­ SCID mice and study of metastasis. Yan Ke Xue Bao = Eye science. 2005;21(3):185–91. Epub 2006/12/14. 35. Rolstad B.  The athymic nude rat: an animal experimental model to reveal novel aspects of innate immune responses? Immunol Rev. 2001;184:136–44. Epub 2002/03/29. 36. Li Z, Wu X, Li J, et al. Antitumor activity of celastrol nanoparticles in a xenograft retinoblastoma tumor

96 model. Int J Nanomedicine. 2012;7:2389–2398. Epub 2012/06/05. https://doi.org/10.2147/ijn.s29945. PubMed PMID: 22661892; PMCID: PMC3357982. 37. Song X, Zhou Y, Jia R, et al. Inhibition of retinoblastoma in vitro and in vivo with conditionally replicating oncolytic adenovirus H101. Invest Ophthalmol Vis Sci. 2010;51(5):2626–35. Epub 2009/12/17. https://doi.org/10.1167/iovs.09-3516. 38. Ruiz S, Segrelles C, Bravo A, et al.  Abnormal epidermal differentiation and impaired epithelial-­ mesenchymal tissue interactions in mice lacking the retinoblastoma relatives p107 and p130. Development. 2003;130(11):2341–53. Epub 2003/04/19. 39. Sabet SJ, Darjatmoko SR, Lindstrom MJ, et al. Antineoplastic effect and toxicity of 1,25-­dihydroxy-­ 16-ene-23-yne-vitamin D3 in athymic mice with Y-79 human retinoblastoma tumors. Arch Ophthalmol. 1999;117(3):365–70. Epub 1999/03/24. 40. Madreperla SA, Whittum-Hudson JA, Prendergast RA, et al. Intraocular tumor suppression of retinoblastoma gene-reconstituted retinoblastoma cells. Cancer Res. 1991;51(23 Pt 1):6381–4. Epub 1991/12/01. 41. Chevez-Barrios P, Hurwitz MY, Louie K, et al. Metastatic and nonmetastatic models of retinoblastoma. Am J Pathol. 2000;157(4):1405–1412. Epub 2000/10/06. https://doi.org/10.1016/s00029440(10)64653-6. PubMed PMID: 11021842; PMCID: PMC1850157. 42. Assayag F, Nicolas A, Vacher S, et al.  Combination of carboplatin and bevacizumab is an efficient therapeutic approach in retinoblastoma patientderived xenografts. Invest Ophthalmol Vis Sci. 2016;57(11):4916–4926. Epub 2016/09/23. https:// doi.org/10.1167/iovs.15-18725. PubMed PMID: 27654418. 43. del Cerro M, Seigel GM, Lazar E, et al. Transplantation of Y79 cells into rat eyes: an in  vivo model of human retinoblastomas. Invest Ophthalmol Vis Sci. 1993;34(12):3336–46. Epub 1993/11/01. PubMed PMID: 8225869. 44. Kang SJ, Grossniklaus HE. Rabbit model of retinoblastoma. J Biomed Biotechnol. 2011;2011:394730. Epub 2011/01/22. https://doi.org/10.1155/2011/394730. PubMed PMID: 21253494; PMCID: PMC3022222. 45. Kim JW, Jacobsen B, Zolfaghari E, et al.  Rabbit model of ocular indirect photodynamic therapy using a retinoblastoma xenograft. Graefes Arch Clin Exp Ophthalmol. 2017;255(12):2363–73. Epub 2017/10/04. https://doi.org/10.1007/s00417-017-3805-8. 46. Chan HS, DeBoer G, Thiessen JJ, et al.  Combining cyclosporin with chemotherapy controls intraocular retinoblastoma without ­requiring radiation. Clin Cancer Res. 1996;2(9):1499–508. Epub 1996/09/01. 47. Mohney BG, Elner VM, Smith AB, et al. Preclinical acute ocular safety study of combined intravitreal carboplatin and etoposide phosphate for retinoblastoma. Ophthalmic Surg Lasers Imaging Retina 2017;48(2):151–159. Epub 2017/02/15. https:// doi.org/10.3928/23258160-­20170130-09. PubMed PMID: 28195618.

T. A. Mendel and A. B. Daniels 48. Houston SK, Lampidis TJ, Murray TG.  Models and discovery strategies for new therapies of retinoblastoma. Expert Opin Drug Discovery. 2013;8(4):383– 394. Epub 2013/02/23. https://doi.org/10.1517/17460 441.2013.772975. PubMed PMID: 23427911. 49. Schaiquevich P, Fabius AW, Francis JH, et al. Ocular pharmacology of chemotherapy for retinoblastoma. Retina 2017;37(1):1–10. Epub 2016/09/13. https:// doi.org/10.1097/iae.0000000000001275. PubMed PMID: 27617542. 50. Pritchard EM, Stewart E, Zhu F, et al. Pharmacokinetics and efficacy of the spleen tyrosine kinase inhibitor r406 after ocular delivery for retinoblastoma. Pharm Res. 2014;31(11):3060–3072. Epub 2014/06/08. https://doi.org/10.1007/s11095-014-1399-y. PubMed PMID: 24906597; PMCID: PMC4213378. 51. Kim JH, Kim C, Kim JH, et al. Absence of intravitreal bevacizumab-induced neuronal toxicity in the retina. Neurotoxicology. 2008;29(6):1131– 5. Epub 2008/07/22. https://doi.org/10.1016/j. neuro.2008.06.006. 52. Zhang Q, Cheng Y, Huang L, et al. Inhibitory effect of carboplatin in combination with bevacizumab on human retinoblastoma in an in  vitro and in  vivo model. Oncol Lett. 2017;14(5):5326–5332. Epub 2017/11/04. https://doi.org/10.3892/ol.2017.6827. PubMed PMID: 29098028; PMCID: PMC5652222. 53. Buitrago E, Winter U, Williams G, et al. Pharmacokinetics of melphalan after intravitreal injection in a rabbit model. J Ocul Pharmacol Ther. 2016;32(4):230–5. Epub 2016/01/20. https://doi. org/10.1089/jop.2015.0088. 54. Buitrago E, Del Sole MJ, Torbidoni A, et al. Ocular and systemic toxicity of intravitreal topotecan in rabbits for potential treatment of retinoblastoma. Exp Eye Res. 2013;108:103–9. Epub 2013/01/22. https:// doi.org/10.1016/j.exer.2013.01.002. 55. Francis JH, Schaiquevich P, Buitrago E, et al. Local and systemic toxicity of intravitreal melphalan for vitreous seeding in retinoblastoma: a preclinical and clinical study. Ophthalmology. 2014;121(9):1810– 7. Epub 2014/05/14. https://doi.org/10.1016/j. ophtha.2014.03.028. 56. Winter U, Buitrago E, Mena HA, et al. Pharmacokinetics, safety, and efficacy of intravitreal digoxin in preclinical models for retinoblastoma. Invest Ophthalmol Vis Sci. 2015;56(8):4382–93. Epub 2015/07/16. https://doi.org/10.1167/iovs.14-16239. 57. Schaiquevich P, Buitrago E, Taich P, et al. Pharmacokinetic analysis of melphalan after superselective ophthalmic artery infusion in preclinical models and retinoblastoma patients. Invest Ophthalmol Vis Sci. 2012;53(7):4205–12. Epub 2012/05/26. https://doi.org/10.1167/iovs.12-9501. 58. Hayden BC, Jockovich ME, Murray TG, et al. Pharmacokinetics of systemic versus focal Carboplatin chemotherapy in the rabbit eye: possible implication in the treatment of retinoblastoma. Invest Ophthalmol Vis Sci. 2004;45(10):3644–9. Epub 2004/09/29. https://doi.org/10.1167/iovs.04-0228.

7  Animal Models in Retinoblastoma Research 59. Edelhauser HF, Rowe-Rendleman CL, Robinson MR, et al. Ophthalmic drug delivery systems for the treatment of retinal diseases: basic research to clinical applications. Invest Ophthalmol Vis Sci. 2010;51(11):5403–5420. Epub 2010/10/29. https:// doi.org/10.1167/iovs.10-5392. PubMed PMID: 20980702; PMCID: PMC3061492. 60. Marr BP, Dunkel IJ, Linker A, et al. Periocular carboplatin for retinoblastoma: long-­ term report (12 years) on efficacy and toxicity. Br J Ophthalmol 2012;96(6):881–883. Epub 2012/02/14. https://doi. org/10.1136/bjophthalmol-2011-300517. 61. Zhang F, Tagen M, Throm S, et al. Whole-body physiologically based pharmacokinetic model for nutlin-3a in mice after intravenous and oral administration. Drug Metab Dispos. 2011;39(1):15–21. Epub 2010/10/16. https://doi.org/10.1124/dmd.110.035915. PubMed PMID: 20947617; PMCID: PMC3014266. 62. Kim JH, Kim JH, Kim KW, et al. Intravenously administered gold nanoparticles pass through the blood-­retinal barrier depending on the particle size, and induce no retinal toxicity. Nanotechnology. 2009;20(50):505101. Epub 2009/11/20. https://doi. org/10.1088/0957-4484/20/50/505101. 63. Pascual-Pasto G, Olaciregui NG, Opezzo JAW, et al. Increased delivery of chemotherapy to the vitreous by inhibition of the blood-retinal barrier. J Control Release. 2017;264:34–44. Epub 2017/08/24. https:// doi.org/10.1016/j.jconrel.2017.08.018. 64. Schaiquevich P, Buitrago E, Ceciliano A, et al. Pharmacokinetic analysis of topotecan after superselective ophthalmic artery infusion and periocular administration in a porcine model. Retina. 2012;32(2):387–95. Epub 2011/09/01. https://doi. org/10.1097/IAE.0b013e31821e9f8a.

97 65. Wilson MW, Jackson JS, Phillips BX, et al. Real-time ophthalmoscopic findings of superselective intraophthalmic artery chemotherapy in a nonhuman primate model. Arch Ophthalmol. 2011;129(11):1458-1465. Epub 2011/11/16. https://doi.org/10.1001/archophthalmol.2011.330. PubMed PMID: 22084215; PMCID: PMC3527084. 66. Tschulakow AV, Schraermeyer U, Rodemann HP, et al. Establishment of a novel retinoblastoma (Rb) nude mouse model by intravitreal injection of human Rb Y79 cells  - comparison of in  vivo analysis versus histological follow up. Biol Open. 2016;5(11):1625– 1630. Epub 2016/10/04. https://doi.org/10.1242/ bio.019976. PubMed PMID: 27694105; PMCID: PMC5155534. 67. Cassoux N, Thuleau A, Assayag F, et al. Establishment of an Orthotopic Xenograft Mice Model of Retinoblastoma Suitable for Preclinical Testing. Ocul Oncol Pathol. 2015;1(3):200–6. Epub 2016/05/14. https://doi.org/10.1159/000370156. PubMed PMID: 27171982; PMCID: PMC 4847680. 68. Aerts I, Leuraud P, Blais J, et al. In vivo efficacy of photodynamic therapy in three new xenograft models of human retinoblastoma. Photodiagn Photodyn Ther. 2010;7(4):275–83. Epub 2010/11/30. https:// doi.org/10.1016/j.pdpdt.2010.09.003. 69. Bajenaru ML, Pina Y, Murray TG, et al. Gelatinase expression in retinoblastoma: modulation of LH(BETA)T(AG) retinal tumor development by anecortave acetate. Invest Ophthalmol Vis Sci. 2010;51(6):2860–2864. Epub 2010/01/29. https:// doi.org/10.1167/iovs.09-4500. PubMed PMID: 20107171; PMCID: PMC2891454.

8

Retinocytoma or Retinoma Randy C. Bowen, Christina Stathopoulos, Francis L. Munier, and Arun D. Singh

Introduction The term retinoma was first introduced in 1982 by Gallie et al. to define nonprogressive retinal lesions observed in patients known to carry the gene for retinoblastoma [1]. Those lesions were described as translucent, gray, elevated retinal masses frequently associated with calcifications and pigment epithelium alterations [1]. Later, histopathologic studies demonstrated that these tumors are composed of well-differentiated, benign-appearing mature retinal cells with characteristic absence of mitoses and necrosis [2]. Based on the nomenclature used to classify pineal body tumors (benign, pineocytoma, and malignant, pineoblastoma), an alternate term retinocytoma has been also used to describe these tumors. Other less frequently used terminologies include spontaneously regressed retinoblastoma, arrested retinoblastoma, and R. C. Bowen Department of Ophthalmology, University of Wisconsin, Madison, WI, USA C. Stathopoulos · F. L. Munier Department of Ophthalmology, Jules-Gonin Eye Hospital, University of Lausanne, Lausanne, Vaud, Switzerland e-mail: [email protected]; [email protected] A. D. Singh (*) Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA e-mail: [email protected] © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_8

retinoblastoma group 0 [3–5]. Within this chapter, we will use the term retinocytoma. The incidence of retinocytoma in the general population is not known but ranges from 1.8% to 10% among those diagnosed with or having a positive family history of retinoblastoma [1, 5– 8]. Most of the cases are diagnosed in the parents or other first-degree relatives of a child with retinoblastoma [1, 5, 6, 9]. Nonetheless retinocytoma and retinoblastoma can coexist in the same patient (with retinoblastoma in one eye and retinocytoma in the fellow eye) [1, 6] or in the same eye [10]. Because they carry lifelong risk for malignant transformation, retinocytoma should be regularly followed [6, 7, 11].

Etiology and Pathogenesis Retinocytoma is considered to be a benign manifestation of the RB1 gene mutation [1, 4, 12, 13]. Historically, Knudson’s two-hit hypothesis has been applied to explain the pathogenesis of retinocytoma and retinoblastoma [14]. The theory states that both alleles of the RB1 gene must be mutated to convert normal retinal cells into neoplastic retinoblastoma cells. However, as genetic testing improved, retinocytoma was then thought to be the result of low expressivity and low penetrance caused by less severe mutations such as missense mutations. These missense mutations only reduce RB protein function, causing u­ nilateral tumors, retinocytomas, or 99

100

carrier state (without manifestation), whereas severe mutations caused by nonsense or frameshift mutation of RB1 gene result in complete loss of RB protein and are consequently associated with multifocal and bilateral retinoblastomas [15]. However, most recent evidence suggests that genetic instability and aneuploidy are instead the likely decisive factors separating retinoblastoma from its precursor retinocytoma tumor and that RB1 mutations causing retinocytoma are undistinguishable from those associated with retinoblastoma [16–18]. While both tumors may be homozygous null for RB1 (Rb -/-), retinocytoma display lower levels of aneuploidy and higher levels of senescence proteins [16, 19]. Cytoplasmic p16INK4a appears to be a uniquely expressed senescence protein in retinocytoma after loss of Rb -/-, which is not seen in normal retina or retinoblastoma. Cytoplasmic p16INK4a interacts with RB family proteins p130 and/or p107 to provide this cellular senescence [16]. When these senescence pathways fail and increasing genetic instability reaches a threshold through altered gene copies of oncogenes (MDM4 and KIF14 on chromosome 1q, MYCN on chromosome 2p, E2F3 and DEK on chromosome 6p) and tumor suppression genes (p75NTR on chromosome 11, CDH11 on 16q), among others, tumor cells become fully proliferative, resulting in retinoblastoma [16]. Loss of RB1, while necessary, is not sufficient to induce retinoblastoma. A senescence response to the mutation can result instead in nonproliferative retinocytoma. In theory then, all retinoblastomas progress through a stage of retinocytoma, however brief, before accrued genetic instability leads to uncontrolled proliferation (retinoblastoma).

R. C. Bowen et al.

because of minimal or complete lack of response to systemic chemotherapy or radiotherapy. In that case, the lesions fail also to grow when all treatments have been discontinued. Finally, retinocytoma reactivated in retinoblastoma can be suspected in cases of retinoblastoma presenting at an older age [11, 20].

Characteristics The ophthalmoscopic appearance of the retinocytoma resembles the spectrum of retinoblastoma regression patterns observed after irradiation (Box 8.1) [21]. Presence of a translucent-gray retinal mass (88%), calcification (63%), retinal pigment epithelial alteration (54%), and chorioretinal atrophy (54%) are four diagnostic ophthalmoscopic features of retinocytoma (Fig. 8.1) [1, 6, 7]. Any one of the four features listed above is present in all cases. However, majority (80%) of cases have at least two of the four features with only 10% of cases having all four features [7]. Other clinical manifestations of retinocytoma include seeding of calcifications into the vitreous [10, 22] and well-defined translucent cavities, a feature of presumed well-differentiated retinoblastoma [23, 24] (Fig.  8.2). Rarely, cases of phthisis bulbi have been reported in series of spontaneously regressed retinoblastoma [9, 12].

Clinical Features and Diagnosis The majority of patients with retinocytoma are asymptomatic, and the diagnosis occurs either on routine eye examination or when the diagnosis of retinoblastoma is made in another family member prompting an eye examination [1, 7]. Leukocoria, a common initial feature of retinoblastoma, is not a presenting feature of retinocytoma [7]. Sometimes, retinocytoma are diagnosed retrospectively during the treatment of retinoblastoma

Fig. 8.1  The ophthalmoscopic appearance of a retinocytoma Type 3. Note translucent grayish retinal mass, calcification, retinal pigment epithelial alteration, and chorioretinal atrophy

101

8  Retinocytoma or Retinoma

a

b

Fig. 8.2  A one and a half-year-old girl with bilateral multifocal retinoblastoma who had been treated with chemotherapy (VEC × 6 cycles) without focal therapy. The left eye shows two tumors (a). The tumor located temporal to the fovea had a clear cystoid space and was considered as

Box 8.1 Salient Features of Retinocytoma

Retinocytoma is a benign manifestation of RB1 gene mutation. The ophthalmoscopic appearance resembles the spectrum of retinoblastoma regression patterns observed after irradiation. Presence of a translucent grayish retinal mass, calcification, retinal pigment epithelial alteration, and chorioretinal atrophy with or without associated staphyloma are four diagnostic features. Retinocytoma is not associated with retinal exudation or prominent feeder vessels but may be associated with tortuous sclerosed feeder vessels. Retinocytoma lacks growth over short periods of observation (weeks to months). Retinocytoma can undergo malignant transformation into retinoblastoma.

The majority of retinocytomas remain stable. Photographic regression of retinocytoma with increasing chorioretinal atrophy over prolonged follow-up has been occasionally documented [7, 25]. The mechanisms of tumor regression in retinocytoma are unknown but might involve apoptosis [26] rather than ischemia or immune-­ mediated necrosis [27].

retinocytoma. The tumor was observed without focal therapy. Over a period of 3 years, the presumed retinocytoma remained stable in size with clumps of retinal epithelial proliferation along the anterior margin (b)

Diagnostic Evaluation In most of the cases, the diagnosis of retinocytoma can be done with appropriate patient and family history together with indirect ophthalmoscopy. However, fluorescein angiography, ultrasonography, and optical coherence tomography (OCT) can be useful ancillary studies.

Fluorescein Angiography Fluorescein angiography of retinocytoma shows prominent superficial network of fine vessels in the arterial phase without significant leakage in the venous or late phase [9]. Ultrasonography Ultrasonography is useful to measure tumor size and to demonstrate calcified lesions that show characteristic features including acoustic solidity and shadowing due to calcification within the mass on B-scan ultrasonography. A-scan ultrasonography shows a sharp anterior border, high internal reflectivity, and attenuation of orbital echoes posterior to the tumor.  ptical Coherence Tomography (OCT) O OCT is particularly helpful for the differential diagnosis of small- to medium-sized retinal tumors. Thus, small retinocytoma appear as round hyperreflective lesions involving the outer

102

retinal layers, while the overlying inner retinal layers drape over the lesion, and the nerve fiber layer remains intact, unlike astrocytic hamartoma or myelinated nerve fibers that affect the nerve fiber layers (see Sect. 8.5). OCT can also reveal small invisible foci of calcifications. Finally, it enables to accurately evaluate and follow up the dimensions of small lesions. It remains an area of active debate whether OCT can differentiate between retinocytoma and retinoblastoma.

Histopathology Hematoxylin and eosin of retinocytoma shows well-differentiated prominent photoreceptors (fleurettes) with normal-sized nuclei with no pleomorphism or mitotic activity (Fig. 8.3) [28].

a

R. C. Bowen et al.

Interestingly, histopathologic analysis of the eyes enucleated for retinoblastoma has revealed the presence of retinocytoma adjacent to both normal retina and in up to 16% of retinoblastoma tumors [16], conforming the hypothesis that retinoblastoma is the final malignant result on a continuum of clonal progression from normal to benign to malignant cells after loss of RB1 [16–18].

Differential Diagnosis Despite characteristic ophthalmoscopic features of retinocytoma outlined above, certain entities can closely resemble retinocytoma. Clinical features allowing to differentiate retinocytoma form retinoblastoma, astrocytic hamartoma, and myelinated nerve fibers are summarized in Table 8.1.

b

c

Fig. 8.3  Histopathology of retinocytoma. Macroscopic view showing pseudocystic appearance (a). On light microscopy, the tumor is composed of benign cells (b). Note photoreceptor differentiation on electronmicrophotograph (c)

103

8  Retinocytoma or Retinoma Table 8.1  Differential diagnosis of retinocytoma Feature Calcification Chorioretinal atrophy RPE changes Feeder vessels

Retinoblastoma White, chunky Absent

Exudation Growtha Association

Absent Present 13 q deletion syndrome

Present Present

Retinocytoma White, chunky Present in older patients but absent in early retinocytoma Present Absent (except sclerosed and tortuous) Absent Absent 13 q deletion syndrome

Astrocytic hamartoma Yellow, spherical Absent

Myelinated nerve fibers Absent Absent

Absent Absent

Absent Vessels obscured

May be present Absent Tuberous sclerosis

Absent Absent None

RPE retinal pigment epithelium Short-term growth observed over weeks to months

a

Retinoblastoma From a clinical standpoint, it is of utmost importance to differentiate retinocytoma from retinoblastoma. Retinoblastoma is usually diagnosed prior to the age of 5, while retinocytoma is usually diagnosed in adults. Although calcification is seen in both tumors, areas of chorioretinal atrophy and associated retinal pigment epithelial changes are uncommon in untreated retinoblastoma. In addition, dilated, tortuous retinal feeder vessels are a feature of retinoblastoma rather than retinocytoma. Despite these differences, it may be impossible to differentiate a small retinoblastoma from retinocytoma. Characteristically, retinoblastoma will show growth within 4–6 weeks, whereas retinocytoma will appear unchanged. In cases of doubt, it may be more prudent to treat a small tumor as a retinoblastoma rather than observe for growth especially, if treatment is not expected to lead to significant visual loss.

Astrocytic Hamartoma Astrocytic hamartoma, a benign retinal tumor, can also closely resemble retinocytoma. Calcification when present can demonstrate subtle differences, as they tend to be dull and chalky white in a retinocytoma, whereas calcifications in an astrocytic hamartoma are more glistening yellow resembling fish eggs. Surrounding retinal pigment epithelial alterations are a common find-

ing in retinocytoma but are typically absent in astrocytic hamartoma as they are situated superficially in the retina. OCT shows typically a thickened nerve fiber layer by the tumor with moth-eaten optically empty spaces, whereas the underlying retinal layers are intact [29]. In contrast, small retinocytoma or retinoblastoma appears as a homogenous sphere within the retina [30]. Although uncommon, presence of hard exudation supports the diagnosis of astrocytic hamartoma rather than retinocytoma [31].

Myelinated Nerve Fibers Myelinated nerve fibers sometimes mimic a retinocytoma. However, myelinated nerve fibers are usually located at or adjacent to the optic disc margin, show a more fibrillated margin, are flat without any elevation, and are not calcified.

Treatment and Follow-Up While many retinocytomas remain benign for the lifetime of an individual, malignant transformation into retinoblastoma can also occur with an estimated transformation rate of 4% [7, 9, 11]. Thus if suspected, retinocytoma should initially be closely followed, especially in children, to confirm the absence of growth. Later, dilated ocular examination should be performed at least once a year. Patients should be warned for possible ocular

104

symptoms (floaters, vision/visual field decrease, etc.) necessitating an earlier control. First-degree relatives should be ophthalmoscopically screened for similar lesions. Genetic testing for RB1 mutations and genetic counseling related to the risk of retinoblastoma to the offsprings should be offered. Although retinocytoma carries similar genetic implications as germ line retinoblastoma, a review of the large published series of patients with retinocytoma suggests that second malignant neoplasms are rare in patients with retinocytoma [1, 5, 6, 32, 33]. It is possible that mechanisms that play a protective role in inducing retinocytoma also protect the extraocular cells from the development of second malignant neoplasms [34, 35]. In cases of malignant transformation, patients should be referred to a specialized ocular oncology center to ensure appropriate management. Treatment may entail chemotherapy, radiotherapy, focal treatments, or even enucleation depending upon the extent of the disease.

References 1. Gallie BL, Ellsworth RM, Abramson DH, et al. Retinoma: spontaneous regression of retinoblastoma or benign manifestation of the mutation? Br J Cancer. 1982;45(4):513–21. 2. Margo C, Hidayat A, Kopelman J, et al. Retinocytoma. A benign variant of retinoblastoma. Arch Ophthalmol. 1983;101(10):1519–31. 3. Aaby AA, Price RL, Zakov ZN.  Spontaneously regressing retinoblastomas, retinoma, or retinoblastoma group 0. Am J Ophthalmol. 1983;96(3):315–20. 4. Abramson DH.  Retinoma, retinocytoma, and the retinoblastoma gene. Arch Ophthalmol. 1983;101(10):1517–8. 5. Lommatzsch PK, Zimmermann W, Lommatzsch R.  Spontaneous growth inhibition in retinoblastoma. Klin Monatsbl Augenheilkd. 1993;202(3):218–23. 6. Balmer A, Munier F, Gailloud C. Retinoma. Case studies. Ophthalmic Paediatr Genet. 1991;12(3):131–7. 7. Singh AD, Santos CM, Shields CL, et al. Observations on 17 patients with retinocytoma. Arch Ophthalmol. 2000;118(2):199–205. 8. Zhang Q, Chen Y, Wu Z, et al. Retinoma and phthisis bulbi of retinoblastoma. 1. Clinical and genetic analysis. Yan Ke Xue Bao. 1992;8(3):117–21. 9. Abouzeid H, Balmer A, Moulin AP, et al. Phenotypic variability of retinocytomas: preregression and postregression growth patterns. Br J Ophthalmol. 2012;96(6):884–9.

R. C. Bowen et al. 10. Lueder GT, Heon E, Gallie BL.  Retinoma asso ciated with vitreous seeding. Am J Ophthalmol. 1995;119(4):522–3. 11. Eagle RC Jr, Shields JA, Donoso L, et al. Malignant transformation of spontaneously regressed retinoblastoma, retinoma/retinocytoma variant. Ophthalmology. 1989;96(9):1389–95. 12. Gallie BL, Phillips RA, Ellsworth RM, et al. Significance of retinoma and phthisis bulbi for retinoblastoma. Ophthalmology. 1982;89(12):1393–9. 13. Balmer A, Munier F, Gailloud C. Retinoma and phtisis bulbi: benign expression of retinoblastoma. Klin Monatsbl Augenheilkd. 1992;200(5):436–9. 14. Knudson AG Jr. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A. 1971;68(4):820–3. 15. Parma D, Ferrer M, Luce L, et al. RB1 gene mutations in Argentine retinoblastoma patients. Implications for genetic counseling. PloS One. 2017;12(12):e0189736. 16. Dimaras H, Khetan V, Halliday W, et  al. Loss of RB1 induces non-proliferative retinoma: increasing genomic instability correlates with progression to retinoblastoma. Hum Mol Genet. 2008;17(10): 1363–72. 17. Mastrangelo D, De Francesco S, Di Leonardo A, et al. Does the evidence matter in medicine? The retinoblastoma paradigm. Int J Cancer. 2007;121(11):2501–5. 18. Abouzeid H, Schorderet DF, Balmer A, et al. Germline mutations in retinoma patients: relevance to low-penetrance and low-expressivity molecular basis. Mol Vis. 2009;15:771–7. 19. Sampieri K, Amenduni M, Papa FT, et  al. Array comparative genomic hybridization in retinoma and retinoblastoma tissues. Cancer Sci. 2009;100(3): 465–71. 20. Mataftsi A, Zografos L, Balmer A, et al. Chiasmatic infiltration secondary to late malignant transformation of retinoma. Ophthalmic Genet. 2012;33(3):155–8. 21. Singh AD, Garway-Heath D, Love S, et al. Relationship of regression pattern to recurrence in retinoblastoma. Br J Ophthalmol. 1993;77(1):12–6. 22. Garoon RB, Medina CA, Scelfo C, et al. Retinocytoma with vitreous seeding: new insights from enhanced depth imaging optical coherence tomography and high-resolution posterior segment ultrasonography. Retin Cases Brief Rep. 2018. 23. Mashayekhi A, Shields CL, Eagle RC Jr, et al. Cavitary changes in retinoblastoma: relationship to chemoresistance. Ophthalmology. 2005;112(6):1145–50. 24. Morris WE, LaPiana FG.  Spontaneous regression of bilateral multifocal retinoblastoma with preservation of normal visual acuity. Ann Ophthalmol. 1974;6(11):1192–4. 25. Lam A, Shields CL, Manquez ME, et al. Progressive resorption of a presumed ­spontaneously regressed retinoblastoma over 20 years. Retina. 2005;25(2):230–1. 26. Nork TM, Poulsen GL, Millecchia LL, et al. p53 regulates apoptosis in human retinoblastoma. Arch Ophthalmol. 1997;115(2):213–9.

8  Retinocytoma or Retinoma 27. Papac RJ. Spontaneous regression of cancer. Cancer Treat Rev. 1996;22(6):395–423. 28. Mendoza PR, Specht CS, Hubbard GB, et  al. Histopathologic grading of anaplasia in retinoblastoma. Am J Ophthalmol. 2015;159(4):764–76. 29. Shields CL, Say EAT, Fuller T, et al. Retinal astrocytic hamartoma arises in nerve fiber layer and shows “Moth-Eaten” optically empty spaces on optical coherence tomography. Ophthalmology. 2016;123(8):1809–16. 30. Soliman SE, VandenHoven C, MacKeen LD, et al. Optical coherence tomography-­ guided decisions in retinoblastoma management. Ophthalmology. 2017;124(6):859–72.

105 31. Giles J, Singh AD, Rundle PA, et al. Retinal astrocytic hamartoma with exudation. Eye. 2005;19(6):724–5. 32. Moll AC, Imhof SM, Bouter LM, et al. Second primary tumors in patients with retinoblastoma. A review of the literature. Ophthalmic Genet. 1997;18(1):27–34. 33. Korswagen LA, Moll AC, Imhof SM, et al. A second primary tumor in a patient with retinoma. Ophthalmic Genet. 2004;25(1):45–8. 34. Bremner R, Du DC, Connolly-Wilson MJ, et  al. Deletion of RB exons 24 and 25 causes low-­penetrance retinoblastoma. Am J Hum Genet. 1997;61(3):556–70. 35. Sakai T, Ohtani N, McGee TL, et al. Oncogenic germline mutations in Sp1 and ATF sites in the human retinoblastoma gene. Nature. 1991;353(6339):83–6.

9

Retinoblastoma: Genetic Counseling and Testing Meghan J. DeBenedictis and Arun D. Singh

Introduction Genetic counseling for retinoblastoma is simple at first glance but complex in practice. The appearance of simplicity, in part, lies with the monogenic nature of disease, with the RB1 gene being the only gene involved for the vast majority of cases (>98%). The fact that almost all patients with bilateral retinoblastoma have a detectable germline mutation in RB1 reinforces that perception. However, for those with unilateral retinoblastoma, genetic counseling is less straightforward as only about 15–20% will have a germline mutation [1, 2]. Although most patients are young children whose parents have specific concerns, genetic counselors must also be prepared to counsel adult RB survivors who are at risk for second, non-ocular cancers. Mosaicism and low-penetrance mutations may make risk assessments difficult. The optimum surveillance strategy for second primary cancers in retinoblastoma survivors has not been developed making it difficult to offer guidance to mutation carriers who want to manage their cancer risk. Finally, pregnant patients whose fetuses are at risk for retinoblastoma pose challenges for the geneticist since RB1 gene testing must be completed within a narrow time frame. M. J. DeBenedictis (*) · A. D. Singh Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA e-mail: [email protected] © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_9

In some centers, a genetics professional works with the retinoblastoma team. In other settings, the treating ophthalmologist must communicate genetic information to the patient, in some instances without having had special training or experience in the field of genetics. Further, this diagnosis is often devastating to the parents of the patient. A lack of mental health professionals or social workers on the retinoblastoma team may require the treating ophthalmologist to assist with not only the emotional fallout of the diagnosis and prognosis but also with the confusion that results from the complex genetic information provided. Genetic counselors have specialized training in genetics and in psychosocial counseling and support, which makes them a valuable part of the retinoblastoma patient care team. The impact of genetic counseling and testing reaches across generations and even into the future as it affects reproductive decisions regarding future children. The purpose of this chapter is to help the professional providing genetic counseling, whether experienced or otherwise, to be successful when encountering this particular patient population.

Background  ho Is the Patient in Genetic W Counseling? It may be helpful to redefine the concept of the patient in the context of genetic counseling. The 107

108

“patient” in fact consists of multiple people: the parents/guardians and the affected individual. Often, depending on the family, the patient list can further extend to include siblings, the extended family, and their offspring. It is important for the geneticist or medical professional providing the counseling to keep in mind that the family as a whole is the patient and multiple members may rely on the information being provided. This point of view gives the provider a unique perspective and fosters patient care and communication.

 ho Should Be Referred for Genetic W Counseling and Testing? Approximately 30% of probands with a negative family history will have bilateral disease, and 60% of probands with a negative family history will have unilateral retinoblastoma. Individuals with a family history and/or with bilateral retinoblastoma may be more likely to be referred for genetic counseling because they almost always have an underlying germline RB1 gene mutation. Individuals with unilateral retinoblastoma, though, have just as much to gain from genetic counseling. A common mistake in counseling is minimizing the risk for a germline mutation in the individual with unilateral retinoblastoma and no family history. These patients may mistakenly be advised that the chance of an RB1 germline mutation is so low that testing is not warranted. Only 10% of children with either unilateral or bilateral retinoblastoma have a positive family history, and a germline mutation is found in about 15% or 1 in 7 of sporadic unilateral retinoblastoma patients. Retinoblastoma patients, especially those with a unilateral tumor, may be discouraged from RB1 gene testing because DNA studies are expensive and the yield is low. The potential problems, such as non-informative results if a tumor sample is not available or the chance of undetected mosaicism, may be given as further justification for not offering RB1 gene analysis. However, when a germline mutation is detected, all aspects of care  – treatment, prognosis for second tumors, and reproductive counseling  – are impacted.

M. J. DeBenedictis and A. D. Singh

Conversely, individuals with sporadic unilateral retinoblastoma who do not have a germline mutation will need fewer examinations under anesthesia and less intense monitoring for retinoblastoma in the unaffected eye. Identification of the RB1 mutation reduces overall healthcare expenditures by identifying those children who are at risk for additional intraocular tumors and sparing patients and their relatives who test negative for the RB1 germline mutation from unnecessary screening evaluations. Thus, genetic testing for both bilateral and unilateral cases may not only affect the management of the proband but also detect at-risk members of the family who may harbor the RB1 mutation. This information is important for families planning on having additional children and for adult RB survivors who plan to become pregnant.

The Role of the Genetic Counselor The role of the genetic counselor is twofold. First, the counselor must assess risk and communicate this complex and difficult information to the patient and guide the family through the genetic testing process when they choose to go forward. The second part of the job, which is equally challenging, is to communicate the genetic information to the other members of the team and help structure a personalized plan for treatment and long-term management. To be effective in both aspects of the job requires familiarity with this rare disorder and integration into the retinoblastoma team. Not surprisingly, this can be more difficult than it might appear at first glance. Retinoblastoma is rare, and, because of this, few genetic counselors or geneticists outside of pediatric cancer centers have counseled more than an occasional family with this disorder. In some situations, the treating ophthalmologist fills this role. The genetic counselor may be at a disadvantage due to inexperience with retinoblastoma, while the ophthalmologist is at a disadvantage because of lack of knowledge about the limitations of genetic testing and genetic counseling. Commonly, but not optimally, a pregnancy inspires the initial referral for genetic

9  Retinoblastoma: Genetic Counseling and Testing

counseling for a parent of an affected child or in an adult retinoblastoma survivor. In this situation, genetic counseling may take place in the context of a prenatal diagnosis clinic, and the focus may be inappropriately limited to reproductive issues. For instance, a patient might be told that it is not worth pursuing RB1 mutation analysis because results would not be back in time to make a diagnosis in the current pregnancy. However, the implications of RB1 gene testing are much broader and deserve a comprehensive approach. The consequences of having a germline RB1 mutation are lifelong and serious. For these reasons, it is better to start the genetic counseling process prior to a pregnancy. Finally, a genetic counselor, who is both familiar with the disease and affiliated with a comprehensive retinoblastoma team, should provide detailed and thorough counseling. Genetic information is valuable to all members of the team. The well-integrated team incorporates family and genetic information into their regular protocols. For example, with low-­penetrance mutations in mind, the ophthalmologist would routinely examine the parents and siblings of a retinoblastoma patient to check for retinocytoma/retinoma. The ophthalmologist can use the genetic information to modify the management plan accordingly, perhaps by performing office examinations with ultrasound instead of examinations under anesthesia after the likelihood of an RB1 germline mutation has been reduced. Genetic counselors not only explain the genetic basis of retinoblastoma to patients but can also assist in the management care plan of patients and their families, interpret and relay genetic test results, provide recurrence risk information, and serve as a liaison between the ophthalmologist and other medical specialists. They can also ensure that the appropriate type of genetic testing is conducted.

Preparing the Family/Patient for Genetic Counseling Preparation, both for the clinician/counselor and for the family of a retinoblastoma patient, can improve the genetic counseling experience

109

for all concerned. Experience has shown us that without adequate preparation, clinicians and counselors may not address critical issues, and families may misinterpret and mistrust genetic information. The first step in preparing families is giving them a picture of the multidisciplinary retinoblastoma team, introducing the team members and describing the roles that they play. The team includes their pediatric ophthalmologist, pediatric oncologist, as well as nurses, often an ocularist, and social worker. A comprehensive team also includes genetic professionals, such as genetic counselors and/or medical geneticists. If these roles, and how they integrate with one another, are explained early in the process of diagnosis and treatment, genetic counseling will become a normal and expected part of the family’s experience. Without this broad view, genetic counseling can become one more frightening and unexpected event that families experience. Under these circumstances, genetic counseling is difficult and less effective for all concerned. It is also useful to emphasize the protocols that the family will encounter with each specialist. This gives the family a “road map” of care and the expectation of what the experience will entail. When viewed from a large perspective, parents can see where their child is in the “big picture,” and there often are fewer surprises. Instead of increasing anxiety, this approach, especially when adopted from the outset and presented with compassions and empathy, may in fact decrease anxiety. The parents see the terrain ahead and work with the clinicians who guide them through unfamiliar territory.

The Pedigree and Family History The family history, documented in a three or more generation pedigree, is the fundamental working tool for a genetics professional. As a graphic representation of the family history, it neatly summarizes information that would otherwise be scattered throughout the chart. The patient’s age at the time of onset of the retinoblastoma and the tumor laterality should be recorded. The ages of

110 Fig. 9.1  This pedigree represents an apparently sporadic bilateral retinoblastoma in the proband, but genetic testing of the parent revealed that the unaffected mother is mosaic for the RB1 mutation. In this case, the mother must be counseled regarding her own reproductive and cancer risks. This case highlights the importance of testing all unaffected parents for the germline mutation found in their child

M. J. DeBenedictis and A. D. Singh Three generation pedigree. Unafftected parent of girl with bilateral retinoblastoma was found to be mosaic with known RB1 gene mutation.

28

30

2 Diagnosed age 4 months = Billateral retinoblastoma (proband) RB1 mutation carrier (mosaic), = unaffected

the parents and any cancers in the family should be noted. The pedigree should be referred to at each visit and updated regularly. It is incomplete until parents and siblings have had eye examinations to rule out retinocytoma/retinoma. It is modified with test results. If the disease in the proband advances from unilateral to bilateral retinoblastoma or when there is a positive family history of retinoblastoma in previous generations, the implications for subsequent generations are clear at a glance (Fig. 9.1).

Confounding Factors Many confounding factors can complicate what may seem to be a simple pedigree or an apparently nonfamilial case. A complete genetic assessment is important for any child with retinoblastoma and unexpected findings or developmental delay. When these scenarios are understood, the counselor will be better prepared to avoid these common pitfalls.

Chromosome 13q14 Deletions Deletion of chromosome 13q14, the locus of the RB1 gene, and neighboring regions on the long

arm of chromosome 13 can lead to intellectual disability and retinoblastoma. The size of the deletions and the phenotype may vary. However, genotype-phenotype correlation between the size of the deletion and the severity of the clinical presentation has not been established. Some individuals with small deletions of this area are developmentally normal. In children with a deletion involving chromosome 13q14, developmental delay or intellectual disability may be appreciated before retinoblastoma is diagnosed (Fig. 9.2). Any child diagnosed with a 13q deletion who has not had an ophthalmic examination requires one urgently. Children with chromosome13q14 deletions may develop retinoblastoma at a somewhat later age, and often they develop only one tumor. One may speculate that when the “first hit” is a large deletion, gene conversion, a common pathway for the “second hit” in retinoblastoma, may lead to premature cell death instead of cancer. Paradoxically by reducing cell viability with the “second hit,” large mutations can act as low-­ penetrance mutations. High-resolution chromosome analysis and fluorescence in situ hybridization (FISH) for 13q14 can rule out both macroscopic and microscopic deletions [3]. When a chromosome deletion is discovered in an affected patient, a microarray

111

9  Retinoblastoma: Genetic Counseling and Testing

Fig. 9.2  This boy was referred to a pediatric neurologist for hypotonia before his retinoblastoma was diagnosed. He has a chromosome 13q deletion visible on routine banding. He has an ocular prosthesis following enucleation of his left eye for retinoblastoma. He also has developmental delay and mild facial dysmorphism: broad forehead with frontal bossing, arched eyebrows, hypertelorism, small mouth

should be done to determine the size and extent of the deletion. The associated intellectual disability has been mapped to the nearby NUFIP1 and PCDH8 genes. The parents should also have testing with microarray or FISH for 13q14 to rule out a similar deletion, inversion, or other heritable chromosome anomaly as this would influence recurrence risks for future children. All children with retinoblastoma should be monitored for age-appropriate developmental milestones. However, intellectual disability in children with retinoblastoma is not always due to a deletion of chromosome 13q14. Those who are developmentally delayed deserve a prompt and thorough evaluation.

Mosaicism Mosaicism, in which the RB1 mutation is present in some but not all cells of the affected patient, is

common in the first affected member of a family with retinoblastoma [4–10]. Patients who are mosaic for RB1 germline mutations often have unilateral retinoblastoma with a later onset. They also tend to lack a family history of the disease. However, mosaicism has also been seen in patients with bilateral retinoblastoma and in unaffected parents of affected children. There is no technique that will reliably detect all cases of mosaicism. Gene sequencing may miss mosaicism when less than 20% of cells have a mutation. Even when more than one tissue is studied, mosaicism can never be completely ruled out in the first affected member of the family. When a child with unilateral sporadic retinoblastoma has normal RB1 gene test results, the chance of a germline mutation is never zero. There is always a small residual risk for undetected mosaicism. This residual risk figure will depend on the test and laboratory used for the genetic testing. The counseling session should include a discussion of the possibility of low-level germline mosaicism (Fig. 9.3). -

-

+

+

Fig. 9.3 Germline/gonadal mosaicism. The proband (arrow) in this family was diagnosed with bilateral multifocal retinoblastoma at 9  months of age. RB1 analysis detected a mutation. Her parents’ leukocyte DNA testing was negative. Her brother was diagnosed with unilateral retinoblastoma at 10 days of age and contralateral retinoblastoma at 6 weeks. He carried the same RB1 mutation as the proband. One of the proband’s parents must have germline mosaicism. Squares, males; circles, females; and white symbols, unaffected members. +: RB1 mutation-­ positive; −: RB1 mutation-negative. (Reprinted from Moline and Singh [29]. With permission from Future Medicine Ltd.)

112

M. J. DeBenedictis and A. D. Singh

A genetic counselor should consider mosaicism in a multi-generation retinoblastoma family when the first affected individual, the “founder,” has unilateral retinoblastoma or a late-­onset tumor yet their affected offspring have bilateral, early-onset retinoblastoma. In a retinoblastoma family with two or more affected generations, including a parent and child, it is always best to start the testing process in the child from the second affected generation. This avoids the possibility of a false-negative result due to undetected mosaicism in the first affected member of the family. Mosaicism, when it occurs, is limited to the first affected member of the pedigree. Mosaicism is not hereditary. The affected child of a mosaic individual inherits the deleterious mutation and is not mosaic.

penetrance. Penetrance varies with the in-frame (low penetrance) or out-of-frame (high penetrance) nature of splice site mutations. Another mechanism for variable penetrance occurs when exon 1 mutations produce functional transcripts through alternate mRNA transcription [14, 15]. Dilated eye exams for the parents of RB patients should be conducted in order to rule out the presence of a retinocytoma/retinoma. A parent who is unaffected and healthy can share the same RB1 germline mutation with their affected child. So-called pseudo low penetrance may also lead the counselor astray. This refers to two affected relatives in a large pedigree giving the appearance of familial retinoblastoma, when in fact, the retinoblastoma tumors arose from independent and unrelated sporadic RB1 mutations [16].

Low-Penetrance Mutations and Variable Expressivity

 volving Phenotypes and Changing E Pedigrees

Low-penetrance mutations, often due to missense mutations in RB1 that do not truncate the protein product, and variable expressivity are well documented in the retinoblastoma literature (Fig. 9.4) [11–13]. The specific type of RB1 mutation is important in determining whether to expect complete penetrance, high penetrance, or variable

Evolving phenotypes and changing pedigrees sometimes make it necessary to revise risks and re-counsel families. After counseling a patient with unilateral sporadic retinoblastoma, the counselor may have to revise their risk assessment after the discovery of a second affected individual in the family or a second tumor in the

Unilateral retinoblastoma Bilateral retinoblastoma

Fig. 9.4  Low-penetrance family. The proband (arrow) was diagnosed with unilateral retinoblastoma at 6 months of age and was found to carry an RB1 missense mutation. The mutation was inherited from her father who had no signs of retinoblastoma or retinocytoma. All at-risk relatives should have genetic testing for the familial mutation

to determine risk of tumor development as well as transmission risks. Squares, males; circles, females; and white symbols, unaffected members. (Reprinted from Moline and Singh [29]. With permission from Future Medicine Ltd.)

9  Retinoblastoma: Genetic Counseling and Testing

proband. This is a special concern when counseling the parents of a young infant with a unilateral tumor. This also highlights the importance of educating families to provide their retinoblastoma care team with new cancer or tumor diagnoses in the family. Counseling should always include the parents, whose status may change as a result of genetic testing. An unaffected parent may be surprised to find that they harbor a germline mutation and are at increased risk for non-ocular cancers. This parent now needs to be monitored and counseled. It is also important to counsel the parent who had unilateral RB themselves but did not know their RB1 mutation status until the birth of a child with bilateral RB. This affected parent is now clearly a germline mutation carrier and at risk for secondary tumors. This illustrates the need to consider the affected parent as a patient, even at a time when most of the medical attention is focused on the affected child.

 he Isolated Case of Unilateral T Retinoblastoma It is the isolated case of unilateral retinoblastoma that is most problematic for the genetic counselor. The lack of a family history and an older age at onset of unilateral sporadic retinoblastoma do not exclude a germline RB1 mutation [17]. Although most children with unilateral retinoblastoma will not have heritable retinoblastoma, a significant minority (15%) will have a germline mutation in the RB1 gene. If a germline RB1 mutation is detected, modification of tumor and cancer surveillance and the provision of reproductive counseling are needed. Reducing the morbidity and mortality associated with retinoblastoma is an important goal of genetic counseling.

Prenatal Diagnosis When a parent has a known RB1 germline mutation, options for testing an embryo or fetus including prenatal diagnosis after natural conception and preimplantation genetic diagnosis

113

(PGD) should be discussed with parents. PGD is performed by testing an embryo for the presence of the RB1 mutation following in vitro fertilization (IVF). Assisted reproductive technology including in  vitro fertilization is often seen, at first glance, as an attractive option for a parent who has an RB1 mutation. Using donor eggs or sperm may remove the risk of retinoblastoma in the pregnancy. When using the gametes from the couple, PGD and selective implantation of only unaffected embryos obviate the need for the consideration of a pregnancy termination. However, IVF is expensive, and it is associated with more pregnancy complications such as preterm delivery, low birth weight, and multiple gestation compared to a natural conception. Furthermore, the risk for chromosome anomalies and birth defects of all types is increased by about 25% in IVF-conceived fetuses, possibly related to the use of ovarian-stimulating drugs or other aspects of the procedure (e.g., intracytoplasmic sperm injection). PGD itself is not foolproof: it has its own risks and limitations. Both false-positive and false-negative PGD results have occurred for various chromosomal and genetic disorders, so prenatal diagnosis with chorionic villous sampling (CVS) or amniocentesis might be considered to confirm normal findings. In fact, it is important to note that de novo RB1 mutations have occurred in children conceived by IVF [18]. Therefore, in spite of the benefits they bring, the risks and limitations associated with IVF and PGD may make this option less attractive than a natural c­ onception. Chorionic villus sampling (at 11–13  weeks gestation) and amniocentesis (after 15  weeks gestation) are diagnostic tests with high sensitivity and specificity. Targeted mutation analysis can be done on tissue obtained from these methods. Normal fetal results can provide reassurance which is a benefit to the anxious couple, which is often unfairly minimized. When the presence of an RB1 mutation is established in the fetus, this information can be used to either plan for surveillance or make decisions regarding continuing the pregnancy. In experienced hands, these procedures are associated with a  ≤1% risk of complications, including miscarriage. Maternal cell contamination and other risks and

114

limitations of the procedures should be discussed by the genetic counselor as part of the informed consent process. If the mutation is identified prenatally, fetal ultrasound can be used to identify large intraocular tumors. Preterm delivery of an affected infant with an ocular tumor evident on fetal ultrasound exam may offer some benefits by allowing for early treatment of tumors or early ocular examination [19]. Whether prenatal diagnosis is performed or not, cord blood or an infant’s peripheral blood may be used for diagnostic or confirmatory testing after delivery. Since genetic test results are not always available in a short time, at-risk children should have ocular evaluation by an experienced ophthalmologist soon after delivery.

 he Limits of Technology and NonT informative Results The limits of technology need to be reviewed in detail with the family prior to genetic testing as part of the informed consent process. The possibility of undetectable mutations, false-positive and false-negative results, and mosaicism should be discussed with the family.

Mutation Detection When direct DNA testing shows an abnormal result, the family can be counseled accordingly. However, normal results should always be interpreted with caution as sensitivity for mutation detection is not 100%. Technical limitations of gene sequencing analysis contribute to this lack of sensitivity because this method does not reliably detect large deletions or lowlevel mosaicism. Patients also need to be aware of the possibility of non-informative results. This refers to the situation in which RB1 gene analysis in blood appears to be negative but a cryptic mutation is in fact present. This situation can usually be avoided if the tumor tissue is tested at the same time as the blood sample.

M. J. DeBenedictis and A. D. Singh

As retinoblastoma tumors will typically contain two RB1 gene mutations, by starting the testing process with a tumor sample, the sensitivity of the testing technique to detect the mutations in question can be determined. When DNA testing on tumor tissue does not reveal both mutations, it is evident that the same mutation would likely not be detectable in blood. This lack of sensitivity cannot be discerned when only blood is studied. For this reason, in all unilateral retinoblastoma cases treated with enucleation, fresh tumor tissue should be frozen so that it is available for genetic analysis later. Even in the best laboratories, using a variety of DNA techniques, RB1 gene analysis yields a detection rate of about 96% (Table 9.1). With this in mind, the chance of misinterpreting an undetected RB1 mutation as a normal result (false negative) should be discussed whenever blood alone is studied. This is particularly applicable as fewer eyes are being enucleated with new therapies.

 he Future in DNA Testing T for Retinoblastoma We have discussed germline mutation testing, but there is a parallel body of work on genetic analysis of the tumor itself. Germline mutations are found in the RB1 gene in body tissues outside the tumor; however, there are a variety of other somatic mutations (gene mutations and chromosome changes found in the tumor itself). Specific chromosome changes in addition to those found on chromosome 13 such as +1q, +2p, +6p, −13, and −16 have been recognized in retinoblastoma tumors since the early 1980s [20]. Recent reports suggest that DNA analysis in these and other chromosome regions may shed light on progression of malignancy events in retinoblastoma [21– 24]. These findings may have clinical relevance in the future. Of similar potential clinical interest is the finding that loss of specific metastasis suppressor genes (MSGs) has been associated with a much higher risk for metastatic growth in other human cancers [25–27].

115

9  Retinoblastoma: Genetic Counseling and Testing Table 9.1  Retinoblastoma gene testing techniques: limitations and detection rates Technique Cytogenetic analysis

Chromosome analysis FISH for 13q14 Microarray

Direct DNA analysis

Indirect DNA analysis

RB1 gene sequence analysis RB1 quantitative multiplex PCR RB1 allele-specific PCR Methylation of RB1 promoter Primary MYCN amplification Linkage analysis

Limitations Limited to detection of chromosome 13 translocations, rearrangements, and very large deletions Should be done in conjunction with FISH for 13q14 Limited to detection of large RB1 gene deletions Should be done in conjunction with chromosome analysis Detects small deletions in RB1 Should be done on all deletion cases Limited to detection of small sequence variations Detects small deletions, insertions, point mutations Does not reliably detect mosaicism or splice site changes Limited to detection of deletions and gene rearrangements Limited to cases in which familial mutation is known or tumor mutation is known and mosaicism is suspected Limited to nonhereditary, sporadic unilateral retinoblastoma Limited to nonhereditary, sporadic unilateral retinoblastoma without a RB1 mutation Limited to multigenerational families Mosaicism in proband can lead to false-positive result for unaffected offspring

Detection rate (%)a 5

Greater than 70 20

11

RB1 retinoblastoma gene, FISH fluorescent in situ hybridization, PCR polymerase chain reaction a These percentages are approximations. Actual detection rates may vary by laboratory. Clinicians should consult with a genetic counselor or the lab directly for their current methods and detection rates

RB1+/+ MYCNA Tumors A rare and genetically unique form of retinoblastoma has been recently discovered in which both alleles of the RB1 gene are normal, and instead expression of the MYCN gene on chromosome 2p is considerably amplified [28]. This form of tumor presented with advanced unilateral disease at a young age; however it is otherwise clinically indistinguishable from RB1 mutated tumors. It has exclusively been discovered in 1–3% of children with unilateral retinoblastoma, especially in those under the age of 12 months. Histopathology demonstrates undifferentiated cells with large, prominent, multiple nucleoli and necrosis, apoptosis, and little calcification. The histopathology feature of this form of retinoblastoma resembles that of other MYCN-amplified tumors such as neuroblastoma. The characteristics of this type of

tumor strongly suggest nonhereditary disease, without risk for retinoblastoma in the other eye, for other malignant cancers throughout life, and risk of familial transmission. These tumors are considered more aggressive histopathologically.

Summary The genetics of retinoblastoma is complex and unique. Genetic counseling and RB1 genetic testing have value for patients, especially those with unilateral retinoblastoma. This information is also valuable for the other members of the ­retinoblastoma team who can manage patients whose RB1 status has been clarified more effectively. The genetic counseling process is improved when both patients and physicians are prepared and understand the benefits and limitations of the molecular technology and the cancer

116

M. J. DeBenedictis and A. D. Singh

retinoblastoma associated with dysmorphic features. Genet Couns. 2005;16(1):91–3. 11. Dalamon V, Surace E, Giliberto F, et  al. Detection of germline mutations in argentine retinoblastoma patients: low and full penetrance retinoblastoma caused by the same germline truncating mutation. J Biochem Mol Biol. 2004;37(2):246–53. 12. Huang Q, Dryja TP, Yandell DW.  Distinct Rb gene point mutations in families showing low penetrance of hereditary retinoblastoma. Zhonghua Yi Xue Yi Chuan Xue Za Zhi. 1998;15(3):139–42. 13. Lohmann DR, Brandt B, Hopping W, et  al. Distinct RB1 gene mutations with low penetrance in hereditary retinoblastoma. Hum Genet. 1994;94(4): 349–54. 14. Taylor M, Dehainault C, Desjardins L, et  al. References Genotype-phenotype correlations in hereditary familial retinoblastoma. Hum Mutat. 2007;28(3):284–93. 1. Albrecht P, Ansperger-Rescher B, Schuler A, et  al. 15. Sanchez-Sanchez F, Ramirez-Castillejo C, Weekes Spectrum of gross deletions and insertions in the DB, et  al. Attenuation of disease phenotype RB1 gene in patients with retinoblastoma and assothrough alternative translation initiation in lowciation with phenotypic expression. Hum Mutat. penetrance retinoblastoma. Hum Mutat. 2007;28(2): 2005;26(5):437–45. 159–67. 2. Lohmann DR, Gallie BL. Retinoblastoma: revisiting 16. Munier FL, Wang MX, Spence MA, et  al. Pseudo the model prototype of inherited cancer. Am J Med low penetrance in retinoblastoma. Fortuitous familGenet C Semin Med Genet. 2004;129(1):23–8. ial aggregation of sporadic cases caused by indepen 3. Kallioniemi A, Kallioniemi OP, Waldman FM, et al. dently derived mutations in two large pedigrees. Arch Detection of retinoblastoma gene copy number in Ophthalmol. 1993;111(11):1507–11. metaphase chromosomes and interphase nuclei by 17. Brichard B, Heusterspreute M, De Potter P, et  al. fluorescence in situ hybridization. Cytogenet Cell Unilateral retinoblastoma, lack of familial history and Genet. 1992;60(3–4):190–3. older age does not exclude germline RB1 gene muta 4. Dudin G, Nasr A, Traboulsi E, et al. Hereditary retinotion. Eur J Cancer. 2006;42(1):65–72. blastoma and 13q–mosaicism. Cytogenet Cell Genet. 18. Barbosa RH, Vargas FR, Lucena E, et al. Constitutive 1984;38(3):235–7. RB1 mutation in a child conceived by in vitro fertil 5. Greger V, Passarge E, Horsthemke B. Somatic mosaization: implications for genetic counseling. BMC icism in a patient with bilateral retinoblastoma. Am J Med Genet. 2009;10:75. Hum Genet. 1990;46(6):1187–93. 19. National Retinoblastoma Strategy Canadian 6. Kivela T, Tuppurainen K, Riikonen P, et al. Guidelines for Care. Strategie therapeutique du retinoRetinoblastoma associated with chromosomal blastome guide clinique canadien. Can J Ophthalmol. 13q14 deletion mosaicism. Ophthalmology. 2009;44(Suppl 2):S1–88. 2003;110(10):1983–8. 20. Squire J, Gallie BL, Phillips RA.  A detailed analy 7. Munier FL, Thonney F, Girardet A, et  al. Evidence sis of chromosomal changes in heritable and non-­ of somatic and germinal mosaicism in pseudo-low-­ heritable retinoblastoma. Hum Genet. 1985;70(4): penetrant hereditary retinoblastoma, by constitu291–301. tional and single-sperm mutation analysis. Am J Hum 21. Grasemann C, Gratias S, Stephan H, et al. Gains and Genet. 1998;63(6):1903–8. overexpression identify DEK and E2F3 as targets of 8. Ribeiro MC, Andrade JA, Erwenne CM, et al. Bilateral chromosome 6p gains in retinoblastoma. Oncogene. retinoblastoma associated with 13q-­ mosaicism. 2005;24(42):6441–9. Possible manifestation of a germinal mutation. Cancer 22. Gratias S, Schuler A, Hitpass LK, et  al. Genomic Genet Cytogenet. 1988;32(2):169–75. gains on chromosome 1q in retinoblastoma: con 9. Sippel KC, Fraioli RE, Smith GD, et  al. Frequency sequences on gene expression and association with of somatic and germ-line mosaicism in retinoblasclinical manifestation. Int J Cancer. 2005;116(4): toma: implications for genetic counseling. Am J Hum 555–63. Genet. 1998;62(3):610–9. 23. Marchong MN, Chen D, Corson TW, et al. Minimal 10. Van Esch H, Aerssens P, Fryns JP. The importance of 16q genomic loss implicates cadherin-11  in retinoexcluding 13q deletion mosaicism in the diagnosis of blastoma. Mol Cancer Res. 2004;2(9):495–503.

surveillance strategies that are currently available. Even when the facts are mastered, genetic counseling for retinoblastoma is further complicated by the psychological and emotional aspects of this disorder. Geneticists, ophthalmologists, psychologists, social workers, and other mental health professionals work best when they work together to help families grapple with the lifelong implications of the information they have been given.

9  Retinoblastoma: Genetic Counseling and Testing 24. Orlic M, Spencer CE, Wang L, et al. Expression analysis of 6p22 genomic gain in retinoblastoma. Genes Chromosomes Cancer. 2006;45(1):72–82. 25. Berger JC, Vander Griend DJ, Robinson VL, et  al. Metastasis suppressor genes: from gene identification to protein function and regulation. Cancer Biol Ther. 2005;4(8):805–12. 26. Kauffman EC, Robinson VL, Stadler WM, et  al. Metastasis suppression: the evolving role of metastasis suppressor genes for regulating cancer cell growth at the secondary site. J Urol. 2003;169(3): 1122–33.

117 27. Keller ET. Metastasis suppressor genes: a role for raf kinase inhibitor protein (RKIP). Anti-Cancer Drugs. 2004;15(7):663–9. 28. Rushlow DE, Mol BM, Kennett JY, et  al. Characterisation of retinoblastomas without RB1 mutations: genomic, gene expression, and clinical studies. Lancet Oncol. 2013;14(4):327–34. 29. Moline J, Singh AD.  Genetics of retinoblastoma and genetic counseling. In: Murray TG, editor. Retinoblastoma: clinical advances and emerging treatment strategies. Published Online 14 Aug 2013. p. 36–48

Retinoblastoma: Treatment Options

10

Jonathan W. Kim, A. Linn Murphree, and Arun D. Singh

Introduction Survival rates for retinoblastoma patients have increased dramatically over the past century, with documented 5-year survival reaching 95–99% in developed countries [1, 2]. Similarly, there have been significant advances in the treatment approaches for intraocular retinoblastoma, driven by a motivation to increase salvage rates and decrease complications. Over the past 50 years, there have been major paradigm shifts in the approaches for managing intraocular retinoblastoma. In the 1960s, external radiation therapy was the primary vision-saving modality for treating the ocular tumors [3]. In the mid-1990s, systemic chemotherapy combined with focal modalities became the dominant treatment strategy, emphaJ. W. Kim (*) Department of Ophthalmology, USC Roski Eye Institute, Keck School of Medicine of USC Children’s Hospital Los Angeles, Los Angeles, CA, USA e-mail: [email protected] A. L. Murphree Department of Ophthalmology, Children’s Hospital Los Angeles, University of Southern California, Los Angeles, CA, USA USC Roski Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA A. D. Singh Department of Ophthalmic Oncology, Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_10

sizing multiple drug chemoreduction protocols and minimizing the use of external beam radiation [4]. Over the past 10  years, there has been growing interest in  local or regional therapies, delivering chemotherapeutic agents directly to the globe or through regional arteries in an attempt to improve cure rates and reduce the morbidity of less selective modalities. In this chapter, we summarize current management approaches for intraocular retinoblastoma, emphasizing the clinical indications for intravenous chemotherapy, external beam radiation, brachytherapy, focal modalities, intra-arterial chemotherapy, and intravitreal injection of chemotherapy.

Classification (Grouping) The International Intraocular Retinoblastoma Classification (IIRC) system offers guidance to clinicians in deciding when and how to manage intraocular retinoblastoma (Chap. 3) [5]. For group A disease, focal modalities are typically adequate to cure the eye (i.e., photocoagulation and cryotherapy). Most group B eyes will require another modality in addition to focal treatments to control the disease and achieve optimal visual outcomes. For example, most group B disease can be successfully treated with 3–6 cycles of intravenous chemotherapy combined with focal consolidation or occasionally a radioactive plaque if the tumor is away from 119

J. W. Kim et al.

120

the posterior pole. Group C retinoblastoma is managed with a similar approach (3–6 cycles of chemotherapy, occasionally brachytherapy), although success rates are slightly lower because of the presence of localized seeding. Modern ocular oncology centers will attempt to save group A–C eyes, even if central vision is poor and the patient has unilateral disease, because of the relatively high success rates achieved with current approaches (80–100%). Conversely, the likelihood of salvaging group D eyes with intensive chemotherapy treatment ranges from 47 to 90% [6], which creates a dilemma for patients with normal or near-­normal vision in the contralateral eye. If the visual potential is poor (i.e., macula destroyed by tumor), a strong case can be made for recommending enucleation for unilateral group D disease and sparing the child the morbidity of 6 months of systemic chemotherapy (or 3–4 months of intra-arterial chemotherapy). For bilateral patients with one group D eye and at least a group B diagnosis in the contralateral eye, the decision to treat with 6 cycles of systemic chemotherapy is less controversial since the better eye also requires treatment. In general, group E eyes should be enucleated. This is because the chance of salvaging such an advanced eye is low (despite all treatments), the visual prognosis is dismal, and the odds that a group E eye harbors high-risk pathologic features are significant (i.e., 24%) [7]. If a group E eye demonstrates optic nerve invasion on the staging MRI study, most centers advocate immediate enucleation with adjuvant chemotherapy being given if high-risk features are confirmed on histopathology. However, in Los Angeles, we have demonstrated that patients with proximal optic nerve enhancement treated with five cycles of neo-adjuvant chemotherapy prior to enucleation achieve good outcomes [8]. The latter strategy was adopted to prevent a positive posterior (i.e., cut end) optic nerve margin and spare the child the morbidity of external beam radiation to the orbit. It should be kept in mind that for bilateral patients, the eye with the more advanced group classification should dictate the intensity of the treatment regimen. However, it should not be assumed that the eye with the

lesser grade tumor will ultimately be the betterseeing eye after the completion of therapy. The American Joint Committee on Cancer (AJCC) publishes a universal staging system for all malignancies including retinoblastoma using the TNM criteria. The most recent 8th edition of the AJCC includes an updated classification system for retinoblastoma that is becoming more widely adopted in the literature [9]. The AJCC classification system for retinoblastoma is similar to the international classification system with some minor differences. Category cT1a includes small tumors of 3  mm or less in any dimension, with cT1b being defined by either larger tumors or tumors within 1.5 mm of the optic nerve or fovea. Category cT2 includes eyes with more than 5  mm of subretinal fluid (cT2a) or tumor seeding (cT2b); category cT2b includes eyes with tumor seeding any distance from the main tumor. Therefore, Group C tumors in the international classification system (with localized seeding) are grouped together with Group D tumors (with distant seeding) in the AJCC classification. Also, there are no size criteria for the advanced cT3 category, which is essentially the same as Group E in the international classification since it includes eyes with ocular phthisis, invasion of the anterior segment, neovascular glaucoma, intraocular hemorrhage, and orbital cellulitis. It should become apparent that cT2b is the most common diagnosis in the AJCC classification since it includes eyes with even localized seeding as well as advanced eyes with tumors filling the majority of the posterior segment but without neovascular glaucoma, intraocular hemorrhage, anterior segment invasion, etc.

Intravenous Chemotherapy In the mid-1990s, there was an effort to increase the use of intravenous chemotherapy and focal treatments and to avoid the use of external beam radiation therapy, mainly because of the growing awareness of the risk for second tumors in retinoblastoma patients (Chap. 11). Prior to 1990, most centers reserved intravenous chemotherapy

10  Retinoblastoma: Treatment Options

for patients requiring adjunctive treatment after enucleation or rescue therapy for extraorbital or metastatic disease (Chap. 18). The recognition of the increased risk for second tumors in patients treated with external beam radiation stimulated many groups in the 1990s to use chemotherapy to treat intraocular retinoblastoma. Over the past two decades, intravenous chemotherapy became the most important conservative (eye-sparing) treatment approach for intraocular retinoblastoma [10–15]. Intravenous chemotherapy is currently used to treat tumors that are too large or widespread to treat with focal modalities such as cryotherapy, thermotherapy, or brachytherapy. Although external beam radiotherapy remains an excellent option for preserving vision in patients with retinoblastoma, most clinicians now use EBR as only as the last resort treatment due to the availability of other safer options. The visual outcomes of intravenous chemotherapy also appear to be comparable to external beam radiation for patients with bilateral disease and visual potential in one or both eyes [3].

Treatment Parameters Although intravenous chemotherapy protocols vary slightly between institutions, most centers are currently treating intraocular retinoblastoma with carboplatin, vincristine, and etoposide as a three-drug regimen given in 3–6 cycles. The regimen of carboplatin, etoposide, and vincristine has been used successfully against extraocular retinoblastoma, as well as other primitive neuroectodermal tumors such as neuroblastoma [16, 17]. Carboplatin, an analogue of cisplatin with less nephrotoxicity and ototoxicity, is an active agent against many brain tumors and is known to cross the blood–brain barrier [18]. Many centers have utilized a two-agent regimen of carboplatin with either etoposide or vincristine with similar outcomes as the three-drug regimen [11, 19–22]. A single-agent chemotherapy regimen (with carboplatin) has also been used successfully by Abramson’s group [23, 24]. The addition of cyclosporine as a P-glycoprotein inhibitor was

121

suggested by Chan to increase chemosensitivity [25, 26]. Their group has demonstrated that some patients who became resistant to multiple cycles of chemotherapy respond to the same regimen given with cyclosporine [10]. Although the success rate in their series of patients seemed to correlate with higher doses of cyclosporine, most centers have not incorporated cyclosporine into their chemotherapy regimens. At CHLA, we utilize the standard three-drug regimen at the following doses: carboplatin (13  mg/kg/ day), vincristine (5.0 mg/kg/day), and etoposide (0.05 mg/kg/day) given for 2 sequential days. For patients less than 6  months of age, carboplatin and vincristine are given at 50% dose, without vincristine to avoid ileus. As previously discussed, group A eyes can typically be managed with focal modalities alone, and group E eyes are usually enucleated. For group B eyes, the standard approach is 3 cycles of chemotherapy and for group C eyes, 3–6 cycles depending on the clinical response. For group D eyes, 6 cycles are almost always given, unless a decision is made to recommend EBR or enucleation before the completion of therapy.

Concomitant Focal Therapy Although the ideal regimen for intravenous chemotherapy has not been determined, most authors agree that chemotherapy must be combined with focal modalities for adequate tumor control (Chap. 10). It is rare for a tumor to be cured with chemotherapy alone, even after 6 cycles [24, 27, 28]. Without laser treatment or cryotherapy, Wilson found that 92% of eyes progressed after completion of chemotherapy [21]. Abramson suggested that focal treatments can usually be delivered after 2–3 cycles, since cumulative reduction in tumor area was near maximal after 2 cycles and the mean reduction in tumor area for the third treatment alone was only 5.4% [24]. Gallie also suggested focal treatment after two cycles if the clinical examination confirmed adequate tumor reduction and resolution of subretinal fluid [10]. Our recommendation is to begin using local modalities when clinical

J. W. Kim et al.

122

judgment indicates that tumor consolidation can be achieved safely with laser treatment, cryotherapy, or brachytherapy (1–3 cycles), waiting for further regression for larger tumors and those in the macula. For tumors in the macula, tumors should be maximally regressed before beginning laser therapy to minimize the size of the posttreatment scotoma. Often, extensive tumors in the posterior pole will shrink enough with chemotherapy to allow treatment with focal modalities in an attempt to preserve at least a portion of the central vision. Typically, laser treatment is used to treat tumors in the posterior pole or small tumors (2 disc diameters or less) in the periphery. For larger tumors in the periphery, peripheral tumors in the periphery not responding to laser therapy, or peripheral tumors with nascent vitreous seeding at the apex, cryotherapy is the modality of choice (Chap. 10).

Efficacy Similar to other treatment modalities, clinical studies examining the efficacy of intravenous chemotherapy demonstrate a correlation with the stage of the intraocular disease. The strategy of

combining intravenous chemotherapy with focal therapy has been associated with a 90–100% chance of avoiding radiation or enucleation for eyes with group A and B tumors (or Reese– Ellsworth groups I–III) [28, 29]. Results for patients with groups D and E (or Reese–Ellsworth groups IV and V eyes) have been less encouraging. Our clinical series suggests that group D eyes treated with primary chemoreduction have a 47% chance of avoiding external beam radiation or enucleation (Table 10.1) [6], although the rate increases to 75% when combined with intravitreal chemotherapy injections for recalcitrant vitreous seeding. In general, group E eyes should be enucleated, particularly if the patient has unilateral disease. There may be clinical situations where a patient with group E disease and bilateral advanced tumors may be treated with intravenous chemotherapy, but clinicians should be prepared to enucleate the group E eye if there is suboptimal clinical response after two cycles. The presence of subretinal or vitreous seeds is a common cause of treatment failure in eyes undergoing intravenous chemotherapy. Systemic chemotherapy definitely causes regression of some tumor seeds in the vitreous cavity and subretinal space, although the response is variable,

Table 10.1  Summary of clinical studies on intravenous chemotherapy of retinoblastoma Author Gallie [9] Kingston [26] Murphree [17] Shields [32] Greenwald [18] Bornfeld [11] Shields [27] Gunduz [14] Friedman [13] Beck [10] Wilson [20] Shields [29] Brichard [12] Rodriguez [21] Schiavetti [19] Antoneli [99] Totals

No. eyes 40 24 35 31 11 12 52 27 75 33 36 158 24 43 58 145 804

Regimen VRES VRE VRE VRE RE VRES VRE VRE VRE RE VR VRE VRE VR RE VRE

Cycles 2–4 3 2 6–7 3 2.6 2.6 6 2–5 6 6 2–6 8 4–8 2–6

V eyes 18 24 21 22 6 7 36 27 30 13 14 75 12 15 17 74 411

EBR 4 20 7 9 5 2 19 16 13 7 8 32 0 8 4 ?

Enuc 1 6 17 0 3 1 8 10 9 5 5 32 2 6 11 ?

None 13 18 0 13 1 4 9 5 14 2 5 27 10 4 1 30 156

F/U (months) 3 60 ? 6 23 7 17 25 13 31 19 28 21 32 53 ?

V vincristine, R carboplatin, E etoposide, S cyclosporine, V eyes Reese–Ellsworth group V eyes, EBR external beam radiation, Enuc enucleated eyes, None eyes avoiding enucleation and external beam radiation, F/U mean follow-up, ? not reported

10  Retinoblastoma: Treatment Options

unpredictable, and typically not complete [19, 28, 30]. In 1996, Shields reported that calcification occurred in 50% of vitreous seeds and 78% of subretinal seeds following 2 cycles of chemotherapy, similar to the findings after external beam radiation [30]. Overall, eyes with vitreous and subretinal seeds developed tumor recurrence at a rate of approximately 46 and 62% at 3 years and 5  years of follow-up, respectively [29]. If some vitreous or subretinal seeds are viable after six cycles of chemotherapy, they will inevitably cause new retinal tumor recurrences. Wilson has also pointed out that seed dispersion can be induced or worsened by chemotherapy: as tumors regress during the initial cycles of intravenous chemotherapy, tumors can fragment and release seeds into the vitreous cavity [21]. Persistence of seeds may also represent inadequate penetration of the chemotherapy to the avascular sites in the vitreous cavity and subretinal space. Another possible cause of treatment failure with intravenous chemotherapy is the development of new tumors. In Lee’s study of single-agent carboplatin, 47% of eyes developed additional tumors during the period of follow-up, and 37% of eyes had new tumors only 1  month after the initial cycle of chemotherapy [23]. The risk of new tumor formation was more than twice as likely if the child was treated before the age of 6 months of age, and nearly all new tumors occurred in the first 2 years of life. These findings confirmed that systemic chemotherapy does not appear to have a protective or prophylactic effect against the development of new tumors, even in the immediate posttreatment period. Therefore, patients undergoing intravenous chemotherapy need to be monitored for the development of new tumor foci in both eyes before, during, and after the completion of intravenous chemotherapy.

Complications Short-term systemic side effects of chemotherapy are common, including fatigue, nausea, and vomiting, and hematologic problems such as leukopenia, thrombocytopenia, and anemia. Occasionally, patients require admission for

123

transfusions or work-ups for neutropenic fevers, but hematologic suppression rarely requires a delay of chemotherapy doses [10]. Ototoxicity has been reported in 0–25% of patients with therapeutic doses of carboplatin, with the risk being the highest in children below 6  months of age [31, 32]. Baseline hearing testing is encouraged in all patients [33]. Several ophthalmic complications have been reported in patients undergoing intravenous chemotherapy and focal therapy for retinoblastoma, although in general these cases are rare. Anagnoste reported three cases of rhegmatogenous retinal detachments, and Gombos reported a case of intraocular cholesterosis following intravenous chemotherapy [34, 35]. Parents should be warned, however, that iatrogenic iris injury is a remote possibility in eyes that undergo repeated laser treatments. A rare but potentially life-threatening complication of intravenous chemotherapy is the development of secondary nonocular cancers, particularly hematologic malignancies such as leukemia. Acute myelogenous leukemia (AML) has been reported following the use of etoposide with relatively short latency periods of 1–7 years [36]. In lymphoblastic leukemia patients undergoing chemotherapy, the long-term risk of AML has been reported to be 2–3% for intensive weekly or twice weekly schedules of teniposide or etoposide [36]. A survey by Gombos et  al. identified 12 cases of AML in retinoblastoma patients undergoing intravenous chemotherapy [37]. This survey included a questionnaire of retinoblastoma specialists practicing throughout the Americas and Europe, as well as a database of 1601 patients from the National Institutes of Health, the Department of Health and Human Services, and the Ophthalmic Oncology Service at the Memorial Sloan Kettering Cancer Center. Among the 12 identified cases, 9 patients had bilateral or multifocal retinoblastoma, and 8 patients had received an epipodophyllotoxin (etoposide or teniposide). Although a causative link between AML and epipodophyllotoxin therapy in retinoblastoma patients has not been established, it is concerning that prior to the intravenous chemotherapy era, the development of leukemia was thought to be a rare event. In a study published

124

J. W. Kim et al.

in 1984, Abramson et al. observed only one case of leukemia among 1900 survivors of retinoblastoma [37]. Without knowing the total number of retinoblastoma patients treated with intravenous chemotherapy in the modern era, it is not possible to calculate the risk of developing AML in this population of patients or even to conclude that a definite association exists. However, this report by Gombos et  al. suggests that further investigation will be required to fully assess the validity of this risk.

Periocular Chemotherapy (Injections and Exoplants) There is a great need to consistently achieve therapeutic levels of therapeutic agents on a regimen that is not limited by systemic toxicity. The chemotherapeutic agents currently used (carboplatin, etoposide, and vincristine) are small molecules and should enter the eye easily in an appropriate trans-scleral delivery system. In 1998, Mendelsohn and Abramson showed that peribulbar and episcleral injection of carboplatin could achieve higher vitreous concentrations than intravenous administration in primates [38]. a

In clinical practice, however, periocular chemotherapy injections have largely fallen out of favor due to poor clinical efficacy and significant local side effects. Carvalho and colleagues have described a promising closed trans-scleral delivery system that consists of a small, impermeable refillable silicone reservoir that can be firmly attached to the episclera with minimally invasive conjunctival surgery (Fig. 10.1) [39]. Once in place the reservoir can be filled and refilled as often as necessary by simple transconjunctival injections. These authors have demonstrated the superiority of this trans-scleral protected delivery system in delivering agents to the posterior vitreous and retina when directly compared to agent delivery via subtenon injection. In addition, much less delivered agent gains access to the plasma. As many as four reservoirs can be attached to the episclera of a single eye, allowing the concurrent delivery from multiple devices. The simplicity of the placement and recharging of the reservoir, the sustained delivery of high levels of agent to the vitreous and the posterior retina, and the potential for an inexpensive route for delivering tumor-­ targeted biotherapies make this type of trans-­scleral delivery very promising.

b

Fig. 10.1  Schematic of the eye (not to scale) that shows the positioning of the episcleral reservoir (a). This image shows a rigid reservoir held in place by scleral sutures. The most current version of the reservoir is made of flexible silicone. Indenting the reservoir creates a suction that securely attaches the implant to the sclera. Tissue adhe-

sive can also be used to assist in maintaining its position. A higher magnification of the implant (b). The round soft refill port can be palpated through the overlying conjunctiva for refilling of the reservoir with a small 30-gauge needle

125

10  Retinoblastoma: Treatment Options

Fig. 10.2  Periocular edema 2 days after bilateral periocular (subtenon injection of 10  mg/ml carboplatin × 2 sites). The edema resolved within a week with a short course of oral steroids without residual deficit

examination showed focal areas of ischemic necrosis and atrophy in the retrobulbar optic nerve along with dystrophic calcification and inflammation in the surrounding fibrovascular tissue. Mulvihill and colleagues reported ten patients with ocular motility restriction following subtenon carboplatin injection, diagnosed by forced duction testing [43]. They reported that subtenon carboplatin injection was associated with significant fibrosis of orbital soft tissues, restricting eye movement and making subsequent enucleation difficult. Because of the potential for local scarring and toxicity, most centers no longer routinely perform periocular carboplatin injections. In an attempt to reduce periocular complications of carboplatin, some authors have investigated the clinical use of topotecan and of fibrin sealant based upon favorable preclinical pharmacokinetic data [44–46].

Efficacy Murray and colleagues showed a dose-dependent inhibition of tumor growth with subconjunctivally delivered carboplatin in transgenic retinoblastoma mice [40]. The first clinical trial was performed by Abramson on 11 children with bilateral retinoblastoma, using a median of three injections per eye with an interval of 21  days between injections [41]. In that trial, a major clinical response was observed in three of five eyes with vitreous seeds and two of five eyes with retinal tumors. Periorbital edema and redness after injection were observed in 4 of 13 eyes, and 1 patient developed optic neuropathy (Fig. 10.2).

Complications Although the clinical experience with periocular injection has been encouraging, there have also been reports of local complications with this delivery method. Schmack and colleagues reported four cases of optic nerve atrophy in eyes that had been enucleated following periocular carboplatin injection [42]. The enucleated eyes had received between three and seven periocular carboplatin injections. Histopathologic

Selective Intra-arterial Chemotherapy (IAC) As early as 1953, Kupfer described a case of retinoblastoma treated with nitrogen mustard injected directly into the periocular circulation [47]. Later, in the 1960s and 1970s, Reese and Ellsworth combined external beam radiotherapy with intracarotid chemotherapeutic agents [48]. In the 1980s, Kaneko at the National Cancer Institute in Tokyo, Japan, began working on a new method to administer ocular chemotherapy – he described it as selective ophthalmic arterial infusion (SOAI) [49]. With this approach, developed primarily to avoid enucleation, a balloon catheter was inserted in the femoral artery, past the internal carotid, and guided just past the origin of the ophthalmic artery. The balloon was then inflated and melphalan injected into the arterial vasculature. Often adjuvant treatments were also administered, but more than half of the treated eyes were preserved. In 2008, Abramson and colleagues at Memorial Sloan Kettering Cancer Center (MSKCC) modified this technique with direct insertion of the cannula just past the ostium of the artery (Chap. 12) [50].

J. W. Kim et al.

126

Over the past 10 years, selective intra-arterial infusion of chemotherapy (IAC) has emerged as an important new modality for treating eyes with advanced intraocular retinoblastoma [15, 51–61]. IAC is currently the first choice for primary therapy for unilateral disease at most major centers, including MSKCC.  Doses used for IAC have ranged between 3.0 and 7.5  mg of melphalan per treatment in primary cases, and a multidrug regimen of carboplatin, melphalan, and topotecan has been used in salvage cases. Although the majority of treated patients with IAC have unilateral retinoblastoma, the Abramson group have recently published their results on treating bilateral patients with tandem, simultaneous IAC to both eyes. Francis et  al. showed eventfree globe survival of 91.3% in all treated eyes following tandem IAC, with no significant ocular or systemic side effects [62]. However, there are lingering concerns regarding the total dose of melphalan given to patients receiving tandem IAC as well as the cumulative radiation dose from fluoroscopy [63, 64], particularly given that bilateral patients tend to be younger and always have germinal RB1 mutations with a corresponding risk for second cancers.

Efficacy The initial phase I/II trial of ten patients with group V retinoblastoma salvaged seven eyes that would have otherwise been enucleated [50]. While the initial series used melphalan, additional followup reports have infused other agents including carboplatin and topotecan (alone or in combination) with good results. As mentioned, the technique has been used successfully in unilateral and bilateral cases and as both a primary and salvage approach. Follow-up electroretinogram (ERG) data suggests improved ERG findings in some very advanced cases with the resolution of the retinal detachment [65]. Defining an event as “enucleation or need for radiotherapy,” 4-year data from the Abramson group demonstrated an 81.7% event-free (no radiation or enucleation) survival for eyes that received intra-arterial chemotherapy as primary treatment and 58.4% for

eyes that had previous treatment failure with intravenous chemotherapy and/or external beam radiation therapy [55]. Recently in 2018, the Abramson group published their 10-year experience with IAC, showing event-free ocular survival rate estimate of 93% at 1 year when combined with intravitreal chemotherapy injections [66]. Approximately 25% of eyes initially treated with IAC in this series developed tumor recurrences, the majority within the first year of treatment. Despite the high success rates reported with IAC for intraocular retinoblastoma, close follow-up and multimodal treatment using focal consolidation and intravitreal chemotherapy injections are needed to optimize outcomes.

Complications Despite the encouraging clinical data for IAC, there is also growing evidence for potential ocular toxicity with this therapy, ranging from minor side effects (periocular edema, transient lash loss, forehead hyperemia) to more serious complications such as retinal artery occlusion and vitreous hemorrhage [57, 67–69]. Fortunately, neurologic complications related to the catheterization process appear to be extremely rare with this technique. There has been a single case of a delayed cerebral infarction following IAC in a child with an incomplete circle of Willis [70]. Systemic neutropenia has been reported in a minority of children with IAC [55]. Concern has also been raised regarding the clinical significance of low-­ dose radiation exposure from the fluoroscopy used during the IAC procedure [64, 71, 72]. Finally, IAC is not widely available even in some developed countries, and clinical success rates appear to vary between centers, perhaps related to the technical proficiency of the interventional neuroradiologist performing the procedure. When IAC is compared to intravenous chemotherapy as primary therapy, the benefit of avoiding systemic treatment in young children with retinoblastoma must be weighed against the higher risk of local complications and the complexity of the catheterization procedure. The latter risks are minimized but not completely reduced when

10  Retinoblastoma: Treatment Options

centers have accumulated years of experience in treating retinoblastoma with IAC. When considering the use of IAC as salvage therapy, the different set of potential side effects between IAC, enucleation, and EBR must be carefully weighed in any individual case, taking into account the patient age and whether the recurrent disease is unilateral or bilateral. For unilateral patients, the decision to enucleate for recurrent disease after primary therapy has failed is often weighed against using IAC for salvage therapy in a poorly seeing eye. For patients with bilateral retinoblastoma, IAC would typically be used before EBR for salvage therapy. In the current era, EBR is only used to salvage the second remaining eye after all other modalities have failed, and enucleation is the only remaining option. Finally, concern has been raised regarding the possibility of a slightly higher risk of extraocular relapse when using IAC to treat advanced intraocular disease compared to intravenous chemotherapy or enucleation [73]. In contrast to IAC, intravenous chemotherapy has the potential of providing systemic coverage of occult metastases in patients with Group D or E disease. Despite rare, anecdotal reports of extraocular relapse in children with advanced retinoblastoma undergoing attempted globe salvage with IAC [73], a recent survey of six retinoblastoma centers showed no increased risk of metastatic death from retinoblastoma following IAC [74]. However, clinicians should keep in mind that attempting to salvage an eye with advanced retinoblastoma rather than performing a primary enucleation always poses an inherently higher risk of tumor relapse and subsequent tumor dissemination, even if the increased risk is small and difficult to evaluate.

Intravitreal Chemotherapy Intravitreal chemotherapy (IVC) injection has reemerged as an effective, new modality to salvage eyes with residual vitreous seeding after systemic or intra-arterial chemotherapy (Chap. 12). For many years, the field of retinoblastoma management has avoided intraocular injection

127

due to widespread concerns that a needle entering an eye with active retinoblastoma would lead to the extraocular spread of cancer cells. There is histopathologic evidence of tumor cells in needle tracks of eyes with active retinoblastoma following fine needle aspiration biopsy, although documented cases of clinical extraocular relapse are rare [75, 76]. There are, however, documented cases of extraocular spread after the performance of vitrectomy (with positive-pressure infusion) in eyes with unsuspected retinoblastoma [77].

Efficacy Seregard initially reported on a series of three children with recurrent retinoblastoma being treated with IVC in 1995 [78]. Since then, there have been other reports of IVC being used successfully for children with active vitreous seeding. Suzuki and Kaneko reported on 237 eyes of 227 patients treated with 896 IVC injections of melphalan, with only a 0.4% rate of extraocular spread with a mean follow-up of 91 months [49, 79]. Then in 2012, Munier used 135 IVC injections of melphalan in 30 eyes of 30 children to safely treat patients who had failed systemic chemotherapy [80, 81]. In his series, IVC injections were given with a 32-gauge needle using several important safety measures including (1) ultrasound biomicroscopy (UBM) to rule out pars plana involvement at the site of injection, (2) preinjection paracentesis, and (3) postinjection cryotherapy at the site of injection [81]. Importantly, no child in the Munier series had evidence of extraocular spread during the period of follow-­up, and approximately 80% of eyes with vitreous seeding were salvaged [80–82]. Even with these precautions, repeated injections into the vitreous cavity do carry the potential for tumor spread, and it is important to strictly follow published protocols when using this technique. The majority of centers currently use 15–25  ug of melphalan per injection diluted at a concentration of 20 ug/0.1 ml, given every 1–2 weeks until clinical response has been achieved. The total number of injections typically ranges between 3 and 6 per eye, with more injections typically

J. W. Kim et al.

128

being needed for the clinical category of “cloud,” whereas fewer injections are needed to clear eyes with “dust” [83, 84]. The candidates for IVC melphalan injections are those patients with isolated vitreous seeding and minimal tumor load after chemotherapy and/ or radiation. Our experience in Los Angeles over the past 6 years with intravitreal injection of melphalan has been very encouraging with a high salvage rate for eyes with vitreous seeding of approximately 90%, with essentially no concern for extraocular spread if the protocol is strictly followed. Globe salvage rates in treating vitreous seeding have ranged between 90 and 100% in various series with IVC [82, 85, 86], with most failures occurring due to retinal tumor recurrences. Intravitreal melphalan injections do not appear to be effective for retina-based tumors, although a recent paper by Abramson et al. demonstrates some efficacy when treating both subretinal seeds and retinal tumors [87]. However this particular series also used other modalities such as IAC and focal therapy, and therefore definite conclusions are difficult to draw. Because of serious long-term side effects, external beam radiotherapy is now reserved for patients who have failed IVC, especially in children younger than 12–18 months of age (see section on EBR).

Complications To avoid complications with this technique, it is important to carefully follow the protocol published by Munier [80, 81]. By using small 32- or 33-gauge needles, performing a paracentesis, and applying cryotherapy at the site of injection, the risk of extraocular spread should be remote. A recent retrospective cohort study on the safety of intravitreal chemotherapy injections performed for retinoblastoma at 10 retinoblastoma centers worldwide showed no cases of extraocular extension following 3553 intravitreous injections [88]. In our experience, peripheral chorioretinal atrophy commonly occurs at the site of injection and does not appear to affect visual function. However, we have also observed rare cases of profound retinal toxicity, presenting as an

acute hemorrhagic retinopathy involving the entire posterior pole with a resultant flat electroretinogram [89]. There have been some theories on how this complication occurs, such as inadvertently injecting melphalan into the retrohyaloid space in the presence of a posterior vitreous detachment. However, there does not appear to be any identifiable correlation with the individual or cumulative dose of melphalan, presence of posterior vitreous detachment, or other clinical factors when this complication is observed. The current recommendation is to inject into the center of the vitreous cavity to avoid the possibility of a retrohyaloid injection. Conversely, injecting too close to the lens can lead to anterior segment complications such as cataract and iris atrophy. Finally, rare cases of severe ocular phthisis have been reported with IVC, and the risk seems to be correlated with higher doses of melphalan [90]. Although more investigation is needed, intravitreal chemotherapy appears to be an effective option for salvaging selected eyes with isolated vitreous seeding as success rates appear to be higher than any other modality and the safety profile is more favorable than with EBR. However, parents should be aware of the possibility of rare ocular complications with IVC.

Laser Therapy Laser therapy is used for the following indications in the management of intraocular retinoblastoma: (1) for accomplishing consolidation of large tumors after systemic chemotherapy (i.e., chemoreduction), (2) for treating small peripheral or posterior tumors as the sole modality, and (3) for eradicating small tumor recurrences within or adjacent to scars following chemotherapy and/or radiotherapy (Chap. 10). When used in conjunction with primary chemotherapy for intraocular retinoblastoma, focal consolidation can be accomplished with either the green 532 nm argon or the 810 nm diode infrared laser. The shorter wavelength green 532 nm argon laser is more readily absorbed in the relatively nonpigmented retinoblastoma tumor, while the longer wavelength 810 nm diode infrared laser achieves

10  Retinoblastoma: Treatment Options

deeper penetration in the presence of intact retinal pigment epithelium. The technique we find useful with the argon 532  nm is essentially the same for both primary treatment of group A lesions and consolidation following primary chemotherapy in groups B–D. In general, focal consolidation begins after the first or second cycle of systemic chemotherapy after the tumor volume has been reduced.

Treatment Parameters The goal of the therapy is to completely cover each lesion with 30% overlap during at least three different sessions. We choose initial power settings of 250–300  mW, with durations of 300–500  ms. The power and time settings are kept low to prevent tumor disruption and hemorrhage that may be associated with excessive energy delivery. The first burns are placed at the edge of the lesion with the spot half on and half off the tumor. The power and/ or duration can be adjusted to achieve gentle whitening of the tumor. We do not recommend exceeding 500–600  mW and 700  ms with the 532  mm laser. Once the lesion is outlined, the entire lesion including any type I regression-associated calcium is covered with overlapping rows of burns. Larger lesions undergoing chemoreduction may require 200–400 burns for good coverage. The burns over the thicker areas of the tumor may be virtually invisible compared with those placed at the edge of the lesion. The power or duration should not be increased to compensate for the decreased absorption over the thicker parts of the lesion. Repeat the laser coverage at 2–4 week intervals during and/or after the administration of systemic chemotherapy until the entire lesion has been covered on at least three different occasions. Because the infrared 810 nm diode laser has a longer wavelength than the argon laser, it penetrates deeper and is absorbed mainly by the retinal pigment epithelium. Therefore, it is particularly useful if retinal pigmented epithelium (RPE) is intact under the lesion to be treated. Another advantage of the diode laser is that its larger spot size allows more rapid coverage of the lesion and a lower risk of delivering excessive energy that

129

might cause bleeding or tumor disruption. The diode laser can also be used for photocoagulation or transpupillary thermotherapy, depending on the settings utilized (Chap. 10). The endpoint of energy application is, like that for the argon laser, a gentle whitening of a spot placed half on and half off the tumor. Because of the larger spot size, the power is generally set initially at between 300 and 500 mW for 500 ms. The power can be adjusted upward to 700–800  mW if required to achieve the desired endpoint. If the active tumor focus demonstrates growth after first session of laser treatment, a second application should be attempted at a high-­power level. Persistent growth after the second laser session is an indication that another modality will be needed to eradicate the tumor.

Efficacy Laser photocoagulation is an appropriate method of management in cases where the tumor is located posteriorly, the media are clear, and the tumor is 3.0 mm or less in diameter and 2.0 mm or less in thickness without seeding into the adjacent vitreous. In a series of 188 tumors that had mean tumor diameter of 3.0  mm and thickness of 2.0  mm, tumor regression was achieved in 86% with a recurrence rate of 14% [91]. Using the diode laser on a continuous mode with 45–60 sec treatment duration (i.e., transpupillary thermotherapy or TTT), Abramson was able to achieve complete regression in 84 out of 91 tumors (92%) [92]. Larger tumors are at greater risk for complications such as focal iris atrophy and focal paraxial lens opacity because they require more intense therapy as compared to smaller tumors. Abramson and others have also reported success in treating recalcitrant retinal tumors by giving patients intravenous indocyanine green (ICG) to increase absorption of the 810 nm laser [93–95]. ICG infusion 1  minute prior to applying TTT may be considered for patients with tumors with suboptimal response to standard TTT, recurrence after standard TTT, or minimally pigmented fundus with poor standard TTT uptake [96].

J. W. Kim et al.

130

Complications Complications of focal laser consolidation include burns of the iris at the pupillary margin and focal lens opacities, both of which are rare in experienced hands. Other complications that are associated with excessive energy delivered to the tumor include subhyaloid and vitreous hemorrhage. Theoretically, it is possible to mechanically disrupt the tumor and create vitreous seeding of the tumor by using excessive energy (power Å ~ time) levels but that complication has also been a very rare event in patient care if the above cautions are exercised. In approximately 1000 lesions in more than 300 eyes treated in Los Angeles, we have seen tumor disruption by the laser on only one tumor. In that case, early in the series, treatment was done before chemotherapy was given, and the laser power was increased to approximately 900 mW. Vitreous hemorrhage and tumor seeding ultimately resulted in loss of the eye. The most significant long-term complication of focal consolidation is decreased vision from RPE scar migration or “creep” in lesions near the foveola. A judicious approach is required when applying laser on the foveal side of a tumor near fixation to minimize the resulting scotoma. Lee and colleagues demonstrated an increase in the size of laser scars following red diode laser application [97]. It is reasonable to consider close observation after sufficient primary chemotherapy of a small tumor located near the fovea until documented growth is seen. In some instances, central vision can be spared if regrowth does not occur. Tumors that exist in the maculopapular bundle can be managed in this fashion, especially if the contralateral eye has been enucleated or has poor visual prognosis.

Cryotherapy Destruction of retinoblastoma tumors by cryotherapy results from disruption of cellular membranes following the freeze–thaw cycle. It can also have a local vaso-occlusive effect on the tumor and nearby retina/choroid. Cryotherapy is useful for small peripheral tumors and can be used successfully for lesions up to 3.0 mm in diameter and 2.0 mm in thickness (Chap. 10). Cryotherapy can also be used

to eradicate small tumors with localized vitreous seeding near the apex, assuming that the ice ball from the treatment can encompass both the tumor and the seeds. The difficulty in using cryotherapy as local treatment for posterior pole tumors is that a surgical procedure is required to open the conjunctiva so that accurate placement of the probe can be achieved. In addition, because the probe tip cannot be visualized while the freezing is taking place, it is theoretically possible to freeze the macula or optic nerve. An important consideration to keep in mind is that cryotherapy routinely destroys a great deal of normal retina surrounding the lesion, thereby increasing the visual deficit from the resulting chorioretinal scar. Therefore, the location and size of the tumor are important considerations when using cryotherapy for retinoblastoma.

Treatment Parameters The treatment begins by confirming that the cryotherapy probe and foot pedal are functioning properly. Using indirect ophthalmoscopy, the probe of the cryotherapy unit is used to localize and elevate the tumor on the tip of the probe. Once the probe is directly beneath the tumor, freezing is initiated and the ice ball maintained until it encompasses the entire tumor mass. After the treatment covers the apex of the tumor for 2 mm, the ice ball is allowed to thaw, and this freeze–thaw cycle is repeated for a total of two or three applications. To avoid iatrogenic injury to the globe, it is important not to move the probe on the sclera until the ice ball has completely resolved.

Efficacy Cryotherapy is indicated for anteriorly located tumors with clear media, and the highest success rate is achieved for primary, small tumors without seeding. Proper patient selection and utilization of careful technique are important factors in achieving a high success rate. In a series of 138 tumors treated with cryotherapy by Abramson, 70% of tumors overall were cured with cryotherapy [98]. For primary tumors without previ-

10  Retinoblastoma: Treatment Options

ous therapy, the cure rate was 95%, but all of the tumors at the vitreous base with seeding failed.

Complications Complications of cryotherapy include vitreous hemorrhage, development of subretinal fluid, and retinal holes. Very rarely, retinal breaks from cryotherapy can result from a combination of the atrophic retina and vitreous traction, particularly at the edges of the treated area. We have observed several cases of rhegmatogenous retinal detachment when large superior, partially calcified tumors were treated extensively with cryotherapy. Extensive cryotherapy can also cause atrophy of the sclera, with formation of a pseudocoloboma of the sclera. The presence of preexisting subretinal fluid in the region of proposed cryotherapy is a relative contraindication. The use of proper technique is critical in avoiding these complications, and it is particularly important to not move the cryotherapy probe until there is visual confirmation through the indirect ophthalmoscope that the ice ball has completely dissipated.

131

outcome, although in most cases the tumor has already destroyed central vision. Diffuse vitreous or subretinal seeding will not respond to brachytherapy, although it may be possible to treat the distant seeding with other modalities such as intravitreal injection.

Treatment Parameters Iodine-125 is currently the most commonly used isotope for brachytherapy in the United States. Other source materials such as Ruthenium-106 have been used successfully in Europe [99]. When creating an Iodine-125 plaque for a child with retinoblastoma, radioactive seeds are placed into a custom-built plaque designed to treat the specific shape and size of the tumor. Plaque placement is confirmed with the indirect ophthalmoscope, and the active plaque is inserted in the operating room. The regression response most commonly seen after removal is a type 4 pattern (flat scar). With Iodine plaques, the radiation dose is 4000–4500 cGy to the apex of the tumor at a rate of 50–150 cGy/h. The plaque is removed in a second operation 2–3 days later, depending on the isotope used and the size of the tumor.

Brachytherapy In current treatment regimens for intraocular retinoblastoma, episcleral brachytherapy is a treatment option for focal tumors that are too large for cryotherapy or laser treatment (Chap. 10). Unlike external beam radiotherapy, radiation exposure is limited to the ocular structures, and there is no increased risk of second nonocular cancers or orbital hypoplasia. As is the case with other modalities, proper selection of patients is critical for success with brachytherapy. The ideal candidate for a radioactive plaque is a patient with a focal tumor (8 mm or less in thickness), without vitreous or subretinal seeds and more than 2 disc diameters away from the macula or optic nerve. Tumors with localized seeding (100  Gy) [99]. Brachytherapy is also effective as a salvage technique in eyes that have failed other types of therapy including external beam radiation, photocoagulation, or cryotherapy, as long as the seeding is absent or limited. Used as salvage therapy for eyes that have failed other treatment methods, Abramson reported an overall success rate for brachytherapy of 50%, utilizing

J. W. Kim et al.

132

cobalt plaques [101]. Merchant and colleagues recently reported a salvage rate of 60% in eyes that had failed chemotherapy or external beam radiotherapy [102]. Risks for tumor recurrence following brachytherapy include the presence of tumor seeds in the vitreous and subretinal space, large tumor size, prior failure of external beam radiation, lower dose of radiation (30  Gy using lenssparing techniques experienced some facial asymmetry [108]. Although long-term data are not available, it is hoped that conformal radiation techniques such as intensity-modulated radiotherapy (IMRT) will decrease the severity of orbital and facial hypoplasia. Less serious complications which have been reported with EBR include keratitis sicca, corneal ulceration, keratinization of the conjunctiva and sclera, lacrimal gland atrophy/fibrosis, loss of lashes, fat atrophy in the orbit, and prolonged skin erythema within the area of the radiation portal. Severe keratitis sicca is very common in the first 3 months after treatment, and we recommend performing prophylactic silicone punctual plug placement in all children undergoing EBR to reduce photophobia and ocular discomfort.

Enucleation Despite the progress of conservative modalities, enucleation remains the most commonly employed technique for treating retinoblastoma worldwide (Chap. 15). Retinoblastoma typically responds, at least partially, to all eye-conserving treatment modalities. However, tumor regrowth is a common cause of treatment failure, necessitating constant surveillance and monitoring. When all conservative strategies have failed, enucleation is typically curative unless the tumor extends to the optic nerve margin or invades the sclera. The vast majority of retinoblastoma cases are sporadic (nonfamilial), and many children do not present for medical care until the eye is filled with tumor, causing leukocoria, strabismus, or glaucoma. Typically these eyes have very limited visual potential, even with aggressive treatment. If the other eye is not involved or can be treated

J. W. Kim et al.

with focal therapies, there is little reason to subject the patient to the toxicities of systemic chemotherapy or external beam radiation. Patients considered for enucleation are those with unilateral group D disease, unilateral or bilateral group E disease, and any patient with active tumor following the completion of primary therapy in a blind eye. Patients are also considered for enucleation if the eye contains suspected active tumor and cannot be followed with fundoscopy due to obscured media. Greater than 95% of patients with unilateral retinoblastoma without extraocular disease are cured by enucleation, a rare situation in surgical oncology [117]. The decision to enucleate an eye with retinoblastoma should be made in consultation with the family, and several key issues should be discussed. First, it should be emphasized that the eye has not had useful vision for a prolonged period and the child will not experience any functional limitations from enucleation. Second, the operation is not painful and can usually be performed on an outpatient basis. Finally, the family should understand that enucleation is being considered because tumor control cannot be accomplished with any of the available modalities and that the risk of keeping a blind eye cannot be ­justified when there is a risk for tumor spread and metastasis. Critical elements of the surgery include avoiding any perforations of the globe and obtaining a long section of optic nerve of at least 15 mm. Different techniques have been described for obtaining a long section of optic nerve during enucleation. Although most experienced ocular oncologists routinely obtain 15–20 mm of optic nerve with enucleation, one of the newer techniques is to sever the optic nerve under direct visualization through a superior orbital approach, utilizing a small upper lid incision [118]. Shrinkage of the optic nerve segment typically occurs with processing, and this should be kept in mind when evaluating the results of different surgical techniques [119]. A variety of orbital implants are available to reestablish the orbital volume, including silicone, hydroxyapatite, Medpor, and dermis fat graft. When considering implant choices, the silicone sphere is widely available, has the lowest

135

10  Retinoblastoma: Treatment Options

incidence of complications, and provides acceptable motility. Porous implants such as hydroxyapatite and high-density polyethylene (Medpor) have gained in popularity due to the low rates of implant migration and the potential for better motility if the implant is pegged to allow coupling with the prosthesis. However, no study has demonstrated a motility advantage for non-pegged porous implants (hydroxyapatite, Medpor) when compared to nonporous implants (silicone). In addition, porous orbital implants have higher rates of implant exposure and infection compared to silicone spheres, as well as higher costs [120]. No matter which implant is chosen, the largest implant that can be fit into the orbit should be selected (16–18  mm), both to encourage orbital growth and to obviate the need to place a secondary implant when the child grows. Postoperative infections and other complications are extremely rare with modern surgical techniques. After 4 weeks, patients can be fitted with a prosthesis by the ocularist. Continued monitoring of the child will be necessary in the postoperative period to detect orbital tumor recurrence in the socket, which has a high correlation with systemic metastatic disease [121].

Emerging Therapies There are several emerging therapies for intraocular retinoblastoma on the horizon. Using pars plana vitrectomy (PPV) and endoresection to treat refractory retinal and/or vitreous disease was introduced by Zhao and colleagues in 2018, and the initial clinical series showed that 18 of 21 eyes were salvaged with no cases of extraocular relapse [122]. This technique involves a PPV, endoresection to remove the retinal recurrence, silicone oil instillation, low-dose melphalan in the infusion line, and both subconjunctival and intravitreal melphalan injections following PPV to minimize the risk of tumor dissemination. The technique is explained in more detail in Chap. 18. Photodynamic therapy performed through an indirect delivery system was successfully performed in a rabbit model and may offer a new modality for patients with retinal tumors too large for laser photocoagulation [123].

Finally, the episcleral chemotherapy reservoir was recently used for the first time in a patient with multiple retinal recurrences (personal communication – Gallie B). In the future we expect to evaluate more clinical results from patients treated with these new modalities to learn how they may fit into the treatment algorithm.

Conclusion Strategies for treating intraocular retinoblastoma continue to evolve as new therapies are developed and others fall out of favor. The popularity of chemotherapy during the past decade has spared many young children with retinoblastoma the side effects of external beam radiation. The emergence of local therapies over the past 5  years has improved globe salvage rates while reducing systemic side effects. Intra-arterial chemotherapy has become more widely adopted and is now utilized as primary therapy for advanced unilateral disease at major centers. Intravitreal chemotherapy has revolutionized the treatment of active vitreous seeding, and the use of EBR has dramatically fallen out of favor. Despite this success, there continue to be significant challenges in improving visual outcomes and globe salvage rates in patients with retinoblastoma. Modern centers treating retinoblastoma continue to manage patients with a variety of modalities, individualizing the therapy according to the patient’s presentation and clinical course.

References 1. Novakovic B. U.S. childhood cancer survival, 1973– 1987. Med Pediatr Oncol. 1994;23(6):480–6. 2. Sant M, Capocaccia R, Badioni V, UROCARE Working Group. Survival for retinoblastoma in Europe. Eur J Cancer. 2001;37(6):730–5. 3. Abramson DH, Schefler AC.  Update on retinoblastoma. Retina. 2004;24(6):828–48. 4. Kim JW, Abramson DH, Dunkel IJ. Current management strategies for intraocular retinoblastoma. Drugs. 2007;67(15):2173–85. 5. Murphree AL.  Intraocular retinoblastoma: a case for a new group classification. Ophthalmol Clin N Am. 2005;18:41–53.

136 6. Berry JL, Jubran R, Kim JW, et al. Long-term outcomes of group D eyes in bilateral retinoblastoma patients treated with chemoreduction and low-dose IMRT salvage. Pediatr Blood Cancer. 2013;60(4):688–93. 7. Kaliki S, Shields CL, Rojanaporn D, et al. High-risk retinoblastoma based on international classification of retinoblastoma: analysis of 519 enucleated eyes. Ophthalmology. 2013;120(5):997–1003. 8. Armenian SH, Panigrahy A, Murphree AL, et  al. Management of retinoblastoma with proximal optic nerve enhancement on MRI at diagnosis. Pediatr Blood Cancer. 2008;51(4):479–84. 9. Mallipatna A. Retinoblastoma. In: Edge SB, American Joint Committee on Cancer, editors. AJCC cancer staging manual. 8th ed. New  York City: Springer; 2017. p. 819–31. 10. Gallie BL, Budning A, DeBoer G, et al. Chemotherapy with focal therapy can cure intraocular retinoblastoma without radiotherapy. Arch Ophthalmol. 1996;114(11):1321–8. 11. Beck MN, Balmer A, Dessing C, et al. First-line chemotherapy with local treatment can prevent external-beam irradiation and enucleation in lowstage intraocular retinoblastoma. J Clin Oncol. 2000;18(15):2881–7. 12. Bornfeld N, Schüler A, Bechrakis N, et al. Preliminary results of primary chemotherapy in retinoblastoma. Klin Padiatr. 1997;209(4):216–21. 13. Brichard B, De Bruycker JJ, De Potter P, et al. Combined chemotherapy and local treatment in the management of intraocular retinoblastoma. Med Pediatr Oncol. 2002;38(6):411–5. 14. Gunduz K, Shields C, Shields JA, et al. The outcome of chemoreduction treatment in patients with ReeseEllsworth group V retinoblastoma. Arch Ophthalmol. 1998;116(12):1613–7. 15. Friedman DL, Himelstein B, Shields CL, et al. Chemoreduction and local ophthalmic therapy for intraocular retinoblastoma. J Clin Oncol. 2000;18(1):12–7. 16. Doz F, Khelfaoui F, Mosseri V, et al. The role of chemotherapy in orbital involvement of retinoblastoma. The experience of a single institution with 33 patients. Cancer. 1994;74(2):722–32. 17. Goble RR, McKenzie J, Kingston JE, et al. Orbital recurrence of retinoblastoma successfully treated by combined therapy. Br J Ophthalmol. 1990;74(2):97–8. 18. Murphree AL, Villablanca JG, Deegan WF 3rd, et al. Chemotherapy plus local treatment in the management of intraocular retinoblastoma. Arch Ophthalmol. 1996;114(11):1348–56. 19. Greenwald MJ, Goldman S, Strauss LC. Combined chemoreduction and adjuvant treatment for intraocular retinoblastoma. Ophthalmology. 1998;105(9):1579–81. 20. Schiavetti A, Hadjistilianou T, Clerico A, et al. Conservative therapy in intraocular retinoblastoma: response/recurrence rate. J Pediatr Hematol Oncol. 2005;27(1):3–6. 21. Wilson MW, Rodriguez-Galindo C, Haik BG, et  al. Multiagent chemotherapy as neoadjuvant treatment for

J. W. Kim et al. multifocal intraocular retinoblastoma. Ophthalmology. 2001;108(11):2106–14; discussion 2114–5. 22. Rodriguez-Galindo C, Wilson MW, Haik BG, et al. Treatment of intraocular retinoblastoma with vincristine and carboplatin. J Clin Oncol. 2003;21(10):2019–25. 23. Lee TC, Hayashi N, Dunkel IJ, et al.  New reti noblastoma tumor formation in children initially treated with systemic carboplatin. Ophthalmology. 2003;110(10):1989–94; discussion 1994–5. 24. Abramson DH, Lawrence S, Beaverson KL, et al. Systemic carboplatin for retinoblastoma: change in tumour size over time. Br J Ophthalmol. 2005;89(12):1616–9. 25. Chan HS, Canton M, Gallie BL.  Chemosensitivity and multidrug resistance to antineoplastic drugs in retinoblastoma cell lines. Anticancer Research. 1989;9(2):469–74. 26. Chan HS, DeBoer G, Thiessen JJ, et al. Combining cyclosporin with chemotherapy controls intraocular retinoblastoma without requiring radiation. Clin Cancer Res. 1996;2(9):1499–508. 27. Kingston JE, Hungerford J, Madreperla SA, et al. Results of combined chemotherapy and radiotherapy for advanced intraocular retinoblastoma. Arch Ophthalmol. 1996;114(11):1339–43. 28. Shields CL, Shields J, Needle M, et al. Combined chemoreduction and adjuvant treatment for intraocular retinoblastoma. Ophthalmology. 1997;104(12):2101–11. 29. Shields CL, Mashayekhi A, Cater J, et  al. Chemoreduction for retinoblastoma: analysis of tumor control and risks for recurrence in 457 tumors. Trans Am Ophthalmol Soc. 2004;102:35–44; discussion 44–5. 30. Shields CL, De Potter P, Himelstein BP, et al. Chemoreduction in the initial management of intraocular retinoblastoma. Arch Ophthalmol. 1996;114(11): 1330–8. 31. Qaddoumi I, Bass JK, Wu J, et  al. Carboplatin associated ototoxicity in children with retinoblastoma. J Clin Oncol. 2012;30(10):1034–41. 32. Soliman SE, D’Silva CN, Dimaras H, et al. Clinical and genetic associations for carboplatin-related ototoxicity in children treated for retinoblastoma: a retrospective noncomparative single-institute experience. Pediatr Blood Cancer. 2018;65(5):e26931. 33. Smits C, Swen SJ, Goverts ST, et al. Assessment of hearing in very young children receiving carboplatin for retinoblastoma. Eur J Cancer. 2006;42(4):492–500. 34. Anagnoste SR, Scott IU, Murray TG et  al. Rhegmatogenous retinal detachment in retinoblastoma patients undergoing chemoreduction and cryotherapy. Am J Ophthalmol. 2000;129(6):817–9. 35. Gombos DS, Howes E, O'Brien JM.  Cholesterosis following chemoreduction for advanced retinoblastoma. Arch Ophthalmol. 2000;118(3):440–1. 36. Pui CH, Ribeiro RC, Hancock ML, et  al. Acute myeloid leukemia in children treated with epipodophyllotoxins for acute lymphoblastic leukemia. N Engl J Med. 1991;325(24):1682–7.

10  Retinoblastoma: Treatment Options 37. Abramson DH, Ellsworth RM, Kitchin FD, et  al. Second nonocular tumors in retinoblastoma survivors. Are they radiation-­ induced? Ophthalmology. 1984;91(11):1351–5. 38. Mendelsohn ME, Abramson DH, Madden T, et  al. Intraocular concentrations of chemotherapeutic agents after systemic or local administration. Arch Ophthalmol. 1998;116(9):1209–12. 39. Pontes de Carvalho RA, Krausse ML, Murphree AL, et al. Delivery from episcleral exoplants. Invest Ophthalmol Vis Sci. 2006;47(10):4532–9. 40. Murray TG, Cicciarelli N, O’brien JM, et  al. Subconjunctival carboplatin therapy and cryotherapy in the treatment of transgenic murine retinoblastoma. Arch Ophthalmol. 1997;115(10):1286–90. 41. Abramson DH, Frank CM, Dunkel IJ. A phase I/II study of subconjunctival carboplatin for intraocular retinoblastoma. Ophthalmology. 1999;106(10):1947–50. 42. Schmack I, Hubbard GB, Kang SJ, et  al. Ischemic necrosis and atrophy of the optic nerve after periocular carboplatin injection for intraocular retinoblastoma. Am J Ophthalmol. 2006;142(2):310–5. 43. Mulvihill A, Budning A, Jay V, et  al. Ocular motility changes after subtenon carboplatin chemotherapy for retinoblastoma. Arch Ophthalmol. 2003;121(8):1120–4. 44. Van Quill KR, Dioguardi PK, Tong CT, et  al. Subconjunctival carboplatin in fibrin sealant in the treatment of transgenic murine retinoblastoma. Ophthalmology. 2005;112(6):1151–8. 45. Tsui JY, Dalgard C, Van Quill KR, et  al. Subconjunctival topotecan in fibrin sealant in the treatment of transgenic murine retinoblastoma. Invest Ophthalmol Vis Sci. 2008;49(2):490–6. 46. Yousef YA, Halliday W, Chan HS, et  al. No ocular motility complications after subtenon topotecan with fibrin sealant for retinoblastoma. J Can Ophtalmol. 2013;48(6):524–8. 47. Kupfer C.  Retinoblastoma treated with intravenous nitrogen mustard. Am J Ophthalmol. 1953;36(12):1721–3. 48. Hyman GA, Ellsworth RM, Feind CR, et  al. Combination therapy in retinoblastoma. A 15-year summary of methods and results. Arch Ophthalmol. 1968;80(6):744–6. 49. Kaneko A, Suzuki S.  Eye-preservation treatment of retinoblastoma with vitreous seeding. Jpn J Clin Oncol. 2003;33(12):601–7. 50. Abramson DH, Dunkel IJ, Brodie SE, et  al. A phase I/II study of direct intraarterial (ophthalmic artery) chemotherapy with melphalan for intraocular retinoblastoma initial results. Ophthalmology. 2008;115(8):1398–404, 1404 e1. 51. Thampi S, Hetts SW, Cooke DL, et al. Superselective intra-arterial melphalan therapy for newly diagnosed and refractory retinoblastoma: results from a single institution. Clin Ophthalmol. 2013;7:981–9. 52. Abramson DH, Marr BP, Dunkel IJ, et al. Intra-arterial chemotherapy for retinoblastoma. Ophthalmology. 2012;119(8):1720–1; author reply 1721.

137 53. Abramson DH, Marr BP, Dunkel IJ, et  al. Intra arterial chemotherapy for retinoblastoma in eyes with vitreous and/or subretinal seeding: 2-year results. Br J Ophthalmol. 2012;96(4):499–502. 54. Francis JH, Gobin YP, Brodie SE, et  al. Experience of intra-arterial chemosurgery with single agent carboplatin for retinoblastoma. Br J Ophthalmol. 2012;96(9):1270–1. 55. Gobin YP, Dunkel IJ, Marr BP, et  al. Intra-arterial chemotherapy for the management of retinoblastoma: four-year experience. Arch Ophthalmol. 2011;129(6):732–7. 56. Gobin YP, Dunkel IJ, Marr BP, et  al. Combined, sequential intravenous and intra-arterial chemotherapy (bridge chemotherapy) for young infants with retinoblastoma. PLoS One. 2012;7(9):e44322. 57. Shields CL, Bianciotto CG, Jabbour P, et  al. Intraarterial chemotherapy for retinoblastoma: report no. 2, treatment complications. Arch Ophthalmol. 2011;129(11):1407–15. 58. Shields CL, Kaliki S, Rojanaporn D, et al. Intravenous and intra-arterial chemotherapy for retinoblastoma: what have we learned? Curr Opin Ophthalmol. 2012;23(3):202–9. 59. Shields CL, Kaliki S, Al-Dahmash S, et  al. Management of advanced retinoblastoma with intravenous chemotherapy then intra-­ arterial chemotherapy as alternative to enucleation. Retina. 2013;33(10):2103–9. 60. Ni N, Shields CL, Bianciotto CG, et  al. Complete regression of retinoblastoma following intra-arterial chemotherapy. J Pediatr Ophthalmol Strabismus. 2011;48 Online:e23–5. 61. Shields CL, Bianciotto CG, Jabbour P, et  al. Intraarterial chemotherapy for retinoblastoma: report no. 1, control of retinal tumors, subretinal seeds, and vitreous seeds. Arch Ophthalmol. 2011;129(11):1399–406. 62. Francis JH, Roosipu N, Levin AM, et al. Current treatment of bilateral retinoblastoma: the impact of intraarterial and intravitreous chemotherapy. Neoplasia. 2018;20(8):757–63. 63. Gobin YP, Rosenstein LM, Marr BP, et al. Radiation exposure during intra-­arterial chemotherapy for retinoblastoma. Arch Ophthalmol. 2012;130(3):403–4; author reply 404–5. 64. Vijayakrishnan R, Shields CL, Ramasubramanian A, et al. Irradiation toxic effects during intra-arterial chemotherapy for retinoblastoma: should we be concerned? Arch Ophthalmol. 2010;128(11):1427–31. 65. Brodie SE, Munier FL, Francis JH, et al. Persistence of retinal function after intravitreal melphalan injection for retinoblastoma. Doc Ophthalmol. 2013;126(1):79–84. 66. Francis JH, Levin AM, Zabor EC, et  al. Ten-year experience with ophthalmic artery chemosurgery: ocular and recurrence-free survival. PLoS One. 2018;13(5):e0197081. 67. Marr B, Gobin PY, Dunkel IJ, et  al. Spontaneously resolving periocular erythema and ciliary madarosis following intra-­ arterial chemotherapy for

138 retinoblastoma. Middle East Afr J Ophthalmol. 2010;17(3):207–9. 68. Abramson DH, Marr BP, Brodie SE, et al. Intraocular hemorrhage after intra-arterial chemotherapy for retinoblastoma in sickle cell trait. Open Ophthalmol J. 2012;6:1–3. 69. Munier FL, Beck-Popovic M, Balmer A, et  al. Occurrence of sectoral choroidal occlusive vasculopathy and retinal arteriolar embolization after superselective ophthalmic artery chemotherapy for advanced intraocular retinoblastoma. Retina. 2011;31(3):566–73. 70. De la Huerta I, Seider MI, Hetts SW, et al. Delayed cerebral infarction following intra-arterial chemotherapy for retinoblastoma. JAMA Ophthalmol. 2016;134(6):712–4. 71. Shields CL, Fulco EM, Arias JD, et al. Retinoblastoma frontiers with intravenous, intra-arterial, periocular, and intravitreal chemotherapy. Eye (Lond). 2013;27(2):253–64. 72. Shields CL, Shields JA.  Intra-arterial chemotherapy for retinoblastoma: the beginning of a long journey. Clin Exp Ophthalmol. 2010;38(6):638–43. 73. Yousef YA, Soliman SE, Astudillo PP, et  al. Intraarterial chemotherapy for retinoblastoma: a systematic review. JAMA Ophthalmol. 2016;134:584. 74. Abramson DH, Shields CL, Jabbour P, et al. Metastatic deaths in retinoblastoma patients treated with intraarterial chemotherapy (ophthalmic artery chemosurgery) worldwide. Int J Retina Vitreous. 2017;3:40. 75. Karcioglu ZA, Gordon RA, Karcioglu GL.  Tumor seeding in ocular fine needle aspiration biopsy. Ophthalmology. 1985;92(12):1763–7. 76. Karcioglu ZA. Fine needle aspiration biopsy (FNAB) for retinoblastoma. Retina. 2002;22(6):707–10. 77. Shields CL, Honavar S, Shields JA, et al. Vitrectomy in eyes with unsuspected retinoblastoma. Ophthalmology. 2000;107(12):2250–5. 78. Seregard S, Kock E, af Trampe E. Intravitreal chemotherapy for recurrent retinoblastoma in an only eye. Br J Ophthalmol. 1995;79(2):194–5. 79. Suzuki S, Kaneko A. Management of intraocular retinoblastoma and ocular prognosis. Int J Clin Oncol. 2004;9(1):1–6. 80. Munier FL, Gaillard MC, Balmer A, et al. Intravitreal chemotherapy for vitreous disease in retinoblastoma revisited: from prohibition to conditional indications. Br J Ophthalmol. 2012;96(8):1078–83. 81. Munier FL, Soliman S, Moulin AP, et  al. Profiling safety of intravitreal injections for retinoblastoma using an anti-reflux procedure and sterilisation of the needle track. Br J Ophthalmol. 2012;96(8):1084–7. 82. Munier FL, Gaillard MC, Balmer A, et al. Intravitreal chemotherapy for vitreous seeding in retinoblastoma: recent advances and perspectives. Saudi J Ophthalmol. 2013;27(3):147–50. 83. Francis JH, Abramson DH, Gaillard MC, et  al. The classification of vitreous seeds in retinoblastoma and response to intravitreal melphalan. Ophthalmology. 2015;122(6):1173–9.

J. W. Kim et al. 84. Francis JH, Marr BP, Abramson DH. Classification of vitreous seeds in retinoblastoma: correlations with patient, tumor, and treatment characteristics. Ophthalmology. 2016;123(7):1601–5. 85. Francis JH, Brodie SE, Marr B, et al. Efficacy and toxicity of Intravitreous chemotherapy for retinoblastoma: four-year experience. Ophthalmology. 2017;124(4):488–95. 86. Shields CL, Douglass AM, Beggache M, et  al. Intravitreous chemotherapy for active vitreous seeding from retinoblastoma: outcomes after 192 consecutive injections. The 2015 Howard Naquin Lecture. Retina. 2016;36(6):1184–90. 87. Abramson DH, Ji X, Francis JH, et al. Intravitreal chemotherapy in retinoblastoma: expanded use beyond intravitreal seeds. Br J Ophthalmol. 2018 Jun 6. pii: bjophthalmol-2018-312037. https://doi.org/10.1136/ bjophthalmol-2018-312037 [Epub ahead of print]. 88. Francis JH, Abramson DH, Ji X, et al. Risk of extraocular extension in eyes with retinoblastoma receiving intravitreous chemotherapy. JAMA Ophthalmol. 2017;135(12):1426–9. 89. Aziz HA, Kim JW, Munier FL, et al. Acute hemorrhagic retinopathy following intravitreal Melphalan injection for retinoblastoma: a report of two cases and technical modifications to enhance the prevention of retinal toxicity. Ocul Oncol Pathol. 2017;3(1):34–40. 90. Shields CL, Manjandavida FP, Arepalli S, et  al. Intravitreal melphalan for persistent or recurrent retinoblastoma vitreous seeds: preliminary results. JAMA Ophthalmol. 2014;132:319. 91. Singh AD.  Ocular phototherapy. Eye (Lond). 2013;27(2):190–8. 92. Abramson DH, Schefler AC.  Transpupillary thermotherapy as initial treatment for small intraocular retinoblastoma: technique and predictors of success. Ophthalmology. 2004;111(5):984–91. 93. Francis JH, Abramson DH, Brodie SE, et  al. Indocyanine green enhanced transpupillary thermotherapy in combination with ophthalmic artery chemosurgery for retinoblastoma. Br J Ophthalmol. 2013;97(2):164–8. 94. Ediriwickrema LS, Francis JH, Arslan-Carlon V, et  al. Intravenous injection of indocyanine green results in an artificial transient desaturation by pulse oximetry. Retin Cases Brief Rep. 2015;9(3):252–5. 95. Al-Haddad CE, Abdulaal M, Saab RH, et  al. Indocyanine green-enhanced thermotherapy for retinoblastoma. Ocul Oncol Pathol. 2015;1(2):77–82. 96. Hasanreisoglu M, Saktanasate J, Schwendeman R, et  al. Indocyanine green-enhanced transpupillary thermotherapy for retinoblastoma: analysis of 42 tumors. J Pediatr Ophthalmol Strabismus. 2015;52(6):348–54. 97. Lee TC, Lee SW, Dinkin MJ, et al. Chorioretinal scar growth after 810-nanometer laser treatment for retinoblastoma. Ophthalmology. 2004;111(5):992–6. 98. Abramson DH, Ellsworth RM, Rozakis GW.  Cryotherapy for retinoblastoma. Arch Ophthalmol. 1982;100(8):1253–6.

10  Retinoblastoma: Treatment Options 99. Schueler AO, Flühs D, Anastassiou G, et  al. Beta-ray brachytherapy with 106Ru plaques for retinoblastoma. Int J Radiat Oncol Biol Phys. 2006;65(4):1212–21. 100. Shields CL, Shields JA, Cater J, et al. Plaque radiotherapy for retinoblastoma: long-term tumor control and treatment complications in 208 tumors. Ophthalmology. 2001;108(11):2116–21. 101. Buys RJ, Abramson DH, Ellsworth RM, et  al. Radiation regression patterns after cobalt plaque insertion for retinoblastoma. Arch Ophthalmol. 1983;101(8):1206–8. 102. Merchant TE, Gould CJ, Wilson MW, et  al. Episcleral plaque brachytherapy for retinoblastoma. Pediatr Blood Cancer. 2004;43(2):134–9. 103. Merchant TE, Gould CJ, Hilton NE, et  al. Ocular preservation after 36  Gy external beam radiation therapy for retinoblastoma. J Pediatr Hematol Oncol. 2002;24(4):246–9. 104. McCormick B, Ellsworth R, Abramson D, et  al. Radiation therapy for retinoblastoma: comparison of results with lens-sparing versus lateral beam techniques. Int J Radiat Oncol Biol Phys. 1988;15(3):567–74. 105. McCormick B, Ellsworth R, Abramson D, et  al. Results of external beam radiation for children with retinoblastoma: a comparison of two techniques. J Pediatr Ophthalmol Strabismus. 1989;26(5): 239–43. 106. Zelter M, Damel A, Gonzalez G, et  al. A prospective study on the treatment of retinoblastoma in 72 patients. Cancer. 1991;68(8):1685–90. 107. Shidnia H, Hornback NB, Helveston EM, et  al. Treatment results of retinoblastoma at Indiana University Hospitals. Cancer. 1977;40(6): 2917–22. 108. Foote RL, Garretson BR, Schomberg PJ, et  al. External beam irradiation for retinoblastoma: patterns of failure and dose-­response analysis. Int J Radiat Oncol Biol Phys. 1989;16(3):823–30. 109. Abramson DH, Beaverson KL, Chang ST, et  al. Outcome following initial external beam radiotherapy in patients with Reese-Ellsworth group Vb retinoblastoma. Arch Ophthalmol. 2004;122(9):1316–23. 110. Eng C, Li FP, Abramson DH, et al. Mortality from second tumors among long-term survivors of retinoblastoma. J Natl Cancer Inst. 1993;85(14):1121–8.

139 111. Wong FL, Boice JD Jr, Abramson DH, et al. Cancer incidence after retinoblastoma. Radiation dose and sarcoma risk. JAMA. 1997;278(15):1262–7. 112. Moll AC, Imhof SM, Schouten-Van Meeteren AY, et al. Second primary tumors in hereditary retinoblastoma: a register-based study, 1945–1997: is there an age effect on radiation-related risk? Ophthalmology. 2001;108(6):1109–14. 113. Kleinerman RA, Tucker MA, Tarone RE, et al. Risk of new cancers after radiotherapy in long-term survivors of retinoblastoma: an extended follow-up. J Clin Oncol. 2005;23(10):2272–9. 114. Abramson DH, Frank CM. Second nonocular tumors in survivors of bilateral retinoblastoma: a possible age effect on radiation-related risk. Ophthalmology. 1998;105(4):573–9; discussion 579–80. 115. Abramson DH, Jereb B, Ellsworth RM.  External beam radiation for retinoblastoma. Bull N Y Acad Med. 1981;57(9):787–803. 116. Fontanesi J, Pratt CB, Kun LE, et al. Treatment outcome and dose-­response relationship in infants younger than 1 year treated for retinoblastoma with primary irradiation. Med Pediatr Oncol. 1996;26(5):297–304. 117. Abramson DH, Ellsworth RM.  The surgical management of retinoblastoma. Ophthalmic Surg. 1980;11(9):596–8. 118. Tawfik HA.  Superomedial lid crease approach to the medial intraconal space. Ophthal Plast Reconstr Surg. 2002;18(2):164; author reply 164–5. 119. Abramson DH, Schefler AC, Almeida D, et  al. Optic nerve tissue shrinkage during pathologic processing after enucleation for retinoblastoma. Arch Ophthalmol. 2003;121(1):73–5. 120. Kim JW, Kikkawa DO, Aboy A, et al. Chronic exposure of hydroxyapatite orbital implants: cilia implantation and epithelial downgrowth. Ophthal Plast Reconstr Surg. 2000;16(3):216–22. 121. Kim JW, Kathpalia V, Dunkel IJ, et al. Orbital recurrence of retinoblastoma following enucleation. Br J Ophthalmol. 2009;93(4):463–7. 122. Zhao J, Li Q, Wu S, et al. Pars plana vitrectomy and endoresection of refractory intraocular retinoblastoma. Ophthalmology. 2018;125(2):320–2. 123. Kim JW, Jacobsen B, Zolfaghari E, et  al. Rabbit model of ocular indirect photodynamic therapy using a retinoblastoma xenograft. Graefes Arch Clin Exp Ophthalmol. 2017;255(12):2363–73.

Retinoblastoma: Focal Therapy: Laser Treatment and Cryotherapy

11

Jesse L. Berry and A. Linn Murphree

Introduction

Terminology

In the management of intraocular retinoblastoma, the term “focal therapy” or “consolidation” refers to local treatment modalities such as laser, cryotherapy, and brachytherapy. They can be used as primary treatment for small tumors or in conjunction with intravenous chemotherapy for larger tumors (i.e., chemoreduction followed by consolidation). Focal therapies have the inherent advantage of eradicating focal areas of tumor formation in the retina without any risk of regional or systemic side effects. In this chapter, general guidelines on the use of laser and cryotherapy focal therapies are provided to assist an ophthalmic surgeon who is relatively new to the treatment of retinoblastoma. This chapter might also be of help to those ophthalmologists who would like to compare their current approach with principles and techniques used by other surgeons. Brachytherapy is discussed in chapter 12.

Focal Treatment

J. L. Berry (*) Department of Ophthalmology, USC Roski Eye Institute, Keck School of Medicine of USC, Children’s Hospital Los Angeles, Los Angeles, CA, USA e-mail: [email protected] A. L. Murphree Department of Ophthalmology, Children’s Hospital Los Angeles, University of Southern California, Los Angeles, CA, USA USC Roski Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA © Springer Nature Switzerland AG 2019 J. L. Berry et al. (eds.), Clinical Ophthalmic Oncology, https://doi.org/10.1007/978-3-030-11123-6_11

The term “focal therapy” in the management of retinoblastoma refers to the use of laser treatment, cryotherapy, and brachytherapy (Table  11.1). External beam radiotherapy (EBR) of retinoblastoma (local therapy rather than focal therapy) is discussed elsewhere (Chap. 16).

Focal Primary Treatment Primary treatment refers to focal treatment employed as the sole therapy for a retinoblastoma lesion, typically for very small tumors (group A).

Chemoreduction The term “chemoreduction” is used to describe the induction of tumor shrinkage with chemotherapy, followed by subsequent consolidation treatment with focal therapies, and less commonly with EBR.

Consolidation Treatment In most centers, focal treatment is utilized most frequently following primary intravenous or intra-arte141

142

J. L. Berry and A. L. Murphree

Table 11.1  Local treatment of retinoblastoma Treatment Laser photocoagulation

Indication Primary treatment, consolidation treatment, and for tumor recurrence Tumors not more than 3 mm in diameter, with no evidence of seeding, and located posterior to the equator Laser thermotherapy Primary treatment, consolidation treatment, and for tumor recurrence Tumors not more than 3 mm in diameter, with no evidence of seeding, and located posterior to the equator Thermochemotherapy Consolidation treatment Tumors not more than 12 mm in diameter with no evidence of seeding, and located posterior to the equator Cryotherapy

Brachytherapy

Complications Tumor seeding into vitreous, retinal fibrosis and traction, retinal vascular occlusion

Iris atrophy, focal cataracts, tumor seeding into vitreous, retinal fibrosis and traction, retinal vascular occlusion Iris atrophy, focal cataracts, tumor seeding into vitreous, transient retinal detachment, diffuse choroidal atrophy Primary treatment, consolidation treatment, and for tumor Large area of retinal atrophy, recurrence transient retinal detachment, Tumors not more than 3-5 mm in diameter with no evidence retinal hole, retinal detachment of seeding, and located anterior to the equator. “Cutting cryo” for posterior tumors too large for laser consolidation (e.g., conjunctival incision to allow posterior access) Radiation retinopathy, Primary treatment, residual tumor following radiation optic neuropathy photocoagulation/thermotherapy/thermochemotherapy/ cryotherapy, and for tumor recurrence Tumor less than 15 mm in diameter Presence of diffuse vitreous seeding is contraindication

rial chemotherapy (i.e., chemoreduction) for group B–D tumors. The term “consolidation,” as used in oncology, is a therapy that is used in tandem with primary therapy to further eliminate the tumor cells that were resistant to, or were not inactivated by, the primary therapy. In most other childhood tumors, consolidation involves switching treatment modalities entirely or at least changing to different agents and/or doses of the primary modality. In the case of intraocular retinoblastoma, focal consolidation consists of direct laser photocoagulation, thermotherapy, cryotherapy, or brachytherapy.

Photocoagulation Described by Meyer-Schwickerath in 1949, photocoagulation involves heating of the tumor to temperatures above 65 °C [1].

Hyperthermia Hyperthermia implies raising the tumor temperature to 42–45 °C. Hyperthermia can be induced

by laser, microwave, ultrasound, a localized current field, or ferromagnetic thermoseeds.

Thermotherapy During the thermotherapy, the tumor is heated to a temperature of 45–60  °C.  Oosterhuis and coworkers in 1994 introduced thermotherapy for choroidal melanomas using an infrared red laser through the pupil (transpupillary thermotherapy [TTT]) [2]. Increased depth of tumor necrosis was achieved with TTT as compared with photocoagulation. Unlike hyperthermia, the cytotoxic effects of TTT are irreversible. Transpupillary thermotherapy can be achieved in retinoblastoma tumors using the 810 nm red laser if the continuous mode is used to treat each spot for 45–60 s.

Focal Primary Treatment Group A eyes with small intraretinal lesions away from critical structures are ideal candidates for focal primary therapy such as direct laser photo-

11  Retinoblastoma: Focal Therapy: Laser Treatment and Cryotherapy

coagulation or cryotherapy. Larger focal tumors may be candidates for brachytherapy, and the indication for plaque radiotherapy is discussed below. Tumor foci that have not been treated with systemic chemotherapy may be more fragile and sensitive to intense energy density from the laser. For this reason, small spot size, high power, and prolonged burn duration, all of which contribute to increased power density, should be used with caution to avoid sudden mechanical tumor disruption and dissemination which may lead to seeding.

Focal Consolidation Treatment  hotocoagulation with Argon Green P Laser (532 nm) Background The argon 532 nm (green) laser is particularly useful for very small retinoblastoma tumors (1.5  mm or less) or for focal consolidation after at least one cycle of chemotherapy. As with other uses of retinal photocoagulation, focal consolidation should not be attempted if the retina containing/overlying the lesion is detached given its limited efficacy. We have found that the argon laser midrange visible (532 nm) wavelength is more readily absorbed

a

143

in the nonpigmented retinoblastoma tissue than the longer wavelength infrared 810  nm red laser, which becomes a factor in thick tumors or those occurring over calcified scars. The margins of the treated zone when using the argon laser are also easier to control than with the red laser which is helpful for tumors near critical ocular structures. Its main disadvantage when compared to the red laser is the small spot size. Care must be taken to increase the power density judiciously within the small spot to avoid tumor dissemination, breaks in the retinal pigment epithelium (RPE), or hemorrhages. Tumor disruption may occur in a small spot if the power out of the laser exceeds 700–800 mw for more than 0.3–0.4 s.

Technique The 532 nm green laser is available as a tabletop solid state laser with an indirect laser delivery system that works best for transpupillary laser applications under general anesthesia. The desirable end point for the ophthalmologist is a gentle white spot generated at the boundary between normal retina and tumor edge (Fig.  11.1). The power is initially set between 250 and 350 mw for 0.3–0.5 s. Laser burns are initially placed at the edge of the lesion, half-on and half-off the tumor. The power and/or burn duration is slowly increased until a clear reaction is achieved.

b

Fig. 11.1  Image taken immediately after the third consolidation laser photocoagulation (a). Each lesion was covered with laser burns. Note the distinct gentle white

burn at the lesion edge. There is differential energy uptake. Three weeks later, all lesions are all flat with no clinical evidence of active disease (b)

144

J. L. Berry and A. L. Murphree

Punctate hemorrhage within the treatment spot is a sign that the power density is near the maximum tolerated levels. Once the appropriate power level is set, the edge of the tumor is treated with overlapping burns to establish the perimeter of the lesion. Subsequently the entire lesion should be treated with burns having the same overlap. In the central thicker portions of the tumor, a visible whitening reaction following treatment may not be present. However, neither the power nor the burn duration should be increased once the parameters have been established.

Frequency of Treatment Typically the treatment is repeated every 3–4  weeks, immediately before the next cycle of chemotherapy. A 2–4  week interval can be adopted for consolidation after chemotherapy treatment has been completed. Edge tumor recur-

rence may appear if the laser consolidation process was insufficient, typically within the first 6 months after the last laser session (rarely up to 2 years) (Fig. 11.2).

Mechanism of Action When photocoagulation is used on retinoblastoma lesions and the patient subsequently receives planned systemic chemotherapy (e.g., carboplatin) within 24  hours, two tumor-­ destroying mechanisms may come into play. The first and the most important is direct tumor cell destruction generated by temperatures in excess of 65–70 °C within the treatment spot. The second mechanism occurs in the “donut” or ring of tissue extending for several millimeters outside the laser spot. Heat radiating out from the central spot increases the temperature to the thermotherapy range between 45 and 60 °C. In this region, there is a synergism with the carboplatin, assuming an adequate level

a

b

d

e

Fig. 11.2  Course of consolidation treatment: (a) group B peripapillary tumor at diagnosis, (b) regression noted after one cycle of chemoreduction, laser treatment was initiated at this time (subsequent to this photograph), (c) regression after six  cycles of chemoreduction and laser consolidation, (d) peripheral scar recurrence along the

c

temporal margin of the scar noted 3 months after the completion of chemotherapy which was treated with further focal laser consolidation, (e) complete regression after completion of laser treatment. Both diode and argon lasers were used in the treatment of this lesion

11  Retinoblastoma: Focal Therapy: Laser Treatment and Cryotherapy

145

of carboplatin is achieved in the tumor. To take advantage of the latter mechanism, we typically perform laser treatment during the examination under anesthesia within 24 hours of the next cycle of intravenous chemotherapy.

However, it is more difficult to control the size of the burn with the red laser, and the scar tends to spread beyond the burn that is seen clinically. Thus, extra care is warranted near the optic nerve and fovea with this laser.

Recommendations In Los Angeles, our treatment protocol requires that each lesion be treated completely with laser on at least two occasions, 2–4 weeks apart. In our experience, even a flat chorioretinal scar achieved after one laser session cannot be considered free of disease, and a second treatment should be gently applied. Larger lesions should be lasered at sequential examinations (minimum of three sessions) until the regressed tumor is either flat (type IV scar) or completely calcified (type I). Fleshy type II regression should be lasered until you achieve a type I/IV scar or until you begin to notice retinal contraction. Small areas of type II regression may be left untreated immediately adjacent to the optic nerve or fovea, although the risk for tumor recurrence is always higher with type II vs. type I (or type IV) scars in our experience.

Technique The delivery technique is similar to that with argon green laser photocoagulation. The entire tumor is treated with overlap of the spots similar to that described above for the argon 532 nm laser. The initial power settings with the red laser, especially when the large spot is used, are somewhat higher than for the argon laser. We generally select an initial setting of 300 mw for peripheral and macular lesions undergoing primary therapy and 400–500 mw for large tumors undergoing chemoreduction. If there is little color change induced at the initial power settings for these larger tumors, it is possible to increase the power up to 800 mw, provided that the surgeon is carefully monitoring for complications. The power settings will vary for each patient and for each tumor because of the degree of pigmentation underlying the lesion(s). As with the argon laser, both the power and burn duration can be increased incrementally until the appropriate end point is reached. We typically leave the duration of the treatment on the longest setting (9000 ms) and set the interval to 50 ms; with these parameters the laser is essentially being used in continuous mode and the surgeon can control the duration of each spot treatment with the foot pedal. Using the red laser in this manner allows the surgeon to use the red laser for photocoagulation (1–10  s) or possibly thermotherapy (30–50 s).

 hotocoagulation with Red Laser P (810 nm) Background The 810 nm red laser is most effective when there is intact RPE beneath the tumor. Through a 28D lens, the indirect ophthalmoscope delivery system offers a spot size of 0.35  mm (small spot) and 1.4  mm (large spot). We typically use the large spot indirect system, which provides improved safety and convenience for treating larger tumors versus the argon laser. Safety comes from the reduced likelihood of concentrated power intensity in a small spot creating tumor dissemination. The larger spot size as compared with argon saves treatment time, thus conferring convenience. It is also our impression that the depth of treatment with the red laser is greater than the argon laser.

Frequency of Treatment The treatment is repeated every 3–4  weeks immediately before the next cycle of chemotherapy. A 2–4 week interval can be adopted if the course of chemotherapy has been completed and the laser is for consolidation or scar recurrences.

146

Mechanism of Action The most commonly employed effect is the direct heat-mediated tumor cell kill through photocoagulation. When longer spot duration is utilized (30–50  s), thermotherapy can be employed. The red laser is most effective when intact RPE is present beneath the tumor to be treated. If the RPE has been destroyed, it is believed that most of the 810 nm wavelength energy passes into the orbit without being absorbed by the retinoblastoma (see discussion under TTT below). Recommendation If only one laser has to be chosen for use in delivering local therapy to retinoblastoma, the argon green laser is probably the most versatile in our opinion, although the diode is more frequently used.

Transpupillary Thermotherapy (TTT) Transpupillary thermotherapy describes a laser system that couples large spot size (2–3 mm) and

Fig. 11.3  The base settings for the green laser are shown in the upper laser (power 300/duration 300/interval 300) and the red laser below (power 300/duration 9000/interval 50). With any laser modality position of the surgeon is important. Standing 180 degrees away from the lesion to be treated and ensuring it is in central view on depression are critical. The bed should be adjusted to the surgeon

J. L. Berry and A. L. Murphree

long burn duration (1  min) with low power settings, applied to achieve the end point of gentle whitening in the treatment spot. Initially described for choroidal melanoma, the infrared red laser (810 nm) has been shown to be effective in killing melanoma cells because pigmentation in the tumor allows absorption of the laser energy [2]. However, the long-term efficacy of this approach for treating small choroidal melanomas is under question. Transpupillary thermotherapy is difficult to adapt to the treatment of retinoblastoma because of the inherent lack of pigmentation in the tumor. Initially, intact RPE will absorb the laser energy and generate heat to affect the tumor. However, once the RPE is no longer intact under the retinoblastoma, relatively little of the delivered energy will be absorbed [3, 4]. However, a modified TTT regimen can be used for retinoblastoma by employing the large spot 810  nm red laser in a continuous mode and using burn durations of 30–50 s. The effect of thermotherapy may be enhanced by using indocyanine green (ICG), a chromophore with an absorption peak of 805 nm, which coincides with the red laser emission of 810 nm (Fig. 11.3) [5].

height. Any indirect lens can be used per surgeon preference. The viewing of smaller, posterior lesions is facilitated by the 20D lens, and peripheral lesions are more easily seen with the 28D lens; however, the spot size of delivered laser power is larger with the 28D lens. [Published with parental permission]

11  Retinoblastoma: Focal Therapy: Laser Treatment and Cryotherapy

147

Transscleral Cryotherapy Background The indications for cryotherapy are similar to those for laser thermotherapy (i.e., small tumors) except that it is more suitable for anterior tumors [6]. Approximately 90–95% of carefully selected tumors can be treated successfully with cryotherapy [7]. Overall, small tumors less than 3 mm in diameter, below 2 mm in height, and anterior to the equator are ideal candidates for cryotherapy. Larger tumors can occasionally be treated with cryotherapy alone, but the recurrence rate and risk of complications are higher. For anterior group B tumors, it is usually better to utilize another modality such as intravenous or intra-­arterial chemotherapy to shrink the tumor so that it is amenable to local therapy. For tumors with localized vitreous seeding at the apex of the lesion (within 1–2 mm), cryotherapy can be utilized as primary therapy, although patients should be followed carefully because of the significant risk of recurrence and spread of the vitreous seeds.

Technique The cryotherapy machine should be tested to ensure proper functioning and adequate ice ball formation at the tip before the procedure begins. The cryoprobe tip position is verified by indirect ophthalmoscopy using the standard techniques of scleral indentation. Once the tip is centered directly under the tumor, freezing is begun. The ice ball used to freeze a tumor should cover the apex for 1–2 mm for adequate coverage and to incorporate all of the vitreous near the lesion that may contain the local tumor cell clumps or “seeds.” The lateral spread of the ice ball should be monitored as well as the apex of the tumor to minimize the treatment of the uninvolved retina. Double or triple freeze-thaw cycles of cryotherapy are generally applied. The number of sites treated with cryotherapy at one setting should be limited to two or three because of the risk of creating a secondary serous or rhegmatog-

Fig. 11.4  Cryotherapy scar in the inferotemporal periphery. Note the extensive destruction of peripheral retina

enous retinal detachment with more extensive treatment. It should also be kept in mind that cryotherapy tends to destroy a relatively large amount of peripheral retina. Complications of cryotherapy can include retinal breaks that lead to rhegmatogenous retinal detachment, particularly in tumors located in the superior quadrants and those that have preexisting areas of calcification [8]. Cryotherapy is contraindicated for the treatment of more than one quadrant of disease at the ora serrata. For tumors located posterior to the equator, a small conjunctival incision in the fornix located between the rectus muscles may be necessary to advance the curved cryoprobe posteriorly (“cutdown” cryotherapy). Careful monitoring of the probe position is required when performing posterior cryotherapy to avoid inadvertent treatment of the macula or optic nerve. Following the completion of cryotherapy, it is recommended that a drop of atropine be placed in the eye and subconjunctival injection of marcaine and dexamethasone be given for pain control and episcleral scarring, respectively (Fig. 11.4).

Mechanism of Action Cryotherapy is a local destructive modality that kills tumor cells mechanically via ice crystal disruption of the cell membranes.

148

Frequency of Treatment The treatment is repeated every 3–4 weeks. A flat chorioretinal scar is the desired end point which is usually achieved after 1–2 sessions.

Recommendations Cryotherapy is suitable for the anteriorly located group A tumors and some group B tumors. Excessive cryotherapy should be avoided to minimize risk of complications.

Special Recommendations for Consolidation Therapy The application of consolidation therapy for retinoblastoma is as much of an art as it is a science. While laser therapy (either green argon or red diode) can be exceedingly effective both for consolidation and for scar recurrences, treatment-­ related complications can occur including hemorrhage, epiretinal membrane, and RPE migration. More severe complications including focal scleral thinning leading to orbital relapse has also been reported [9]. Thus, there is a balance between the effective application of laser to be tumoricidal while attempting to reduce morbidity from this treatment. Further, while retention of vision in retinoblastoma treatment falls lower in the priority list after preservation of life and the eye, it does remain a concern and can be directly impacted by focal therapy techniques. Thus, approaches to laser therapy of tumors adjacent to critical structures merit additional consideration. An important consideration is the best time to initiate laser therapy for tumors in the posterior pole. To evaluate this, a retrospective review of timing of laser therapy for group B eyes with macular tumors was undertaken [10]. This review found that the final scar size (final size of regressed tumor and adjacent scar) at the end of treatment was 13% smaller than the original tumor size; greater reductions were seen when chemotherapy preceded laser therapy and when the original tumor was larger (>4.5 mm). The authors of this

J. L. Berry and A. L. Murphree

retrospective review (and also this chapter) suggest the preferred approach for macular group B tumors is to initiate chemotherapy and then augment with laser therapy after 1–2  cycles when initial tumor shrinkage has been observed. The advantage of this approach was found to be less significant for smaller tumors (