X-Ray Imaging Systems for Biomedical Engineering Technology: An Essential Guide 3031462653, 9783031462658


168 59 11MB

English Pages [162]

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Preface
Acknowledgments
Contents
Chapter 1: X-Ray Imaging Systems: An Overview
1.1 Introduction
1.2 Film-Screen Radiography
1.3 Digital X-Ray Imaging Systems
1.3.1 Computed Radiography
1.3.2 Flat-Panel Digital Radiography
1.3.3 Digital Fluoroscopy
1.3.4 Computed Tomography: Major System Components
1.4 Radiation Physics at a Glance
1.5 Imaging Informatics: An Overview
1.6 Quality Assurance and Quality Control
1.7 Radiation Protection Considerations in X-Ray Imaging
References
Chapter 2: Basic Radiation Physics for Diagnostic X-Ray Imaging
2.1 Introduction
2.2 X-Ray Generation
2.3 X-Ray Production
2.3.1 Characteristic Radiation
2.3.2 Bremsstrahlung Radiation
2.4 X-Ray Emission
2.5 X-Ray Attenuation
2.6 Interaction of Diagnostic X-Rays with Matter
2.6.1 Mechanisms of Interaction in Diagnostic X-Ray Imaging
2.6.2 Classical Scattering
2.6.3 Compton Scattering
2.6.4 Photoelectric Absorption
References
Chapter 3: Computed Radiography Imaging: Physical Principles and System Components
3.1 Introduction
3.2 Limitations of Film-Screen Radiography (FSR)
3.3 Computed Radiography (CR): System Components
3.3.1 The CR Imaging Plate
3.3.2 Imaging Plate Exposure: Physics at a Glance
3.3.3 Imaging Plate Reader and Processor
3.4 Image Processing
3.4.1 Pre-processing
3.4.2 Postprocessing
3.5 The CR Workstation and Image Display
3.6 Response of the CR IP to Radiation Exposure
3.7 The Standardized Exposure Indicator (EI)
3.7.1 Use of the Deviation Index in Practice
3.8 The Shortcomings of CR
References
Chapter 4: Flat-Panel Digital Radiography: Principles and System Components
4.1 Introduction
4.2 Flat-Panel Digital Radiography: System Components Overview
4.2.1 The Flat-Panel Detector
4.2.2 The Fill Factor
4.3 Types of Flat-Panel Detectors
4.3.1 Indirect Conversion Digital Detectors: Major Components
4.3.2 Direct Conversion Digital Detectors: Major Components
4.4 Image Processing Considerations
4.4.1 Image Pre-processing
4.4.2 Image Postprocessing
4.5 Image Quality Considerations
4.6 Wireless FPDR Detectors
4.7 Mobile Digital Radiography Systems
References
Chapter 5: Digital Fluoroscopy: System Components and Principles
5.1 Introduction
5.2 Development of Fluoroscopy Systems at a Glance
5.2.1 Image Intensifier-Based Digital Fluoroscopy: Essential Technical Components
5.3 Digital Fluoroscopy with Flat-Panel Digital Detectors
5.3.1 System Configuration
5.3.2 Types of Dynamic FPDs
5.3.3 Technical Features of Dynamic FPDs
5.3.4 Principles of Operation
5.4 Image Post Processing in Digital Fluoroscopy
References
Chapter 6: Digital Image Quality Descriptors and Performance Characteristics
6.1 Introduction
6.2 Image Quality Characteristics
6.2.1 Spatial Resolution
6.2.2 Contrast Resolution
6.2.3 Noise
6.3 Performance Characteristics of Digital Detectors
6.3.1 Detective Quantum Efficiency
6.4 Image Artifacts Overview
References
Chapter 7: Computed Tomography: Basic Physics and Technology
7.1 Introduction
7.2 Essential Physical Principles
7.2.1 Radiation Attenuation
7.2.1.1 Beer–Lambert Law
7.2.1.2 Attenuation and CT Numbers
7.3 Technology
7.3.1 Data Acquisition Principles and Components
7.3.2 Image Reconstruction
7.3.3 Iterative Reconstruction
7.3.4 Image Display, Storage, and Communication
7.4 Multislice CT (MSCT): Fundamental Principles and Technology
7.4.1 Slip-Ring Technology
7.4.2 X-Ray Tube Technology
7.4.3 Interpolation Algorithms
7.4.4 MSCT Detector Technology
7.4.5 Selectable Scan Parameters
7.4.6 Isotropic CT Imaging
7.5 MSCT Image Processing
7.5.1 Image Postprocessing
7.5.1.1 Windowing
7.5.2 Three-Dimensional Image Display Techniques
7.6 Image Quality
7.7 Dual Energy CT
7.8 Photon Counting CT
7.8.1 The Major System Components of a PCD
7.9 A Bit of CT Dosimetry
7.9.1 Factors Affecting Patient Dose
7.10 Radiation Protection
7.11 Artificial Intelligence Applications in CT Image Reconstruction
References
Chapter 8: Imaging Informatics Essentials
8.1 Introduction
8.2 Definitions of Imaging Informatics
8.3 Major Elements of Imaging Informatics
8.3.1 MIMPS: Basic System Components
8.4 Health Information Systems: Definition
8.5 The Electronic Health Record
8.6 Other Concepts in Imaging Informatics
8.7 Enterprise Imaging
8.8 The Biomedical Engineering Technologist as Informaticist: A Potential New Role?
References
Chapter 9: Artificial Intelligence in Medical Imaging: An Overview
9.1 Introduction
9.2 Definitions
9.2.1 Artificial Intelligence
9.2.2 Machine Learning
9.2.3 Deep Learning
9.3 How Does AI Work?
9.4 Which Subset of AI to Use? ML or DL?
9.5 Artificial Intelligence Methods in Medical Imaging
9.6 Medical Imaging Applications
9.6.1 AI in Diagnosis of Diseases
9.7 AI in Computed Tomography
9.8 Ethics of AI in Medical Imaging
9.9 The Biomedical Engineering Technologist and AI
References
Chapter 10: Quality Control in Digital X-Ray Imaging Systems
10.1 Introduction
10.2 General Concepts
10.2.1 Definitions
10.2.2 Essential Components of QC
10.2.3 Steps in Conducting a QC Test
10.2.4 The Tolerance Limits or Acceptance Criteria
10.3 Parameters, Tools, and Frequency for QC Testing in Digital Radiography
10.4 Reject/Retake Analysis
10.4.1 Acceptance Limits for Reject/Retake Analysis
10.5 Display Performance
10.5.1 Visual Assessment of the Electronic Display Monitors
10.6 Examples of QC Tests for CR Using Qualitative Criteria
10.7 American College of Radiology (ACR) CT QC Manual
References
Chapter 11: Radiation Protection in X-Ray Imaging
11.1 Introduction
11.2 Biological Effects of Exposure to Ionizing Radiation
11.3 Goals of Radiation Protection
11.3.1 ICRP Radiation Protection Framework
11.3.2 Personal Actions for Radiation Protection
11.4 Regulatory Radiation Protection Guidance Recommendations
11.5 Personnel Dosimetry
11.5.1 Optically Stimulated Luminescence Dosimetry
11.5.2 Thermoluminescent Dosimetry
11.5.3 Wearing the Personnel Dosimeter
References
Index
Recommend Papers

X-Ray Imaging Systems for Biomedical Engineering Technology: An Essential Guide
 3031462653, 9783031462658

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

Euclid Seeram

X-Ray Imaging Systems for Biomedical Engineering Technology An Essential Guide

X-Ray Imaging Systems for Biomedical Engineering Technology

Euclid Seeram

X-Ray Imaging Systems for Biomedical Engineering Technology An Essential Guide

Euclid Seeram Medical Imaging, Faculty of Health University of Canberra Burnaby, BC, Canada

ISBN 978-3-031-46265-8    ISBN 978-3-031-46266-5 (eBook) https://doi.org/10.1007/978-3-031-46266-5 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 This work is subject to copyright. All rights are solely and exclusively licensed 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, expressed 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 Paper in this product is recyclable.

This book is dedicated to my two smart and beautiful granddaughters Claire and Charlotte With love and blessings to you both forever.

Preface

Biomedical Engineering Technology (BMET) students may have to complete laboratory hands-on exercises in the x-ray imaging laboratory as part of their educational requirements. Additionally, biomedical engineering technologists often service x-ray imaging equipment in the radiology department of their hospitals. X-Ray Imaging Systems include Digital Radiography (DR) and Digital Fluoroscopy (DF) Digital Mammography and Computed Tomography (CT). The imaging modalities are based on physics, engineering, and computer science principles that make them useful clinical tools for imaging the patient. These modalities have experienced several technical innovations in recent years that have not only reduced the radiation dose to the patient but improve the image quality needed for diagnostic interpretation. For example, the fundamental concepts of the wide exposure latitude of DR systems and the standardized exposure indicator established by the International Electrotechnical Commission (IEC) have provided researchers with the motivation to operate DR systems with the goal of optimization of the dose and image quality. Additionally, CT technical innovations, such as new detector technology, Iterative Reconstruction (IR) algorithms, and Artificial Intelligence-based image reconstruction are now offered by several CT vendors, play a significant role in dose reduction and optimization in CT. Furthermore, the introduction of Photon Counting Detectors (PCDs) has solved the major problem of image noise during low-dose CT imaging. This book X-Ray Imaging Systems for Biomedical Engineering Technology: An Essential Guide provides a useful resource to meet the x-ray imaging educational requirements of biomedical engineering students and provide a continuing education resource for practicing technologists. This book is intended to meet fundamental requirements for x-ray imaging systems in the BMET curriculum, in the United States, Canada, South America, Africa, Asia, Australia, and continental Europe. The contents in this book are described in 11 chapters as follows: Chapter 1 introduces the nature and scope of X-Ray Imaging Systems and sets the general framework for the remaining chapters. Whereas Chap. 2 presents a description of the essential Radiation Physics needed for a good understanding of how these systems work and how x-rays interact with the patient to generate images vii

viii

Preface

used for diagnosis. Chapter 3 describes the major technical components of Computed Radiography (CR). Chapter 4 deals with the physical principles and technical components of Flat-Panel Digital Radiography (FPDR). The principles and technology of Digital Fluoroscopy are discussed in Chap. 5. Chapter 6 identifies and outlines Digital Image Quality Descriptors and Performance Characteristics, such as spatial resolution, contrast resolution, and noise as well as the detective quantum efficiency (DQE) which describes the noise characteristics of a digital detector. In Chap. 7, the physical principles and technology of Computed Tomography (CT) at a depth needed for entry-to-practice, specifically the major technical system components of Multi-Slice CT (MSCT), are described. Furthermore, image processing, image quality, and radiation dose and radiation protection are described. Chapter 8 introduces the ideas and major concepts of Imaging Informatics including major topics, such as Medical Image Management and Processing System (MIMPS) formerly referred to as Picture Archiving and Communication Systems (PACS), and related topics, such as Health Information Systems, The Electronic Health Record, Other Concepts in Imaging Informatics, and Enterprise Imaging. Artificial Intelligence (AI) and its subsets, Machine Learning and Deep Learning, as well as applications of AI in medical imaging; AI in Diagnosis of Diseases; AI in Computed Tomography; and Ethics of AI in Medical Imaging are reviewed in Chap. 9. Chapter 10 deals with the major elements of Quality Control in Diagnostic X-Ray Imaging focusing on steps in conducting a QC, tolerance limits or acceptance criteria, reject/retake analysis, visual assessment of the electronic display monitors, and describes briefly four examples of QC Tests for CR using qualitative criteria. These include dark noise, CR Imaging Plate (IP)test for uniformity, spatial accuracy, and erasure thoroughness of the IP. Finally, Chap. 11 presents an overview of Radiation Protection for Diagnostic X-Ray Imaging, including biological effects of exposure to ionizing radiation; goals of radiation protection; regulatory radiation protection guidance recommendations; and personnel dosimetry. Enjoy the pages that follow and remember—your wisdom of ensuring that x-ray imaging equipment performs efficiently, effectively and is absolutely safe for imaging patients. Burnaby, BC, Canada

Euclid Seeram

Acknowledgments

An important satisfying task in writing a book of this nature is to thank those medical physicists, biomedical engineers, computer scientists, and manufacturers who have done the original work in conceptualizing and developing x-ray systems for imaging patients. This book deals with x-ray imaging systems as outlined in the Preface, and in this respect, it is indeed a pleasure to express sincere thanks to all of the scientific and clinical experts. I am most grateful to Dr. Rob Davidson, PhD, MAppSc (MI), BBus, FASMIRT, Professor of Medical Imaging, University of Canberra, Australia. Thanks mate. Furthermore, two medical physicists to whom I am truly grateful for my Digital Radiography (DR) education are Dr. Anthony Siebert, PhD, of the University of California at Davis, and Dr. Charles Willis, PhD, formerly of the University of Texas, MD Anderson Cancer Center, from whom I have learned the physics and technical aspects of digital radiography through their seminars and workshops that I have attended. My X-Ray Imaging education (including CT physics and instrumentation) has its roots in not only attending courses, conferences and workshops, but personal communications with a number of notable medical physicists. Two other notable medical physicists to whom I am particularly grateful are Dr. Perry Sprawls (PhD, FACR, FAAPM, FIOMP, Distinguished Emeritus Professor, Emory University, Director, Sprawls Educational Foundation, http://www.sprawls. org, Co-Director, College on Medical Physics, ICTP, Trieste, Italy, and Co-Editor, Medical Physics International) and Dr. Anthony Wolbarst, PhD, Medical Physics Department, University of Kentucky (Retired) from whose published works in medical physics I have learned a great deal. Another individual to whom I owe a good deal of thanks is Valentina Al Hamouche, MRT(R), MSc, who is the CEO/Founder of VCA Education Solutions for Health Professionals based in Toronto, Canada. Valentina has provided me with recurring opportunities to provide Radiographic Imaging Sciences and CT Physics and Technology in-house lectures and webinars to further educate technologists and students across Canada and internationally as well. Thanks Valentina. I am particularly grateful to Jochen Boehm, MEng, and Dr. Anthony Chan, PhD, both Biomedical Engineers and Faculty in the Biomedical Engineering Technology ix

x

Acknowledgments

Program at the British Columbia Institute of Technology (BCIT); for on-going invitations to present lectures on topics relating to medical imaging, to their students. I am also grateful to all the reviewers of this proposal, including Biomedical Engineer, Dr. Quang Vo, PhD, Faculty in Biomedical Engineering Technology Program at BCIT One individual to whom I am especially grateful is Merry Stuber, Senior Editor, Cell Biology and Biomedical Engineering at Springer, a part of Springer Nature, who did all the hard work in not only reviewing the proposal herself but also in getting external reviews of the proposal that led her to accept it for publication. Merry has provide the needed continuous support and encouragement to bring this work to fruition. Additionally, I am grateful to members of the production team at Springer Nature, who have worked exceptionally hard during the production of this book, especially in the page-proof stages. I humbly acknowledge the support from my beautiful family; first my lovely wife, Trish, a warm, smart, caring, and a very special person in my life, thanks babes. Secondly, my caring and very brilliant son David, the best Dad on the planet, to his two most precious daughters, my granddaughters, to whom this book is dedicated. Thanks for your enduring love, support, and encouragement. Burnaby, BC, Canada

Euclid Seeram

Contents

1

 X-Ray Imaging Systems: An Overview��������������������������������������������������    1 1.1 Introduction��������������������������������������������������������������������������������������    1 1.2 Film-Screen Radiography ����������������������������������������������������������������    2 1.3 Digital X-Ray Imaging Systems ������������������������������������������������������    4 1.3.1 Computed Radiography��������������������������������������������������������    5 1.3.2 Flat-Panel Digital Radiography��������������������������������������������    6 1.3.3 Digital Fluoroscopy��������������������������������������������������������������    6 1.3.4 Computed Tomography: Major System Components����������    8 1.4 Radiation Physics at a Glance����������������������������������������������������������   10 1.5 Imaging Informatics: An Overview��������������������������������������������������   12 1.6 Quality Assurance and Quality Control��������������������������������������������   13 1.7 Radiation Protection Considerations in X-Ray Imaging������������������   13 References��������������������������������������������������������������������������������������������������   15

2

 Basic Radiation Physics for Diagnostic X-Ray Imaging����������������������   17 2.1 Introduction��������������������������������������������������������������������������������������   17 2.2 X-Ray Generation ����������������������������������������������������������������������������   17 2.3 X-Ray Production ����������������������������������������������������������������������������   19 2.3.1 Characteristic Radiation��������������������������������������������������������   22 2.3.2 Bremsstrahlung Radiation����������������������������������������������������   22 2.4 X-Ray Emission��������������������������������������������������������������������������������   23 2.5 X-Ray Attenuation����������������������������������������������������������������������������   24 2.6 Interaction of Diagnostic X-Rays with Matter����������������������������������   26 2.6.1 Mechanisms of Interaction in Diagnostic X-Ray Imaging ��   27 2.6.2 Classical Scattering��������������������������������������������������������������   27 2.6.3 Compton Scattering��������������������������������������������������������������   28 2.6.4 Photoelectric Absorption������������������������������������������������������   29 References��������������������������������������������������������������������������������������������������   30

xi

xii

Contents

3

Computed Radiography Imaging: Physical Principles and System Components ������������������������������������������������������������������������   31 3.1 Introduction��������������������������������������������������������������������������������������   31 3.2 Limitations of Film-Screen Radiography (FSR)������������������������������   33 3.3 Computed Radiography (CR): System Components������������������������   33 3.3.1 The CR Imaging Plate����������������������������������������������������������   33 3.3.2 Imaging Plate Exposure: Physics at a Glance����������������������   35 3.3.3 Imaging Plate Reader and Processor������������������������������������   36 3.4 Image Processing������������������������������������������������������������������������������   37 3.4.1 Pre-processing����������������������������������������������������������������������   37 3.4.2 Postprocessing����������������������������������������������������������������������   39 3.5 The CR Workstation and Image Display������������������������������������������   40 3.6 Response of the CR IP to Radiation Exposure ��������������������������������   40 3.7 The Standardized Exposure Indicator (EI) ��������������������������������������   42 3.7.1 Use of the Deviation Index in Practice ��������������������������������   43 3.8 The Shortcomings of CR������������������������������������������������������������������   44 References��������������������������������������������������������������������������������������������������   44

4

Flat-Panel Digital Radiography: Principles and System Components ������������������������������������������������������   47 4.1 Introduction��������������������������������������������������������������������������������������   47 4.2 Flat-Panel Digital Radiography: System Components Overview����   47 4.2.1 The Flat-Panel Detector��������������������������������������������������������   48 4.2.2 The Fill Factor����������������������������������������������������������������������   49 4.3 Types of Flat-Panel Detectors ����������������������������������������������������������   49 4.3.1 Indirect Conversion Digital Detectors: Major Components���������������������������������������������������������������   50 4.3.2 Direct Conversion Digital Detectors: Major Components���������������������������������������������������������������   50 4.4 Image Processing Considerations ����������������������������������������������������   51 4.4.1 Image Pre-processing������������������������������������������������������������   51 4.4.2 Image Postprocessing�����������������������������������������������������������   52 4.5 Image Quality Considerations����������������������������������������������������������   53 4.6 Wireless FPDR Detectors ����������������������������������������������������������������   54 4.7 Mobile Digital Radiography Systems����������������������������������������������   54 References��������������������������������������������������������������������������������������������������   55

5

 Digital Fluoroscopy: System Components and Principles ������������������   57 5.1 Introduction��������������������������������������������������������������������������������������   57 5.2 Development of Fluoroscopy Systems at a Glance��������������������������   57 5.2.1 Image Intensifier-Based Digital Fluoroscopy: Essential Technical Components������������������������������������������   58 5.3 Digital Fluoroscopy with Flat-Panel Digital Detectors��������������������   60 5.3.1 System Configuration ����������������������������������������������������������   61 5.3.2 Types of Dynamic FPDs ������������������������������������������������������   63 5.3.3 Technical Features of Dynamic FPDs����������������������������������   63

Contents

xiii

5.3.4 Principles of Operation ��������������������������������������������������������   63 5.4 Image Post Processing in Digital Fluoroscopy ��������������������������������   64 References��������������������������������������������������������������������������������������������������   65 6

 Digital Image Quality Descriptors and Performance Characteristics  67 6.1 Introduction��������������������������������������������������������������������������������������   67 6.2 Image Quality Characteristics����������������������������������������������������������   67 6.2.1 Spatial Resolution ����������������������������������������������������������������   69 6.2.2 Contrast Resolution��������������������������������������������������������������   70 6.2.3 Noise ������������������������������������������������������������������������������������   71 6.3 Performance Characteristics of Digital Detectors����������������������������   73 6.3.1 Detective Quantum Efficiency����������������������������������������������   73 6.4 Image Artifacts Overview ����������������������������������������������������������������   75 References��������������������������������������������������������������������������������������������������   76

7

 Computed Tomography: Basic Physics and Technology����������������������   77 7.1 Introduction��������������������������������������������������������������������������������������   77 7.2 Essential Physical Principles������������������������������������������������������������   78 7.2.1 Radiation Attenuation ����������������������������������������������������������   79 7.3 Technology����������������������������������������������������������������������������������������   84 7.3.1 Data Acquisition Principles and Components����������������������   84 7.3.2 Image Reconstruction ����������������������������������������������������������   85 7.3.3 Iterative Reconstruction��������������������������������������������������������   87 7.3.4 Image Display, Storage, and Communication����������������������   89 7.4 Multislice CT (MSCT): Fundamental Principles and Technology ��   89 7.4.1 Slip-Ring Technology ����������������������������������������������������������   89 7.4.2 X-Ray Tube Technology ������������������������������������������������������   90 7.4.3 Interpolation Algorithms������������������������������������������������������   90 7.4.4 MSCT Detector Technology ������������������������������������������������   91 7.4.5 Selectable Scan Parameters��������������������������������������������������   93 7.4.6 Isotropic CT Imaging������������������������������������������������������������   94 7.5 MSCT Image Processing������������������������������������������������������������������   95 7.5.1 Image Postprocessing�����������������������������������������������������������   96 7.5.2 Three-Dimensional Image Display Techniques��������������������   97 7.6 Image Quality������������������������������������������������������������������������������������   98 7.7 Dual Energy CT��������������������������������������������������������������������������������  100 7.8 Photon Counting CT ������������������������������������������������������������������������  100 7.8.1 The Major System Components of a PCD����������������������������  100 7.9 A Bit of CT Dosimetry ��������������������������������������������������������������������  102 7.9.1 Factors Affecting Patient Dose���������������������������������������������  103 7.10 Radiation Protection��������������������������������������������������������������������������  103 7.11 Artificial Intelligence Applications in CT Image Reconstruction����  104 References��������������������������������������������������������������������������������������������������  104

8

Imaging Informatics Essentials��������������������������������������������������������������  107 8.1 Introduction��������������������������������������������������������������������������������������  107

xiv

Contents

8.2 Definitions of Imaging Informatics��������������������������������������������������  108 8.3 Major Elements of Imaging Informatics������������������������������������������  109 8.3.1 MIMPS: Basic System Components������������������������������������  109 8.4 Health Information Systems: Definition ������������������������������������������  113 8.5 The Electronic Health Record����������������������������������������������������������  114 8.6 Other Concepts in Imaging Informatics��������������������������������������������  114 8.7 Enterprise Imaging����������������������������������������������������������������������������  116 8.8 The Biomedical Engineering Technologist as Informaticist: A Potential New Role?������������������������������������������  116 References��������������������������������������������������������������������������������������������������  116 9

 Artificial Intelligence in Medical Imaging: An Overview��������������������  119 9.1 Introduction��������������������������������������������������������������������������������������  119 9.2 Definitions����������������������������������������������������������������������������������������  119 9.2.1 Artificial Intelligence������������������������������������������������������������  120 9.2.2 Machine Learning ����������������������������������������������������������������  122 9.2.3 Deep Learning����������������������������������������������������������������������  123 9.3 How Does AI Work? ������������������������������������������������������������������������  124 9.4 Which Subset of AI to Use? ML or DL?������������������������������������������  125 9.5 Artificial Intelligence Methods in Medical Imaging������������������������  125 9.6 Medical Imaging Applications����������������������������������������������������������  125 9.6.1 AI in Diagnosis of Diseases��������������������������������������������������  127 9.7 AI in Computed Tomography ����������������������������������������������������������  127 9.8 Ethics of AI in Medical Imaging������������������������������������������������������  128 9.9 The Biomedical Engineering Technologist and AI ��������������������������  129 References��������������������������������������������������������������������������������������������������  130

10 Quality  Control in Digital X-Ray Imaging Systems ����������������������������  133 10.1 Introduction������������������������������������������������������������������������������������  133 10.2 General Concepts����������������������������������������������������������������������������  133 10.2.1 Definitions��������������������������������������������������������������������������  133 10.2.2 Essential Components of QC����������������������������������������������  134 10.2.3 Steps in Conducting a QC Test ������������������������������������������  134 10.2.4 The Tolerance Limits or Acceptance Criteria ��������������������  135 10.3 Parameters, Tools, and Frequency for QC Testing in Digital Radiography��������������������������������������������������������������������  136 10.4 Reject/Retake Analysis ������������������������������������������������������������������  137 10.4.1 Acceptance Limits for Reject/Retake Analysis������������������  137 10.5 Display Performance����������������������������������������������������������������������  137 10.5.1 Visual Assessment of the Electronic Display Monitors������  138 10.6 Examples of QC Tests for CR Using Qualitative Criteria��������������  139 10.7 American College of Radiology (ACR) CT QC Manual����������������  141 References��������������������������������������������������������������������������������������������������  142 11 Radiation  Protection in X-Ray Imaging������������������������������������������������  145 11.1 Introduction������������������������������������������������������������������������������������  145

Contents

xv

11.2 Biological Effects of Exposure to Ionizing Radiation��������������������  146 11.3 Goals of Radiation Protection��������������������������������������������������������  146 11.3.1 ICRP Radiation Protection Framework������������������������������  146 11.3.2 Personal Actions for Radiation Protection��������������������������  147 11.4 Regulatory Radiation Protection Guidance Recommendations������  148 11.5 Personnel Dosimetry����������������������������������������������������������������������  149 11.5.1 Optically Stimulated Luminescence Dosimetry ����������������  149 11.5.2 Thermoluminescent Dosimetry������������������������������������������  150 11.5.3 Wearing the Personnel Dosimeter��������������������������������������  151 References��������������������������������������������������������������������������������������������������  151 Index������������������������������������������������������������������������������������������������������������������  153

Chapter 1

X-Ray Imaging Systems: An Overview

1.1 Introduction The Canadian Medical and Biological Engineering Society (CMBES) defines a biomedical engineering technologist as one who works “in large health care facilities providing comprehensive service and support of medical devices and equipment. Activities include inspection, installation, repair, and preventive maintenance of medical devices and complex medical systems. They also provide advice and training on the safe and effective use of medical devices and systems” [1]. An example of such a “complex medical system” is the X-ray imaging system commonly found in medical imaging departments (radiology) worldwide. X-ray imaging systems produce ionizing X-rays to generate patient images. Given that these systems expose patients to radiation, it is mandatory that the biomedical engineering technologist to comprehend the workings of these systems for the safety of patients and related personnel. X-ray imaging systems include film-screen imaging (now obsolete); digital X-ray imaging systems such as computed radiography, flat-panel digital radiography, digital fluoroscopy, digital mammography, and computed tomography (Fig. 1.1). These systems are based on radiation physics which explains how they work. The biomedical engineering technologist ensures the proper and safe functioning of these systems in the care and management of the patient. Additionally, two important considerations that are also essential to the skill set of the biomedical engineering technologist are equipment quality control and radiation protection. The objective of this chapter is to present a comprehensive overview of X-ray imaging systems, including the nature of quality assurance and quality ­control. Additionally, it will provide a summary of radiation physics and radiation protection.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_1

1

2

1  X-Ray Imaging Systems: An Overview

Fig. 1.1  X-Ray imaging systems include film-­ screen imaging (now obsolete), digital X-ray imaging systems such as computed radiography, flat-panel digital radiography, digital fluoroscopy, digital mammography, and computed tomography

Fig. 1.2  The basic steps in the production of a film-screen radiographic image. See text for further explanation

1.2 Film-Screen Radiography Film-Screen Radiography (FSR) served as the primary X-ray imaging system for clinical applications for several years following the discovery of X-rays. In FSR, the detector consists of a film sandwiched between two intensifying screens made of certain phosphors, such as calcium tungstate (CaWO4), and later rare earth phosphors such as gadolinium oxysulfide and lanthanum oxysulfide replaced CaWO4 screens with the goal of reducing the dose to the patient without compromising image quality needed for diagnosis [3, 2]. The basic steps in the production of a film-screen radiographic image are illustrated in Fig. 1.2. X-rays pass through the patient and fall upon the film to form a latent image. The latent image is then rendered visible using chemical processing and subsequently displayed on a light view-box for interpretation by a radiologist. The response of the film to x-ray exposure can be described by what is referred to as the film characteristic curve as illustrated in Fig. 1.3. The curve is a plot of the optical density (OD) to the radiation exposure (or more accurately, the logarithm of the relative exposure) used in the imaging process. The curve shows the degree of contrast or different densities that a film can display using a range of exposures. It has three elements that are significant: the toe, the slope (straight line portion), and the shoulder. While the toe and shoulder indicate underexposure and overexposure,

1.2  Film-Screen Radiography

3

Fig. 1.3  The response of the film to X-ray exposure can be described by what is referred to as the film characteristic curve, which is a plot of the optical density (OD) to the radiation exposure (or more accurately, the logarithm of the relative exposure) used in the imaging process. (From Seeram [4]. Reproduced by permission)

respectively, the slope represents the useful portion of the curve and reflects acceptable exposure or the range of useful densities. The image is black and useless if the exposure falls in the shoulder region of the curve (OD = about 3.2). If the exposure falls within the slope of the curve (OD = 0.3–2.2), the image contrast (density) will be acceptable, and this region of the curve contains the useful range of exposures [4]. The characteristic curve is an important feature of FSR that is applied to digital radiography. FSR is now obsolete and has been replaced by digital radiography, owing to its limitations. A major problem of FSR is its narrow exposure latitude, that is the correct exposure which produces acceptable images must fall within the slope of the curve. Other problems include the fact that film-screen cannot show differences in tissue contrast less than 10%, which leads to limited in its contrast resolution. The spatial resolution [sharpness of the image, measured in line pairs per millimeter (lp/ mm)] for FSR, however, is the highest of all the other imaging modalities, and can range of 5–15  lp/mm [6]. This is the main reason why radiography has been so popular over the years. Another shortcoming of FSR is that as a display medium, the optical range and contrast for film are fixed and limited. Film can only display once, the optical range and contrast determined by the exposure technique factors used to produce the image. In order to change the image display (optical range and contrast), another set of exposure technique factors would have to be used, thus

4

1  X-Ray Imaging Systems: An Overview

Fig. 1.4  The response of the digital detector to radiation exposure. See Text for further explanation

increasing the dose to the patient by virtue of a repeat exposure. Yet another problem with FSR is that as an archive medium, film is usually stored in envelopes and housed in a large room. It thus requires manual handling for archiving and retrieval by an individual. These problems have been overcome by digital x-ray imaging systems. The response of the digital detector to radiation exposure is shown in Fig. 1.4. The response is linear and shows a wide exposure latitude. Images produced by low, intermediate, and high exposures appear to have the same image quality. A consequence of this is that low exposures produce noisy images, while high exposures produce excellent image quality, but the patient receives more dose. This has posed problems for users.

1.3 Digital X-Ray Imaging Systems Digital X-ray imaging systems (also referred to as digital radiography) include modalities such as computed radiography, flat-panel digital radiography, digital fluoroscopy, digital mammography, digital tomosynthesis, and Computed tomography. Digital mammography and digital tomosynthesis will not be described in this book.

1.3  Digital X-Ray Imaging Systems

5

One of the major system components of these modalities is the digital detector. These detectors not only capture the radiation passing through the patient but also convert X-ray attenuation data (from the patient) into electronic signals (analog signals) that are subsequently converted into digital data for processing by a digital computer, the output of which is a digital image. As these imaging modalities operate in the digital domain, it is mandatory that biomedical engineering technologists understand equally as well, what is now popularly known as “imaging informatics” an area of study that involves picture archiving and communication systems (PACS), enterprise imaging, big data, artificial intelligence (AI) and its subsets, machine learning (ML), and deep learning [5]. An overview of each of these will be provided in the next subsections.

1.3.1 Computed Radiography The use of computed radiography (CR) for imaging patients in clinical practice is attributed to Fuji Medical Systems in 1983. The physical principle of CR is based on photostimulable luminescence (PSL) using photostimulable phosphors (PSP), such as barium fluoro halide (BaFX) where the halide (X) can be chlorine (Cl), bromine (Br), iodine (I), or a mixture of them, coated on a plate referred to as an imaging plate (IP) housed in a cassette, similar to FSR cassettes. There are at least six steps in the production of a CR image as illustrated in Fig. 1.5. X-ray exposure of the IP causes electrons to move from their ground state (valence band) to a higher energy level (conducting band) and are trapped there until the PSP plate is exposed to a laser light in a special unit called the image reader. Subsequently, electrons in the higher energy state return to their ground state, and in so doing, emit a bluish-purple light referred to as PSL. This light is collected and concerted into an output electrical signal and is subsequently converted into digital data to produce a CR image. The image is subject to both preprocessing and post processing to be displayed for viewing by a human observer. The IP is then exposed to a bright light to erase it (remove residual latent image) and can be used again.

Fig. 1.5  Six steps in the production of a CR image. See text for further explanation

6

1  X-Ray Imaging Systems: An Overview

Fig. 1.6  The overall steps in the production of an image from flat paned n detectors. See text for further explanation

1.3.2 Flat-Panel Digital Radiography Flat-panel digital radiography (FPDR) was invented to overcome the shortcomings of CR, most importantly, the poor spatial resolution, using detectors based on the physics of semiconductors and constructed as a flat panel. The overall steps in the production of an image from flat paned n detectors are illustrated in Fig. 1.6. First, the flat-panel detector is exposed to X-rays, hence creating an image stored on the detector matrix which is read out to produce analog signals, which are subsequently converted into digital data. Third and fourth steps involve preprocessing and postprocessing, respectively. Finally, the postprocessed image is displayed for viewing by a human observer. There are basically two types of detectors used in FPDR, namely, indirect conversion systems and direct conversion systems. While the former uses an X-ray scintillator (cesium iodide for example) coupled to amorphous silicon photoconductor; the latter uses amorphous selenium photoconductor. Indirect conversion systems first convert X-rays to light which falls upon the silicon photoconductor to produce electrical signals. Direct conversion systems convert X-rays directly into electrical signals. The signals are subsequently digitized and processed by a digital computer which produces images displayed on viewing monitors for interpretation.

1.3.3 Digital Fluoroscopy The physical principles of conventional fluoroscopy and image intensifier-based fluoroscopy have been described in the literature by several authors [4–7]. “Fluoroscopy is an imaging modality that shows anatomical structures and the

1.3  Digital X-Ray Imaging Systems

7

Fig. 1.7  An illustrates of the difference in system components of an image intensifier-based fluoroscopy system and a digital fluoroscopy system. See text for further explanation

motion of organs and the movement of contrast media in blood vessels and organs with the goal of obtaining functional information.” Digital fluoroscopy has replaced intensified-based fluoroscopy, which has been the workhorse for fluoroscopic examinations for decades. Figure 1.7 highlights the difference in system components of an image intensifier-based fluoroscopy and digital fluoroscopy. In Fig.  1.7a, the image intensifier tube is an evacuated glass envelope comprising an input screen, a photocathode, the electrostatic lens, and an output screen, generating the output fluoroscopic image. This image is captured by a television camera tube and associated optics and then transmitted via a coaxial cable to be displayed on a television monitor [4, 5]. Image intensifier-based fluoroscopy will not be further elaborated in this chapter; however, it is noteworthy to mention the problems that lead to the development of digital fluoroscopy. Examples of these problems include veiling glare, vignetting, image lag, and pincushion, and “S” distortions. Furthermore, light and electron scattering within the tube degrade the

8

1  X-Ray Imaging Systems: An Overview

image contrast (veiling glare), and image magnification results in increased dose to the patient [4–7]. Current digital fluoroscopy imaging systems use flat-panel detectors (FPDs) shown in Fig. 1.7b, to display images in real time are now used in gastrointestinal tract fluoroscopy and digital subtraction angiography (DSA). It is not within the scope of this book to describe the details of DSA; however, the interested reader may refer to Bushong [6] for a good coverage of this topic. Digital data produced by the detector goes to the computer for processing to generate digital images displayed on a monitor. In digital fluoroscopy, the FPD produces dynamic images that can be displayed and viewed in real time. For this reason, these detectors are sometimes referred to as dynamic FPDs. Dynamic FPDs are currently of two types, namely, the cesium iodide (CsI) amorphous-silicon (a-Si), thin film transistor (TFT) indirect conversion digital detector, and the a-­selenium (a-Se) TFT direct conversion digital detector. Generally, these detectors are similar in design to the static FPDs used for digital radiography imaging systems.

1.3.4 Computed Tomography: Major System Components Computed tomography (CT) is defined as a sectional imaging modality that produces sectional images of the anatomy. These images are also referred to as transverse axial images and represent sections of the anatomy that are perpendicular to the long axis of the patient. The motivation behind developing CT imaging lies in overcoming the limitations of conventional film-screen tomography [7], resulting in several advantages: (a) excellent low-contrast resolution because of the use of special data collection geometries and electronic detectors that are more sensitive than film-screen image receptors, (b) digital image processing is a central feature of CT, (c) new clinical applications such as CT angiography, CT fluoroscopy, and virtual reality CT are now possible due to new volume data acquisition methods, (d) multi-slice CT increases the volume coverage speed performance, so that larger patient volumes can be scanned faster without compromising image quality, and (e) 3-dimensional (3D) imaging. The CT scanner was invented by Godfrey Hounsfield in the United Kingdom, who shared the Nobel Prize in Medicine and Physiology in 1979, with Allan Cormack, a physicist in South Africa who later worked at Tufts University in Massachusetts. The evolution of the CT scanner has progressed single-slice CT scanners (SSCT) to state-of-the-art multi-slice CT scanners (MSCT). Over the years, technical innovations have led to the increased utilization of CT as a diagnostic tool. As a result, it is well-documented fact that CT delivers the highest collective dose in the United States compared to other medical imaging modalities [6, 7]. The CT scanner comprises three major system components (Fig. 1.8) in the production of CT images. These include the data acquisition, image reconstruction, and

1.3  Digital X-Ray Imaging Systems

Detectors

9

Image Reconstruction

Image Display, Storage, Communications

Attenuation readings

Data Acquisition Fig. 1.8  The CT scanner consists of three major system components. These include the data acquisition, image reconstruction, and image display, storage, and communication

Fig. 1.9  Data acquisition and associated processes in CT imaging. As shown an X-ray tube coupled to a 2-dimensional electronic detector array, rotate 360° around the patient to collect and measure radiation attenuation data as the X-ray beam passes through the patient. See text for further explanation

image display, storage, and communication. The processes related to data acquisition are illustrated in Fig. 1.9. In this illustration, X-ray tube coupled to a 2D electronic detector array rotates 360° around the patient, collecting and measuring radiation attenuation data as the X-ray beam passes through the patient. The collected data are converted into integers (0, a positive number, or a negative number), referred to as CT numbers, utilizing an image reconstruction algorithm to build up the image in numerical format. These CT numbers (numerical image format) are further transformed into a gray-scale image, which is displayed on a monitor for the observer to interpret. The CT numbers are calculated using the following relationship: CT Number 

tissue  water ·K  water



10

1  X-Ray Imaging Systems: An Overview

where K represents a scaling factor. The CT scanner and its system components will be described in detail in Chap. 6.

1.4 Radiation Physics at a Glance Physics is an essential subject in all engineering technology programs, as medical devices are based not only on engineering constructs but also on physics principles that describe how devices work. This text will focus on relevant physical principles of X-ray imaging systems that produce X-rays and create diagnostic images that are used in the care and management of the patient. In X-ray imaging, the scope of radiation physics is wide and varied, and for this reason, this text will address topics such as X-ray generation, X-ray production, X-ray emission, X-ray quantity and quality, X-ray attenuation, and X-ray interaction with matter. These topics will be described in a little more detail in Chap. 2. Additionally, there are applied physics topics relevant to biomedical engineering technologists, especially concerning radiation-­emitting machines. These fall under the subject matter that deals specifically with radiation protection concepts and principles, outlined briefly in Sect. 1.7. X-ray generation uses a specialized electrical circuit to provide electrical power to the operator’s console and to the X-ray tube. This circuit consists of a low voltage section and a high voltage section. While the low voltage circuit consists of electrical components that provide power to the operator’s control console, the high voltage section consists of a high-voltage generator circuit and the X-ray tube circuit. The operator’s console provides an interface between the electronics of the imaging system and the patient and allows the technologist to select and control radiographic exposure factors such as the kilovoltage (kV), the milliamperes-seconds (mAs), and the exposure time in seconds required for the examination. The high voltage circuit includes electrical components such as rectifiers and transformers, to provide the electrical power needed to produce X-rays. X-ray production deals with how X-rays are produced in the X-ray tube. The X-ray tube consists of a cathode and an anode enclosed in an evacuated envelope which houses these components. A major component of the cathode is a filament which when heated with current emits electrons. These electrons are accelerated to high speeds to strike a target located on the anode. The result is the production of X-rays used for imaging the patient. X-ray emission refers to the X-ray photons emitted from the anode of the X-ray tube. A graph showing the number of photons (X-ray intensity) plotted as a function of X-ray energy is called the X-ray emission spectrum. There are two X-ray emission spectra, the continuous spectrum and the characteristic or discrete spectrum. Understanding these spectra provides a clear insight on the effects of kV and mAs on radiation dose to patients, for example, among other technical considerations, such as X-ray quantity and X-ray quality. X-ray quantity and X-ray quality are sometimes referred to as the intensity (I) of the beam from the X-ray tube. While X-ray quantity is the number of photons per

1.4  Radiation Physics at a Glance

11

unit energy, X-ray quality refers to the energy or penetrating power of the X-ray photons in the beam. The main controlling factors for X-ray quantity and quality are the mAs and kV, respectively. The higher the kV, the greater the beam quality. Higher kV will also affect the beam intensity (I) through the following algebraic expression:

I α KV2

This implies that as the kV is doubled, the intensity will increase by a factor of four. Intensity problems can be solved using the following algebraic expression: 2



I1  kV1    I 2  kV2 

where I1 and I2 are intensities at kV1 and kV2, respectively. The flow of electrons from the filament to the target of the X-ray tube is called the tube current (mA) which determines the quantity of photons from the X-ray tube when an exposure is made. The higher the mA, the higher the number of electrons and the greater the quantity of X-rays. X-ray quantity is directly proportional to the mA, algebraically expressed as follows:

I α mA

Doubling the mA doubles the quantity of X-ray photons, and therefore, more radiation dose to the patient. There are other factors affecting X-ray quality and quantity, such as the target material, filtration, and the voltage waveforms; however, these are not under the direct control of the operator. X-ray attenuation is defined as a reduction of the intensity of the X-ray beam as it passes through the patient. Such reduction is as a result of the photons interacting with the tissues of the patient. X-ray interaction with matter include five mechanisms, namely, (1) classical or Rayleigh scattering, (2) Compton scattering, (3) photoelectric absorption or photoelectric effect, (4) pair production, and (5) photodisintegration. The details of first three mechanisms will be described in Chap. 2. However, the latter two mechanisms, pair production and photodisintegration, occur at very high energies beyond those used in diagnostic imaging. As such, they will not be covered in this book.

12

1  X-Ray Imaging Systems: An Overview

1.5 Imaging Informatics: An Overview Imaging informatics is a growing body of knowledge leading to careers in medical imaging and biomedical engineering technology fields, reflecting the dominance of digital domain in healthcare professions. The shift to digital operations in hospitals and medical imaging departments has given rise to the field of imaging informatics. The Society of Imaging Informatics in Medicine (SIIM) defines imaging informatics as “the study and application of processes of information and communications technology for the acquisition, manipulation, analysis, and distribution of image data.” The literature extensively covers imaging informatics [4–8], identifying core topics ranging from information and communication technologies, picture archiving and communication systems (PACS), radiology information systems (RIS), hospital information systems (HIS), the electronic health record, to cloud computing, big data, AI, and its subsets ML, and DL. A brief overview of these topics is provided by Seeram [5] as follows: • “Information technology (IT) refers to the use of computers and computer communications technologies to not only to process, store, secure electronic data but also to communicate these data using computer networking infrastructure.” PACS is an excellent example of an informatics-rich medical device. • PACS includes the imaging image acquisition modalities such as digital radiographic and CT modalities, a computer network database server, storage and archival systems, and display workstations. PACS may be connected to information systems such as the RIS and the hospital information system (HIS). Furthermore, these systems must be fully integrated and secured for efficient and effective data interchange. PACS is now referred to as MIMPS (medical image management and processing system), a recent change made by the US Food and Drug Administration. This will be described further in Chap. 8. • Cloud computing simply provides a means of using the Internet for storage and retrieval (for example) of data using specific software packages. Additionally, emerging topics which will have an impact on the practice of medical imaging include big data, AI, ML, and DL. • Big data is characterized by four Vs: Volume, Variety, Velocity, and Veracity. While Volume refers to the very large amount of data, Variety deals with a wide array of data from multiple sources. Furthermore, Velocity addresses the very high speeds at which the data are generated. Finally, Veracity describes the uncertainty of the data such as the authenticity and credibility. • AI uses computers in an “effort to automate intellectual tasks normally performed by humans.” A subset of AI is ML which includes “a set of methods that automatically detect patterns in data, and then utilize the uncovered patterns to predict future data or enable decision making under uncertain conditions.” A subset of ML is DL which uses algorithms that are “characterized by the use of neural networks with many layers.”

1.7  Radiation Protection Considerations in X-Ray Imaging

13

These emerging technologies will evolve and more importantly become useful tools in medical imaging. “Therefore, students and technologists alike should make every effort to grasp their meaning and applications so that they can communicate effectively with radiologists and medical physics in an effort to participate actively in the management of patient care” [5].

1.6 Quality Assurance and Quality Control Quality assurance (QA) and quality control (QC) are essential tools that are vital to the daily operation of the medical imaging department. As outlined by several authors [4–6], QA addresses management and administrative efforts to ensure the efficient and effective care of the patient. It therefore deals with people. QC on the other hand deals with the technical aspects of medical imaging equipment performance used to image patients. QA and QC programs have evolved into what is now referred to as continuous quality improvement (CQI) which includes total quality management (TQM) [6]. There are at least three important steps in a QC program, namely, acceptance testing, routine performance, and error correction. While acceptance testing ensures that the equipment meets the specifications set by vendors, routine performance requires users to perform the actual QC test on the equipment. Finally, error correction simply requires that if the equipment fails to meet the specified acceptance criteria or the tolerance limit of the QC test then it must be repaired in order to meet the performance standards. An example of the tolerance limit for X-ray beam collimation is it (the beam) should be ±2% of the source-to-image receptor distance. The American Association of Physicists in Medicine (AAPM) has established a set of detailed QC tests for digital radiography. For example, examples of these tests include physical inspection of imaging plate (IP), dark noise and uniformity, exposure indicator (EI) calibration, laser beam function, spatial accuracy, erasure thoroughness, aliasing/grid response, and positioning and collimation errors, among others. The biomedical engineering technologist may have to conduct QC tests in the radiology department in the absence of QC technologists or QC medical physicists.

1.7 Radiation Protection Considerations in X-Ray Imaging An in-depth understanding of radiation protection is mandatory for anyone working with X-ray emitting devices, including biomedical engineering technologists. Radiation protection covers a wide scope of topics ranging from radiobiology to physical principles and concepts, such as radiation quantities and their units, radiation protection criteria and standards, dose limits for occupationally exposed individuals and members of the public, and diagnostic reference levels. Furthermore,

14

1  X-Ray Imaging Systems: An Overview

radiation protection regulations and guidelines are a major activity of not only international organizations, such as the ICRP, but also national organizations, such as the National Council on Radiation Protection and Measurements in the United States, and Radiation Protection Bureau-Health Canada (RPB-HC). Radiobiology is the study of effects of ionizing radiation on biologic systems leading to whole-body biological effects generally categorized as stochastic effects and deterministic effects. Stochastic effects are those for which the probability of the effect increases as the dose increases and for which there is no threshold dose. Examples of these effects include life-span shortening, radiation-induced malignancy, and hereditary effects (late effects that occur in the offspring of the irradiated individual). These effects are also called late effects, since they occur years after the exposure of the individual. Deterministic effects on the other hand are those effects for which the severity of the effect depends on the dose. These effects have a threshold dose and increase with increasing dose. These effects are called early effects since they can occur within minutes, hours, days, weeks, and months after the exposure. Radiation protection has two important objectives, that is, to prevent deterministic effects and minimize the probability of stochastic effects. To accomplish these goals, radiation protection is guided by two triads; radiation protection principles and radiation protection actions. While the former deals with the ICRP’s principles of justification, optimization, and dose limitation, the latter addresses the triad of time, shielding, and distance [5–7, 9, 10]. • Justification requires that there should be a net benefit associated with each and every exposure use in a patient who receives a diagnostic X-ray examination. • Optimization means that the lowest dose should be used for an examination but the image quality must not be compromised. The ICRP refers to this as the ALARA (as low as reasonably achievable) philosophy. • The notion of dose limits has been established for those who are occupationally exposed, and members of the public, as well, such as students in training. These limits are intended to minimize the risks of the stochastic effects of radiation. Time, shielding, and distance constitute the second triad of radiation protection. These are also referred to as radiation protection actions. A summary of these actions is as follows [9, 10]: • The time of exposure to radiation should be kept as short as possible short as possible, since the relationship between time and exposure is proportional; that is, if the time of exposure doubles, the exposure doubles. • Shielding is used to protect patients and related individuals through the use lead shields and aprons. X-ray room walls are also shielding with appropriate materials such as concrete or lead to prevent exposure of members of the public who are in a waiting room. • Distance is used as to protect individuals based on the inverse square law, which states that the exposure is inversely proportional to the distance, that is, the further away individuals are from the source of the radiation, the less exposure they will receive.

References

15

The Diagnostic Reference Level (DRL) is a concept used to address the limits of exposure for patients. DRLs are not equivalent to dose limit s for occupationally and non-occupationally exposed individuals [9, 10]. A formal definition of the DRL given by the American College of Radiology (ACR) is that a DRL is “an investigation level to identify unusually high radiation dose or exposure levels for common diagnostic medical x-ray procedures [11].” DRLs are tools that radiology departments can use to measure and assess radiation doses to patients for a defined set of procedures. If the doses delivered are consistently greater than established DRLs for that facility’s country or region, then the department should be concerned about its radiation protection procedures, investigate why exposures are beyond the established DRLs, and take corrective action [5]. Radiation Protection also focusses on several other relevant topics such as radiation quantities and their associated units. These include exposure, absorbed dose, and effective dose (ED). The units associated with each of these include coulombs per kilogram (C/kg), Grays (Gy), and Sieverts (Sv), respectively [6, 7, 9, 10]. Another significant topic is that of radiation protection regulations and guidelines, as identified in the first paragraph in Sect. 1.7.

References 1. What is a biomedical engineering technologist: The Canadian Medical and Biological Engineering Society (CMBES). 2022. https://www.cmbes.ca/about/what-­is-­a-­biomedical-­ engineering. Accessed 30 Sept 2022. 2. Buchanan RA, Finkelstein SI, Wickersheim KA.  X-ray exposure reduction using rare-earth oxysulfide intensifying screens. Radiology. 1972;105(1):84–92. 3. de Carvalho A, Jørgensen J. Rare-earth and calcium tungstate intensifying screens, a comparative study of relative speed, radiation doses and resolving power. Rofo. 1978;128(3):358–63. https://doi.org/10.1055/s-­0029-­1230861. 4. Seeram E.  Digital radiography: physical principles and quality control. 2nd ed. Singapore: Springer Nature; 2019. 5. Seeram E.  A comprehensive guide to radiographic sciences and technology. Chichester: Wiley; 2021. 6. Bushong S. Radiologic science for technologists. 12th ed. St Louis: Elsevier-Mosby; 2022. 7. Bushberg JT, Seibert JA, Leidholdt EM Jr, Boone JM. The essential physics of medical imaging. 4th ed. Philadelphia: Lippincott Williams and Wilkins; 2022. 8. Seeram E. Computed tomography: physical principles, patient care, clinical applications, and quality control. 5th ed. St Louis: Elsevier; 2023. 9. Seeram E. Rad Tech’s guide to radiation protection. Hoboken: Wiley; 2020. 10. Seeram E, Brennan P. Radiation protection in diagnostic X-ray imaging. Burlington: Jones & Bartlett Learning; 2017. 11. ACR-SPR (American College of Radiology-Society for Pediatric Radiology). ACR-SPR practice parameter for imaging pregnant or potentially pregnant adolescents and women with ionizing radiation. Reston: American College of Radiology; 2018.

Chapter 2

Basic Radiation Physics for Diagnostic X-Ray Imaging

2.1 Introduction Biomedical engineering technologists (BMETs) not only perform maintenance services on X-ray imaging systems but may also conduct quality control testing of major components of the system. In this regard, BMETs must be fully aware of the basic radiation physics concepts involved when energizing the X-ray system for these tasks. This purpose of this chapter is to describe the fundamental principles of X-ray generation, X-ray production, X-ray emission, interaction of radiation with matter, and radiation attenuation. These topics were briefly introduced in Chap. 1.

2.2 X-Ray Generation As discussed in Chap. 1, X-ray generation relies on a specialized electrical circuit (the X-ray generator circuit) to supply electrical power to the operator’s console and the X-ray tube. The physics of the X-ray generation circuit falls mainly within the scope of electricity and magnetism principles, since these are topics reserved for special BMET courses addressing the nature and scope of electrical circuits. In this section, a broad overview of the requirements for X-ray generation will be highlighted. X-rays are generated when high-speed electrons strike a target. These electrons and target are significant parts of the X-ray tube, which is strategically positioned in the electrical circuit. The circuit consists of several electrical components and is divided essentially into two major parts: the low voltage section, and the high voltage section, which contains the X-ray tube requiring high voltage X-ray production. The electrical power supply to the X-ray circuit from the electrical utility company is alternating current (AC) with a frequency of 60  cycles per second or 60 © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_2

17

18

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

Hertz (Hz), in North America. Furthermore, AC power can be classified as single-­ phase AC power and three-phase AC power. While single phase AC power produces an AC waveform consisting of a positive half cycle and a negative half cycle, three phase AC power has three separate AC input lines producing three separate waveforms that have the same frequency and voltage amplitude that are out of phase by one-third of a cycle or by 120°. The production of X-rays is much more efficient with a three-phase generator compared with a single-phase generator, thus enabling greater tube output and the use of shorter exposure times for more sophisticated X-ray examinations. The low voltage section of the circuit provides electrical power to the operator’s console. Low voltage from the alternating current (AC) mains supply to the X-ray room is delivered to an autotransformer (based on the principle of electromagnetic induction), which is tapped appropriately to select and control the radiographic exposure technique {kilovolts (kV), milliamperes (mA), and time (in seconds)} needed for the examination. While the kV selection and timing circuits are coupled to the high voltage transformer, the mA selector circuit is coupled to a step-down transformer that controls the selection of small or large focal spot sizes of the X-ray tube, to be described subsequently. The high voltage circuit includes electrical components such as rectifiers and high-voltage and filament transformers (contained in a tank and immersed in oil for electrical insultation) to provide the electrical power needed to produce X-rays. The purpose of the high-voltage transformer is to increase the low-voltage {volts (V)} input from the autotransformer to high voltage (kilovolts) needed for x-ray production. Since transformers operate with AC, rectifiers are used in the circuit to convert the AC into direct current (DC), which is supplied to the X-ray tube in order to produce X-rays. The filament transformer is a step-down transformer, which requires relatively low voltage and high current (about 10 V and 5 A, respectively). It is important to note that the mA meter is positioned in the high voltage circuit and is connected near the electrical ground (center) of the secondary winding of the high-voltage transformer. The mA meter shows the tube current in milliamperes (mA) and reflects the flow of high-speed electrons across the X-ray tube when a radiographic exposure is made. This is done so that the mA/mAs meter can be placed on the operator’s console, to ensure electrical safety. In diagnostic x-ray imaging systems, rectifiers play a crucial role, and the ones employed are solid-state, semi-conducting materials, such as pure silicon. They are specifically designed to allow electrons to flow from cathode to anode in the X-ray tube. Rectification can take two forms: half-wave rectification or full-wave rectification. The former is no longer utilized in X-ray imaging systems due to it limited X-ray output. Conversely, The full-wave rectification overcomes this limitation by employing the negative half of the AC waveform for X-ray production. This results in X-rays being produced at 120 pulses per second, as opposed to 60 pulses per second in half-wave rectification. Full-wave rectified units produce X-rays more efficiently with shorter exposure times. X-ray generators fall into four categories: single-phase generators, three-phase generators, constant potential generators, and high-frequency generators. Each

2.3  X-Ray Production

19

generator type produces a characteristic voltage waveform or ripple that plays a vital role in the production of the X-ray intensity suitable for different imaging procedures. However, the intricate details of each generator type are beyond the scope of this chapter. For more comprehensive information on the nature and scope of X-ray generators, readers may refer to Bushong [1] and Bushberg et  al. [2]. Nevertheless, the following points are noteworthy: • The high-frequency generator is state-of-the-art generator for use in radiology. The high-frequency generator consists of a number of electrical components that perform the following functions as summarized by Seeram [3]: –– First, low-frequency (60  Hz), low-voltage input is converted into a high-­ frequency (500–25,000  Hz), low-voltage waveform and subsequently to a high-frequency, high-voltage output waveform that travels to the X-ray tube. The waveform to the X-ray tube is almost constant, with a ripple of less than 3%, resulting in a more efficient method of X-ray production with an increase in both X-ray intensity and beam energy. –– Second, the high-frequency generator can use either a single-phase or a three-­ phase power supply. However, the three-phase input lines produce increased power. –– The third point to note is that the high-frequency, high-voltage output from the high-voltage transformer is first rectified and subsequently smoothed by high-voltage capacitors to produce nearly constant voltage to the X-ray tube. High-frequency generators are inexpensive and compact and they are used in portable X-ray units and in modern CT scanners. Other notable advantages of these generators include less radiation dose to patients, providing almost constant voltage to the tube. Finally, an important consideration when describing X-ray generators is the power rating of the generator, which is given in kilowatts (kW). For example, while general X-ray imaging generators typically have power ratings ranging from 30 to 50 kW, special angiography and interventional imaging applications demand power ratings up to about150 kW [1].

2.3 X-Ray Production X-ray production occurs when high-speed electrons strike a target. In X-ray imaging, this occurs in an X-ray tube. The major components of an X-ray tube are shown in Fig. 2.1, all of which are enclosed in an evacuated envelope. The major components of an X-ray tube include the cathode assembly and the anode assembly. The cathode assembly contains a filament that emits electrons when heated with current. These electrons are accelerated to high speeds to strike a target located on the anode, by applying a high voltage (kV) between the negative cathode and positive anode. The result is the production of X-rays used for imaging the patient.

20

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

Tube envelope

Rotating Anode Disk

Stator

+ Rotor



Cathode assembly

Stator

Hot filament Electron stream Anode Target or Focal spot X= Ray Beam

Fig. 2.1  The major components of an X-ray tube. See text for further explanation

Tungsten is used for the filament wire due to its high melting point (3410 °C) and its high atomic number (Z  =  74). The filament is wound in a helical fashion to increase the surface area so that more electrons can be produced. When the filament is heated (3–6 A), electrons “boil-off” through the physical process known as thermionic emission. The electrons are focused to strike the target on the anode, to produce X-rays. Vaporization of the tungsten has been identified as the main cause of X-ray tube failure, as it causes a buildup in the inside of the by envelope resulting in electrical arcing during use [1]. The solution to this problem is to coat the tungsten wire with thorium (thoriated tungsten) to prolong the life of the tube. There are two types of x-ray tubes, the stationary anode tube and the rotating anode tube. The stationary anode is made of a copper block in which a rectangular piece of tungsten is embedded. The target area struck by electrons is referred to as the focal spot. The anode is inclined at an angle, called the target angle, to direct the radiation beam to the patient. Stationary anode tubes may have target angles ranging from 5° to 15°. There are two focal spots, a small and a large focal spot which can withstand greater electrical loads (kV, mA, and time in seconds (mAs) and have higher heat capacities compared with smaller focal spots. Furthermore, stationary anode tubes are limited in their X-ray output (intensity) and heat loading and are used in dental and portable fluoroscopy units. Rotating anode X-ray tubes overcome these limitations. A central feature of a rotating anode tube is the anode disk. As seen in Fig. 2.2, the disk is supported by a molybdenum stem, and its diameter can range from 50 to 200 mm. Modern disks are made of two or more metals such as rhenium (R), zirconium (Z), and titanium (T) used in conjunction with tungsten (T), molybdenum (M), and graphite. The

2.3  X-Ray Production

21

Fig. 2.2  Electron–atom interactions that produce characteristic and Bremstrahlung X-ray photons. (From Sprawls, P. Physical Principles of Medical Imaging. Distinguished Emeritus Professor, Emory University Director, Sprawls Educational Foundation, http://www.sprawls.org Co-Founding Editor, Medical Physics International journal http://www.sprawls.org/ppmi2/XRAYPRO/ Accessed August 2023. Reproduced by permission)

RTM disk for example is made up of 10% rhenium and 90% tungsten (coating layer). Compared to pure tungsten disks, the advantages of compound anode disks include lesser rotational problems because of the lighter weight, resistance to the aging, more heat storage capacity, less roughening of the target track, and high and uniform dose. Rotating anode disks feature are equipped with two focal spot sizes, namely large and small serving distinct roles. The small focal spot contributes to the production of sharp images, while the large focal spot is employed to handle larger patients and more sophisticated examinations, necessitating increased electrical load in terms of mAs and kV. Additionally, the target angles can vary between 5° and 15°. To achieve the necessary rotation, an induction motor is employed, enabling the anode disk to rotate at speeds ranging from 3600 to 10,000 revolutions per minute (rpm). The x-ray tube envelope is meticulously designed to support internal components, such as the anode and cathode structures, ensuring a vacuum to prevent the oxidization of the filament. Finally, the tube insert is encased in a protective housing called the tube housing. This housing serves multiple purposes, providing mechanical support for the insert, along with radiation and electrical insulation. Oil is used between the envelope and the inner walls of the housing to insulate and protect the housing from the high voltage applied to the tube. The oil also helps with heat dissipation. A significant feature

22

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

of the housing is its internal lining made of lead, which prevents radiation from “leaking” through the housing. In accordance with radiation protection standards, the housing must limit leakage radiation to less than 26 μC/kg – hour (100 mR/h) at a distance of 1 m from the X-ray tube [1].

2.3.1 Characteristic Radiation Figure 2.2 illustrates the generation of characteristic and Bremsstrahlung x-rays. While the former is produced when high-speed electrons from the filament interact with inner shell electrons of the anode target atoms, the latter is produced when high-speed electrons interact with the nuclear force field (not the nucleus) of the target atom. According to Sprawls [4], the interaction can occur only if the incoming electron has a kinetic energy greater than the binding energy of the electron within the atom. When this condition is met and a collision occurs, the electron is dislodged from the atom, creating a vacancy that is filled by an electron from a higher energy level. As this filling electron moves down to occupy the vacancy, it releases energy in the form of an X-ray photon. This phenomenon is termed characteristic radiation because the energy of the photon is characteristic of the chemical element serving as the anode material. In the provided example (Fig. 2.2), the dislodged electron originates from a tungsten K-shell electron with a binding energy of 69.5 keV. The vacancy is then filled by an electron from the L shell with a binding energy of 10.2  keV.  Consequently, the energy of the characteristic X-ray photon equals the energy difference between these two levels, amounting to 59.3 keV. X-rays emitted due to the filling of the K-shell are called K-characteristic X-rays, while those resulting from the filling of the L-shell are known as L-characteristic X-rays, and so on. Characteristic X-rays are only useful in diagnostic radiology if they have enough energy to penetrate body tissues and fall upon the image detector. Therefore, characteristic X-rays are best used in soft tissue radiography such as X-ray mammography. Characteristic radiation mainly from k-shell interactions are used to image the breast not only to show small differences in breast tissues but also to obtain maximum image contrast [1].

2.3.2 Bremsstrahlung Radiation Figure 2.2 illustrates the production of Bremsstrahlung radiation. The word “Bremsstrahlung” originates from the German words for “braking” or “slowing down.” This phenomenon occurs when a high-speed electron interacts with the nuclear force field of the nucleus, slowing down or decelerating in the process. As the electron loses its initial energy (kinetic energy = KE) and changes its direction of travel, with less KE, the difference in KE reappears in the form of Bremsstrahlung

2.4  X-Ray Emission

23

radiation (Fig. 2.2). An important feature of Bremsstrahlung radiation is that it has a wide range of energies, which are useful for imaging most body parts in diagnostic X-ray imaging. This range of energies is due to the fact that high-speed electrons from the filament can lose varying amounts of KE as they pass by the nucleus, such that electrons closer to the nucleus will generate high-energy X-rays, compared with electrons that are farther away from the nucleus which results in low-energy X-rays. Bremsstrahlung radiation with energies ranging from 0 to 80 keV is as a result of an incoming high-speed electron having a KE of 8 keV.

2.4 X-Ray Emission The term “X-ray emission” is used to characterized the nature of the X-ray photons emitted from the target of the anode of the X-ray tube, and the number of photons plotted as a function of photon energy is referred to as the X-ray emission spectrum. Figure 2.3 shows two types of X-ray emission spectra, the Bremsstrahlung or continuous spectrum and the characteristic or discrete spectrum. While the Brems spectrum shows that the energy of the photons ranges from 0 to a maximum value, the characteristic spectrum shown photons are emitted only at specific energies. There are several points to note about the continuous spectrum that are important in clinical practice: 1. The maximum energy of the photons is equal to the kVp used by the technologist for the particular examination and is the reason why it is referred to as the kVp(eak) [1].

Fig. 2.3 The X-ray emission spectrum shows two types of X-ray emission spectra, the Bremsstrahlung or continuous spectrum and the characteristic or discrete spectrum

24

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

2. The photons emitted from the tube during a radiographic exposure is generally referred to as the X-ray intensity (I), which includes X-ray quantity and X-ray quality. 3. The area under the curve of the continuous spectrum is referred to as the X-ray quantity (number of photons per unit energy) which is controlled by the mAs used for the examination. The X-ray quantity is directly proportional to the mA, through the following expression: I α mA. If the mA is doubled, the intensity doubles. 4. The X-ray quality refers to the energy or penetrating power of the X-ray photons in the beam. The main controlling factor for X-ray quality is the kV used for the examination. The higher the kV, the greater the beam quality. Higher kV will also affect the beam intensity since it is proportional to the square of the kV change, expressed as I α kV2. This implies that as the kV is doubled, the intensity will increase by a factor of four. 5. mA and kV must therefore be taken into consideration when selecting these factors for imaging, since they both play a role in the dose delivered to the patient. Various factors influence the quality and quantity of x-rays, and while the anode target material, beam filtration, and the voltage waveforms play significant roles, they are not directly controlled by the operator. For an in-depth understanding of how these factors impact the quality and quantity of the X-ray beam, readers referred to Seeram [3].

2.5 X-Ray Attenuation Attenuation is defined as a reduction of the intensity of the x-ray beam as it passes through the patient and is the result of X-ray photons interacting with the tissues of the patient. To fully understand attenuation, it is essential to review at least three parameters, namely, the linear attenuation coefficient, the mass attenuation coefficient, and the half-value layer. The linear attenuation coefficient (μ) is “the fraction of photons lost from the X-ray beam in travelling a unit of distance, measured in cm−1” [4]. The number of photons (n) removed from the original number of photons (N) entering the object of a thickness Δx can be calculated using the following expression:

n   N x

However, as the thickness increases, the above expression cannot be used, and therefore another expression is needed. This algebraic expression for a homogeneous beam of radiation (all photons have the same energy) is as follows:

N  No e x

2.5  X-Ray Attenuation

25

Where No is the original number of incident photons and N is the transmitted number of photons through a thickness x, without interaction [2]. There are several factors influencing μ including the atomic number (Z) and density of the absorber, the energy of the photons in the beam passing through the absorber, and the electrons per gram (g) of absorbing material. For example, if the Z, density, and electrons/gm of the absorber increase, attenuation increases, as the number of transmitted photons decreases. If the kV is increased from say 80 to 100  kV, the number of transmitted photons increases, hence the attenuation decreases. These facts are important when imaging patients, in order to obtain excellent radiographic image quality with minimum dose to the patient. The mass attenuation coefficient is another method to address attenuation in X-ray imaging; however, it is not used routinely. The mass attenuation coefficient is the linear attenuation coefficient (μ) divided by the mass density (ρ) of the material. Mass density is defined as the quantity of material per unit volume and its units are grams/cubic centimeter (g/cm3) or kilograms/cubic meter (kg/m3). “in radiology, we do not usually compare equal masses. Instead, we usually compare regions of an image that corresponds to irradiation of adjacent volumes of tissue. Therefore, density, the mass contained within a given volume, plays an important role. Thus, one can radiographically visualize ice in a cup of water due to the density difference between the ice and the water” [2]. The attenuation can be calculated as follows:

Attenuation     t     /  



where t is the thickness of the absorber. Finally, the half-value layer (HVL) is another parameter that affects attenuation in X-ray imaging, and it represents a simple measure of the energy of the X-rays emitted from the X-ray tube. The HVL of the X-ray beam is defined as “that thickness of absorbing material necessary to reduce the X-ray intensity to half of its original value” [1]. In diagnostic radiography, the absorbing material is aluminum (Al) with a thickness that varies from 3 to 5 mm Al. The significance of the HVL is that it provides information with respect to the energy of the X-ray beam and represents that “most appropriate” tool to describe the beam energy. Beam energy also depends on the kV and the filtration. While the kV affects the penetrability of the beam, that is beam quality (energy of the photons); the filtration affects the beam quantity (number of photons), as well as increases the mean energy of the beam. As the kV increases, the beam energy increases, and hence, the HVL increases as well. For example, the HVLs for 50 kV, 75 kV, and 100 kV, are 1.0 mm Al, 2.8 mm Al, and 3.7 mm Al, respectively [1]. The X-ray beam emitted from an X-ray tube is a heterogeneous beam, that is, the beam consists of both low and high energy photons. Filtration on the other hand is intended to protect the patient from unnecessary radiation by removing the low energy photons from the beam, thus increasing the average energy of the beam. Beam intensity is subsequently reduced. Al is generally used in diagnostic X-ray tubes, however, in specialized X-ray tubes, copper, tin, gadolinium, and holmium have been used as filters.

26

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

Fig. 2.4  The total filtration of a diagnostic X-ray tube is the sum of inherent filtration and added filtration

The total filtration (Fig. 2.4) of a diagnostic X-ray tube is the sum of inherent filtration and added filtration. Inherent filtration, approximately 0.5 mm Al equivalent, arises from the metal envelope and exit window of the X-ray tube. Added filtration involves a thin sheet of Al (about 1.0 mm Al equivalent) placed between the X-ray tube housing and the collimator, and an additional 1.0 mm Al equivalent from the mirror in the collimator box, providing a visual cue of the X-ray beam. The total filtration is typically 2.5  mm Al equivalent, a legal requirement stipulated in International and National X-ray Safety Guidelines and Recommendations.

2.6 Interaction of Diagnostic X-Rays with Matter There are five mechanisms by which X-rays can interact with matter, depending upon not only the energy of the photons which can range from low, moderate, to high energies, but also the composition of the absorbing material. These include classical or Rayleigh scattering, Compton scattering, photoelectric absorption or photoelectric effect, pair production, and photodisintegration [1–4]. Since pair production and photodisintegration occur at very high energies beyond those used in diagnostic X-ray imaging, they will not be described in this book; however, the interested reader may refer to Bushong [1] and Bushberg et al [2] for descriptions of these two processes. In diagnostic X-ray imaging, when a beam of radiation falls upon the patient, some of the photons of the beam will be absorbed, some will be scattered, and some will penetrate the patient (Fig.  2.5) and strike the image receptor, to produce the image.

2.6  Interaction of Diagnostic X-Rays with Matter

27

Fig. 2.5  In diagnostic X-ray imaging, when a beam of radiation falls upon the patient, some of the photons of the beam will be absorbed, some will be scattered, and some will penetrate the patient and fall upon the image receptor, to produce the image

2.6.1 Mechanisms of Interaction in Diagnostic X-Ray Imaging The three mechanisms of X-ray interaction to be described in this section are classical scattering, Compton scattering, and photoelectric absorption. The probability (chance) of either of these three interactions occurring depends on several factors, including the atomic number (Z) of the absorbing material and the energy of the photons.

2.6.2 Classical Scattering Classical scattering or Rayleigh scattering involves the interaction between a low energy photon (about 10 keV) and the atom. As a result, the photon is absorbed by the entire atom leading to excitation, that is the entire atom becomes excited. There is no ionization. The result of excitation is the release of a photon with energy equal to that of the incident photon and scattered in a different direction. This interaction is very small (less than 5%) and therefore has little influence on image formation in diagnostic x-ray imaging. When an exposure in made in diagnostic x-ray imaging, two interactions occur, Compton scattering and photoelectric absorption. The one that predominates depends on the kV used for the examination.

28

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

Fig. 2.6  In Compton scattering, an incident photon interacts with electrons in the outer shell of the atom and subsequently ejects the electron from its shell. The incident photon is scattered in a different direction, as defined by the angle of deflection. Furthermore, the scattered photon energy is less than the incident photon energy

2.6.3 Compton Scattering Compton scattering is an important interaction in diagnostic X-ray imaging. As illustrated in Fig. 2.6, an incident photon interacts with electrons in the outer shell of the atom and subsequently ejects the electron from its shell. The incident photon is scattered in a different direction, as defined by the angle of deflection. Furthermore, the scattered photon energy is less than the incident photon energy. The probability that Compton scattering will occur depends on not only the energy (E) of the incident photon but also on the electron density (number of electrons per gram) of absorber. As the energy of the incident photon and the density of the absorber increase, the probability of Compton scattering increases. The probability of Compton scattering does not depend on the atomic number of the absorber. Generally, Compton scattering predominates at higher kV techniques (lower tissue contrast) and photoelectric absorption predominates at lower kV techniques (higher tissue contrast) [1–4]. While higher kV exposure techniques result in less dose to

2.6  Interaction of Diagnostic X-Rays with Matter

29

Fig. 2.7  In photoelectric absorption, an incoming photon interacts with inner shell electrons (K or L) and transfer all of its energy to these electrons. The electron absorbs this energy and is ejected from the atom. The ejected electron is referred to as a photoelectron

the patient, more scattered radiation is produced which tends to destroy image contrast.

2.6.4 Photoelectric Absorption The crucial interaction of X-rays with matter in diagnostic X-ray imaging is photoelectric absorption, also known as photoelectric effect. Photoelectric absorption is illustrated in Fig. 2.7, where an incoming photon interacts with inner shell electrons (K or L), transferring all of its energy to these electrons. The electron absorbs this energy and is ejected from the atom. The ejected electron is referred to as a photoelectron. The resulting vacancy now left in the inner shell is filled by an electron in the neighboring outer shell, resulting in characteristic X-ray emission. The energy of the photoelectron is equal to the energy of the incident photon minus the binding energy of the ejected electron [1, 2, 4]. As described by Seeram [3], the probability of occurrence of photoelectric absorption depends on the atomic number of the absorber and the energy of the incident photon and is greatest with high atomic number elements and low energy radiation. Thus, bone has an effective Z of 13.8 and will absorb more radiation than soft tissue, which has an effective Z of 7.4. Furthermore, because the probability of photoelectric absorption is inversely proportional to the energy of the incident photons. Increasing this energy by a factor of 2, photoelectric absorption decreases by a factor of 8. Photoelectric absorption is responsible for much of the high contrast; however, at low kV exposure techniques, the photoelectric effect occurs more that the Compton effect and therefore results in more dose to the patient.

30

2  Basic Radiation Physics for Diagnostic X-Ray Imaging

References 1. Bushong S. Radiologic science for technologists. 12th ed. Elsevier: St Louis; 2023. 2. Bushberg JT, Seibert AJ, Leidholdt EM Jr, Boone JM. The essential physics of medical imaging. Philadelphia: Wolters Kluwer/Lippincott Williams & Wilkins; 2020. 3. Seeram E.  A comprehensive guide to radiographic sciences and technology. Oxford: Wiley; 2021. 4. Huda W.  Review of radiologic physics. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2016.

Chapter 3

Computed Radiography Imaging: Physical Principles and System Components

3.1 Introduction Film-screen radiography (FSR), illustrated in Fig. 3.1, has been used to image the patient since the discovery of X-rays by WC Roentgen in 1895. FSR involves three major steps: X-ray exposure of the film creating a latent image of the patient’s anatomy, chemical processing rendering the latent image visible; and display on a light view-box for viewing and interpretation by a radiologist. As described by Seeram [1], “the film image appears with varying degrees of blackening as a result of the amount of exposure transmitted by different parts of the anatomy. While more exposure produces more blackening, less exposure produces less blackening”. This film blackening is referred to as film density, and the differences in densities in the image are referred to as the film contrast. The film optical density (OD) can be measured using a densitometer and is defined as the log of the ratio of the intensity of the view-box (original intensity) to the intensity of the transmitted light. The OD is used to describe the degree of film blackening as a result of radiation exposure. If the OD is plotted as a function of the log of the relative exposure, a characteristic curve (Fig. 3.2) is the result. This curve is important and is used in digital radiography systems (to be described later in the chapter). The overall goal of this characteristic curve is to show the film response to a range of exposures. Three regions of this curve are significant: the toe, the slope (straight line portion), and the shoulder. While the toe and shoulder regions indicate underexposure and overexposure, respectively, the slope represents the acceptable exposures used to create acceptable images. It also indicates that FSR has a narrow exposure latitude (useful range of exposures) or dynamic range. FSR is now obsolete due to several limitations [1], and has been replaced by digital radiography systems which include two X-ray imaging technologies that are now currently used in radiology. These technologies include computed radiography (CR) and flat-panel digital radiography (FPDR). © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_3

31

32

3  Computed Radiography Imaging: Physical Principles and System Components

X-ray Tube

Patient Chemical processing of Exposed film Film-screen cassette

Image Displayed on a Light View-box

Fig. 3.1  The three major steps involved in X-ray film-screen radiographic imaging. See text for further explanation 3.50

Shoulder Region (Overexposure) (Dark Image)

Film Optical Density

3.00 2.50

Straight Line Region (Slope of the Curve) (Useful Density Range) (Correct Exposure) (Acceptable Image)

2.00 1.50

1.00 Toe Region (Underexposure) (Light Image)

.50 Base + Fog Density

Log Relative Radiation Exposure Fig. 3.2  X-ray film response to radiation exposure is described by the characteristic curve. There are three regions of this curve; the toe region (low exposure); slope of the curve (acceptable exposure); and the shoulder region (high exposure). The images in the toe and shoulder regions are light and dark respectively, and are useless in diagnostic interpretation. Film screen radiography has a narrow exposure latitude. (From Seeram [1]. Reproduced by permission)

The purpose of the chapter is to identify the major system components of CR, describe the basic physics and technology of CR image formation, including image processing, and outline the nature and scope of the standardized exposure indicator, and CR image quality descriptors.

3.3  Computed Radiography (CR): System Components

33

3.2 Limitations of Film-Screen Radiography (FSR) FSR major limitation is its narrow exposure latitude for creating acceptable diagnostic images suitable for interpretation, as illustrated in Fig.  3.2. Furthermore, there are other problems with FSR that have provided the motivation to develop other digital radiography systems such as CR and FPDR systems [1]. These include limited contrast resolution, the optical range and contrast are fixed, and in terms of archiving, film is usually stored in envelopes and housed in a large room, hence requiring manual handling by a person for archiving and retrieval [1]. Digital radiography imaging systems overcome these limitations.

3.3 Computed Radiography (CR): System Components Figure 3.3 illustrates the four steps in producing a CR image: image acquisition, image plate scanning and erasure, image processing, and image display. The major system components facilitating these steps include the CR imaging plate (IP), the CR IP reader or processor, and the image display monitor.

3.3.1 The CR Imaging Plate The CR IP consists of a photostimulable storage phosphor (PSP). When exposed to X-rays, a latent image is formed on the IP. This latent image is rendered visible by scanning the IP in the CR image reader. The reader uses a laser to scan the IP in a

Fig. 3.3  The process of producing a CR image consists of four separate steps namely, image acquisition, image plate scanning and erasure, image processing, and image display

34

3  Computed Radiography Imaging: Physical Principles and System Components

systematic manner. Scanning produces a luminescence (light) that is proportional to the stored latent image. This luminescence is referred to as photostimulable luminescence (PSL). After laser scanning using a helium–neon (He–Ne) laser, the IP is erased, by exposing it to a high-intensity light beam, to get rid of any residual latent image, hence allowing the IP to be used again and again for several examinations. The PSPs used in CR imaging typically consist of barium fluorohalide, specifically europium doped (BaFX: Eu2+). The halide (X) can be chlorine (Cl), bromine (Br), or iodine (I), or a mixture of them [2, 3]. Eu2+ acts as an activator, enhancing the efficiency of PSL. Another phosphor variant used in CR is BaFBr/I: Eu2+, and more recently, cesium bromide (CsBr: Eu2+) is employed as a PSP for CR imaging [2, 3]. The basic components of a typical IP are shown in Fig. 3.4. There are two protective layers (top and bottom), a CR phosphor layer, an electroconductive layer, and a light-shielding layer. Furthermore, the purpose of the electroconductive layer is to reduce any problems during transport of the IP to the CR reader, as well any static electricity problems that may degrade image quality. CR vendors typically offer two types of IPs: a standard-resolution IP (thicker phosphor layer and more radiation absorption) and a high-resolution IP (thinner phosphor layer). The latter aims to produce sharper images (compared to the

Fig. 3.4  The basic components of a typical IP. There are two protective layers (top and bottom), a CR phosphor layer, an electroconductive layer, and a light-shielding layer. See text for further explanation

3.3  Computed Radiography (CR): System Components

35

former) by minimizing the lateral spread of laser light. IP sizes vary, with common dimensions including 17″ × 17″ (43 × 43 cm), 17″ × 14″ (43 × 35 cm), 14″ × 17″ (35 × 43 cm), and 14″ × 14″ (35 × 35 cm), while smaller sizes are also available [1].

3.3.2 Imaging Plate Exposure: Physics at a Glance The physics of CR image formation and PSL are complex and beyond the scope of this book. The interested reader, however, may refer to a comprehensive review article by Rowlands [4] for a detailed description of the physics of CR. The basic elements of the physical principles of how PSPs work after X-ray exposure are illustrated in Fig. 3.5. Seeram [1] summarizes the physics as follows: 1. X-ray exposure of the IP first causes fluorescence (light emission after X-ray exposure) for a very brief duration. Second, X-ray exposure causes ionization of the europium atoms of the PSP. 2. The electrons move from the lower energy valence band (ground state) to the higher energy conduction band.

Fig. 3.5  The basic physical principles of how PSPs work after X-ray exposure. See text for further explanation. (From Seeram [1]. Reproduced by permission)

36

3  Computed Radiography Imaging: Physical Principles and System Components

3. Electrons in the conduction band are free to travel to a so-called F-center (from the German word “Farbe” (meaning color)) where they are trapped. 4. The number of trapped electrons is proportional to the absorbed radiation. It is at the point that a latent image is formed. 5. To render the latent image visible, the IP is taken to the CR reader/processor to be scanned by a laser light. A characteristic feature of this light is that it must be capable of being absorbed by the “F-centers.” 6. “This absorption causes the trapped electrons to move up to the conduction band, where they are free to return to the valence band, thus causing the Eu3+to return to the Eu2+ state. 7. This transition of the electrons from a higher energy state to a lower energy state results in an emission of bluish-purple light (~415  nm wavelength). This is referred to as photostimulable luminescence (PSL) in the PSPIP.  This PSL is very different from the fluorescence described earlier. 8. The lasers used today for PSL in CR units are semiconductor lasers that produce light with a 680  nm wavelength compared to He-Ne lasers that produce light with a 633 nm wavelength used in earlier CR units. 9. The PSL from the IP is collected by a special light collection device and sent to a photomultiplier tube that produces an electrical signal. This signal is subsequently digitized and sent to a digital computer for processing and CR image creation” [1].

3.3.3 Imaging Plate Reader and Processor A typical PSP IP reader, also referred to as the CR scanner which is responsible for digitizing X-ray images, comprises several electronic components (Fig.  3.6). Notably, it includes a stimulating laser source, a beam splitter, oscillating beam deflector, lens, cylindrical reflecting mirror, light collection guide, photomultiplier tube (PMT), and light erasure components. As the IP is scanned, light emitted from the process of PSL is then sent to the photodetector (a photomultiplier tube) or charge-coupled device (CCD) which converts the PSL into an electrical signal (analog signal) that is first amplified and subsequently digitized by an analog-to-digital converter (ADC). As noted by Seeram [1], digitization involves both sampling the analog signal and quantization. Depending on the amplification, the ADC will produce 8–16 bits of quantization per pixel, providing discrete gray levels ranging from 28 to 216. Furthermore, dual-sided scanning of the IP is also possible using two sets of photodetectors (dual light collection system) are used to capture PSL from the front and back side of the IP. The purpose of dual-sided scanning is to improve the signal-to-noise ratio and hence, improving image quality. CR image acquisition can be accomplished using a cassette-based system (the IP is housed in a cassette similar to FSR) or a cassetteless system (automated system) using a fixed stationary single IP that is “encased in a special housing that forms a part of the unit. There is also no contact with the IP in the unit when it is read. The

3.4  Image Processing

37 Reference detector Laser Source

f-  lens

Cylindrical mirror Light channeling guide Output Signal PMT ADC

Mirror CR cassette

x= 1279 y= 1333 z= 500 To image processor

Plate translation: Sub-scan direction

Light Erasure

Laser beam: Scan direction Erasure Stage Transport

Fig. 3.6  The major components of a typical PSP IP reader also referred to as the CR scanner which is responsible for digitizing X-ray images. (From AAPM Report Acceptance Testing and Quality Control of Photostimulable Storage Phosphor Imaging Systems Report of AAPM Task Group 10 October 2006. Reproduced by permission)

single fixed IP can accommodate various exposure sizes ranging from 17″ × 17″ and 14″ × 14″ to 10″ × 12″ and 8″ × 10″ or 10″ × 8″. These varying sizes will have varying matrix sizes as well. Once the fixed stationary IP is exposed, the patent image is acquired using a scanning technology appropriate to the system” [1].

3.4 Image Processing 3.4.1 Pre-processing Figure 3.3 shows two types of digital image processing operations in CR imaging: pre-processing and post-processing. Pre-processing, also known as acquisition processing, aims to identify, correct, and scale the raw image data acquired when the IP is scanned in the CR reader, and before post-processing. While specific pre-­ processing algorithms are proprietary and vendor specific, they will not be described further in this book; however, technologists should be familiar with the following fundamental elements of pre-processing [1]:

38

3  Computed Radiography Imaging: Physical Principles and System Components

Fig. 3.7  Exposure field recognition, also referred to as exposure data recognition, includes three basic steps exposure field recognition, histogram creation, and grayscale rendition. See text for further explanation

1. Pre-processing corrects imperfections in the raw data acquired from the IP and the CR reader, such as scratches on the IP, dirt on the light channeling guide, all of which will lead to image artifacts. 2. Pre-processing ensures that the raw data is scaled so that relevant anatomical signals are used in the digitization process to improve image quality. 3. Exposure field recognition, also referred to as exposure data recognition is an important consideration in CR image processing. Exposure field recognition includes three basic steps as illustrated in Fig. 3.7. These include, exposure field recognition, histogram creation, and grayscale rendition. • Exposure field recognition identifies collimation edges and relevant anatomy to be included in the image. Additionally, in this step, an exposure indicator (EI) (sometimes referred to as an exposure index) is computed and appears on the image as a numerical value, to provide a visual cue to the technologist as to whether the correct or incorrect exposure technique factors were used for the examination. The EI will be described later in this chapter. • Histogram creation where the pixel intensity is plotted as a function of number of pixels to produce what is referred to as the scanned or measured histogram to distinguish it from stored histograms, where the identical copy of the images of the anatomy under study stored previously in the CR unit by the vendor. Using these data and special algorithms (anatomy-specific template matching algorithms), CR imaging compensates for underexposure or overexposure by matching the scanned histogram with the stored histogram appropriate known histogram using anatomy-specific template matching algorithm. This means that underexposed and overexposed anatomy will appear with the same image density and contrast, when displayed on the viewing monitor. • Grayscale rendering is a procedure that “maps the raw image values for the least penetrated anatomic region to the largest presentation value for display at maximum luminance. The most penetrated anatomic region of interest is mapped to the smallest presentation value for display at minimum luminance. The intermediate raw values are then mapped to presentation values in a monotonically decreasing fashion. This produces a presentation with a black background similar to that of conventional radiographs” [5].

3.4  Image Processing

39

3.4.2 Postprocessing Postprocessing in CR uses special algorithms to allow the observer to sharpen, smooth, reduce image noise, and enhance the image contrast. Typical algorithms include contrast enhancement, spatial frequency or edge enhancement, multi-scale and multifrequency enhancement, and dual-energy and disease-specific processing. CR vendors offer proprietary image processing software, and therefore the different types available will not be described in this book. The reader is encouraged to contact the vendors for information on their specific software. For example, FujiFilm [6] offers a number of postprocessing software, such as dynamic range control, to improve the visualization of areas with different densities in the same image, energy subtraction, multi-objective frequency processing (frequency enhancement which adjusts both large and small structures independently within the same image simultaneously), flexible noise control (separates noise and image signals, enabling selective suppression of noise levels to enhance diagnostic interpretation), grid pattern removal (removes grid patterns from the image to suppress moiré patterns within an image), and image composition [separate images are stitched (joined together) to form a single image]. Two commonly used postprocessing software in clinical practice are contrast enhancement and edge enhancement (spatial frequency processing). While the former optimizes image contrast and density using a look-up table (pixel values are normalized and rescaled) as shown in Fig. 3.8, the latter adjusts or control the sharpness of detail of an image by adjusting the frequency components of that image. An example of frequency processing to change a blurred image into a sharper image is illustrated in Fig.  3.9. An image in the spatial domain can be transformed into a spatial frequency image using the Fourier Transform (FT). This image contains both high spatial frequencies (sharpness information) at the periphery of the image and low spatial frequencies (contrast information) in the center of the image. Using a high pass digital filter to suppress the low frequencies followed by the inverse FT will result in a much sharper image.

Fig. 3.8  Contrast enhancement optimizes image contrast and density using a look-up table (pixel values are normalized and rescaled)

40

3  Computed Radiography Imaging: Physical Principles and System Components

Fig. 3.9  An example of frequency processing to change a blurred image into a sharper image

3.5 The CR Workstation and Image Display The processed CR image is generally displayed on a liquid crystal display (LCD) monitor, an integral component of the CR workstation. Various monitors are available in different sizes, and several choices of specifications are available from different vendors. An example of typical specifications of a CR monitor is the Barco® Coronis MDCG-3221 3MP 21″ Grayscale LED General Radiology Diagnostic Display [7]. A few specifications include grayscale screen type with a resolution of 3MP (2048 × 1536), a pixel pitch of 0.2115 × 0.2115 mm, and a contrast ratio of 1400:1. Additionally, other specifications include aspect ratio Hv of 4:3, maximum luminance of 1700 cd/m2, auto built-in calibration with back light, front and ambient light sensors. The backlight is LED and the screen technology is ultra-advanced-­ super fine technology. While the active screen size (diagonal) is 541 mm (21.3″), the active screen size is 325 × 433 mm (12.8″ × 17.1″). This monitor is recommended for all digital images, except digital mammography [7]. The CR workstation allows the technologists to interact with the entire CR imaging process through several functions ranging from the input of patient identification or selection of patient data and target exposure and image preview, image processing, quality assurance procedures, image printing, and sending images to the PACS through the Digital Imaging and Communication in Medicine (DICOM) standard. The major components of a typical CR workstation include the image processing computer and image display monitor, keyboard, and mouse. In addition, some workstations offer a barcode reader and a magnetic card reader (not shown). CR software is generally intuitive to provide ease of use of the system.

3.6 Response of the CR IP to Radiation Exposure In FSR, the film exhibits a narrow exposure latitude (Fig.  3.2) and requires that operators must always use the correct exposure for acceptable images of the anatomy under examination. Furthermore, while underexposure of the film results in

3.6  Response of the CR IP to Radiation Exposure

41

Fig. 3.10  The exposure latitudes of film screen detector compared to that of a digital detector. See text for further explanation

noisy images, overexposure will result in high doses to the patient. However, CR imaging has overcome these limitations with a wider exposure latitude compared to film, as illustrated in Fig. 3.10. Seeram [1] points out that “If the exposure is too low or too high, the image quality is still acceptable due to the ability of the CR system to perform digital image processing to adjust the image quality to match the image quality that would be produced by the optimum exposure. A low exposure (underexposure) will produce high noise (that can be detected by the radiologist), while a high exposure (overexposure) will produce very good images (low noise) compared to the optimum image produced by the optimum exposure (appropriate exposure). The problem with high exposures however is related to increased dose to the patient.” An example of these three image display densities for a digital detector is shown in Fig. 3.11. The display densities, which are acceptable to the viewer, appear the same for all three exposures (low, optimum, and high), compared with the narrow exposure latitude of FSR (Fig. 3.2). Furthermore, Seibert [8] notes that “because of the negative feedback due to underexposures, a predictable and unfortunate use of higher exposures, ‘dose creep’ sometimes referred to as ‘exposure creep’ is a typical occurrence. To identify an estimate of the exposure used for a given image, CR manufacturers have devised methods to analyze the digital numbers in the image based upon the calibrated response to known incident exposure.” This is an image preprocessing operation referred to as exposure field recognition (described earlier under Sect. 3.4.1) leading to the computation of the exposure indicator (EI), the amount of radiation falling upon the CR IP, after it leaves the patient. In the past, digital radiography manufacturers used proprietary procedures to compute the EI, leading to different names for EIs, such as “sensitivity” (S) number

42

3  Computed Radiography Imaging: Physical Principles and System Components

Low Noise

Image Quality

Optimum

High Exposure

Optimum Exposure

High Noise

Low Exposure

LOW

Radiation Exposure to the Detector

HIGH

Wide Exposure Latitude

Fig. 3.11  The consequence of a wide exposure latitude of a digital detector is that the display densities which are acceptable to the viewer, appear the same for all three exposures (low, optimum, and high), compared with the narrow exposure latitude of FSR (Fig. 3.2)

(Fuji), “exposure index” (Carestream), and “log of the median of the histogram (lgM).” Additionally, some vendors used a linear proportional scale and an inverse proportional scale in relating EI to the amount of radiation used for the anatomy under study. For example, while Fuji used an inverse proportional scale, Carestream used a linear proportional scale. This created not only confusion but also frustration for operators using digital radiography systems [9]. A solution to this problem would be a standardized EI. Such standardization was championed by the International Electrotechnical Commission (IEC) [10] and the American Association of Physicists in Medicine (AAPM) [11].

3.7 The Standardized Exposure Indicator (EI) Currently, all digital radiography imaging systems use the standardized EI. In this regard, therefore, it is important to understand a number of significant elements including the IEC conditions for the standardized EI, definitions, and the use of the

3.7  The Standardized Exposure Indicator (EI)

43

deviation index (DI). It is not within the scope of this book to describe these elements in detail; however, the following ones are noteworthy: 1. For the standardized EI, IEC requires that the standardized EI is related to the detector exposure, and obtained from the pixel values in the region of interest. Furthermore, the standardized EI must be based on a linear proportional scale related to the detector exposure/signal, that is, if the dose is doubled, the standardized EI value is doubled. 2. The IEC identifies four parameters, EI, target EI (EIT), DI, and the volume of interest (VOI) defined as follows: • EI is a “measure of the detector response to radiation in the relevant image region of an image acquired with a digital x-ray imaging system.” • The target EI (EIT) is the “expected value of the exposure index when exposing the X-ray image receptor properly.” • The deviation index (DI) is a “number quantifying the deviation of the actual exposure index from a target exposure index.” • The VOI is the “central tendency of the original data in the relevant image region. The central tendency is a statistical term depicting generally the center of a distribution. It may refer to a variety of measures such as the mean, median, or the mode.” 3. The IEC [10] and the AAPM [11] suggest that the user (imaging department) establishes an EIT for each body part, view, and procedure for the detector being used. 4. When the digital image is obtained, the DI is calculated as follows: DI = 10log10 (EI/EIT).

3.7.1 Use of the Deviation Index in Practice The outcome of this calculation gives different values. Only one of them is shown in the image. If DI = 0, it means that the correct exposure has been used for the body part and view, that is the EI = EIT, positive and negative DI numbers indicate overexposure, and underexposure respectively. For example. • • • •

A DI number of +1 = an overexposure of 26% more than the desired exposure. A DI number of −1 = an underexposure of 20% less than the desired exposure. A DI number of +3 = 100% more than the desired exposure. A DI number of −3 = 50% less than the desired exposure.

The acceptable range of DI numbers is approximately +1 to −1, and the DR system is able to deliver the EIT established by the department. Furthermore, numbers greater than +1 and less than −1 would indicate gross overexposure and underexposure, respectively [11].

44

3  Computed Radiography Imaging: Physical Principles and System Components

3.8 The Shortcomings of CR There is no doubt that CR imaging technology will be superseded by flat-panel digital radiography. This is the nature of evolution. FSR has been replaced by digital radiography because of the limitations imposed by FSR. CR is not free of shortcomings [1–3, 12]. For example, the detection efficiency of CR is considered poor compared to film screen detectors (intensifying screens), the spatial resolution of CR is about 6 line-pairs/mm (lp/mm) compared to FSR which is about 8 lp/mm3. Other limitations include easily damaged IPs during transportation to the CR image reader and mobile imaging. IPs are also subject to scratches and cracking. Furthermore, CR imaging can deliver high radiation doses to the patient, and nondynamic imaging [12], and can produce a wide range of artifacts, arising from the image acquisition system (including operator errors) and image processing systems [3, 13–16]. Artifacts will not be described in this book; however, the interested reader may refer to references [13–16] for a comprehensive coverage of these artifacts. It is noteworthy to mention that an artifact is “any false visual feature on a medical image that simulates tissue or obscures tissue” [3]. Another definition of an artifact is provided by Willis et al. [13]. who state that “an artifact is a feature in an image that masks or mimics a clinical feature”. Artifacts may pose interpretation problems for radiologists, and others interpreting CR images. The above problems have been solved by digital flat panel radiography systems, which will be described in Chap. 4.

References 1. Seeram E.  Digital radiography: physical principles and quality control. 2nd ed. Singapore: Springer Nature; 2019. 2. Lanca L, Silva A. Digital imaging systems for plain radiography. New York: Springer; 2013. 3. Bushong S. Radiologic science for technologists: physics, biology, and protection. 12th ed. St Louis: Elsevier; 2021. 4. Rowlands JA. The physics of computed radiography. Phys Med Biol. 2002;47:R123–66. 5. Flynn MJ. Processing digital radiographs of specific body parts. In: Samei E, editor. Advances in digital radiology, categorical course in diagnostic radiology physics. Oak Brook: RSNA; 2003. 6. Fuji Film CR Console. http://medi-­diagnostic.com/wp-­content/uploads/2017/06/CR-­Console-­ Catalogue.pdf. Accessed 20 Jan 2023. 7. Barco® Coronis MDCG-3221 3MP 21″ grayscale LED general radiology diagnostic display. https://monitors.com/collections/general-­radiology-­pacs-­displays/products/barco-­coronis-­ mdcg-­3221-­k9301366a. Accessed 30 Jan 2023. 8. Seibert JA.  Computed radiography technology 2004. In: Goldman LW, Yester MV, editors. Specifications, performance, and quality assurance of radiographic and fluoroscopic systems in the digital era. College Park: American Association of Physicists in Medicine (AAPM); 2004. 9. Seibert JA, Morin RL. The standardized exposure index for digital radiography: an opportunity for optimization of radiation dose to the pediatric population. Pediatr Radiol. 2011;41:573–81. 10. International Electrotechnical Commission. IEC 62494-1 Ed. 1 medical electrical equipment exposure index of digital X-ray imaging systems part 1: definitions and requirements for general radiography. Geneva: IEC; 2008.

References

45

11. American Association of Physicists in Medicine (AAPM). An exposure indicator for digital radiography, Report No. 116. College Park: AAPM; 2009. 12. Ou X, Chen X, Xu X, Xie L, Yang H. Recent development in X-ray imaging technology: future and challenges. Research. 2021;2021:98921. https://doi.org/10.34133/2021/9892152. 13. Willis CE, Thompson SK, Shepard SJ.  Artifacts and misadventures in digital radiography. Appl Radiol. 2004;33(1):11–20. 14. Cesar LJ, Schuelar BA, Zink FE, Daly TR, Taubel JP, Jorgenson LL. Artefacts in computed radiography. Brit J Radiol. 2001;4:195–20. 15. Honey ID, Mackenzie A. Artifacts found during quality assurance testing of computed radiography and digital radiography detectors. J Digit Imaging. 2009;22(4):383–92. 16. Shetty CM, Barthur A, Kambadakone A, Narayanan N, Kv R. Computed radiography image artifacts revisited. Am J Roentgenol. 2011;196:W37–47.

Chapter 4

Flat-Panel Digital Radiography: Principles and System Components

4.1 Introduction Chapter 3 highlighted several limitations of computed radiography (CR), including poor detection efficiency, lower spatial resolution, easily damaged imaging plates (IPs) when dropped accidently, cracking and scratches of IPs, and so on. These limitations have provided the motivation to develop flat-panel digital detectors for digital radiography. The aim of this chapter is to explain the overall system components of flat-panel digital detectors, including the fundamental physical principles and technological considerations of these detectors. It is not within the scope of this book to outline and explain the physics of scintillation phosphors, photoconductors, semiconductors, and the electronics of these detectors. The interested readers may refer to Bushberg et al. [1] for a detailed outline of the physics of these materials.

4.2 Flat-Panel Digital Radiography: System Components Overview The major components of a flat-panel digital radiography (FPDR) imaging system are shown in Fig. 4.1. Major components include the X-ray tube and generator, the flat-panel detector (FPD), the host computer, and the image display monitor (part of the FPDR workstation). It is not within the scope of this book to describe the details of all the components in Fig. 4.1; however, the FPD will be described in detail in Sect. 4.2.1. It is important to note the role of the host computer in the system. The host computer facilitates the operator to communicate not only with the X-ray tube and generator, and other system components, such as network communication and image storage (not shown in Fig.  4.1) but also interacts with the © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_4

47

48

4  Flat-Panel Digital Radiography: Principles and System Components

Fig. 4.1  The major components of a Flat-Panel Digital Radiography (FPDR) imaging system. See text for further explanation

FPD.  For instance, the X-ray production is controlled by the host computer. Additionally, the signal readout is also controlled by the host computer. Furthermore, the host computer plays a major role in image processing, such as image correction and image display optimization [2]. Image processing will be described later in this chapter.

4.2.1 The Flat-Panel Detector The FPD is shown in Fig. 4.1 and represents single unit (a thin flat-panel device) that contains what is referred to as an active-matrix detector, and associated electronics such as preamplifiers, switching control, central logic circuits, the ADCs, and internal memory. The active-matrix detector detects the radiation coming the patient during imaging. It is also responsible for digitization of the electronic signal output coming from the flat-panel detector. The FPD is available in different sizes to meet the needs of the varying body sizes being examined. Common dimensions include 43 × 43 cm, 30 × 40 cm, and 18 × 18 cm. Furthermore, the active-matrix size also varies depending on the size of the detector. Common active-matrix sizes include 1920  ×  1536, 2000  ×  2500, 2736 × 2736, 2560 × 3072, 2688 × 2688, 2836 × 2336, and 3121 × 3121 [3]. An important characteristic of the matrix is the pixel size and the pixel pitch (distance from midpoint of one pixel to the midpoint of the adjacent pixel) as they affect the spatial resolution of the image. The number of pixels that make up the active-matrix detector can be computed by the dimensions of the matrix. For example, the number of pixels in a 3000 × 3000 matrix is 9,000,000 and pixel sizes varies from vendor to vendor; however, common sizes include 200 μm, 167 μm, 160 μm, 143 μm, and 139 μm [3].

4.3  Types of Flat-Panel Detectors

49

Fig. 4.2  The active-matrix detector (b) consists of pixels and a 2 × 2 matrix (a) extracted from the active-matrix detector

4.2.2 The Fill Factor Figure 4.2b shows the active-matrix detector which consists of pixels and a 2 × 2 matrix (Fig. 4.2a) extracted from the active-matrix detector. Each pixel of the 2 × 2 matrix consists of three electronic components: a thin film transistor (TFT), a capacitor, and the radiation-sensing area, which is used to define what is referred to as the fill factor. The fill factor is the ratio of sensing area of the pixel to the area of the pixel itself. Furthermore, the fill factor can be expressed as a percentage, where a fill factor of 80% means that 20% of the pixel area is occupied by the detector electronics with 80% representing the sensing area. An important point to note is that image quality afforded by the detector [spatial resolution and contrast resolution (signal-to-noise ratio)] is influenced by the fill factor. High fill factors will provide better image quality compared with low fill factors.

4.3 Types of Flat-Panel Detectors Korner et al. [4] coined the term direct radiography (DR) to refer to the utilization of flat-panel digital detectors in X-ray imaging. There are two types of DR detectors: indirect conversion detectors and direct conversion detectors. The former captures X-ray photons and converts them into electrical charges, while the latter directly converts X-ray photons into electrical charges. The fundamental technical differences between these two are illustrated in Fig. 4.3.

50

4  Flat-Panel Digital Radiography: Principles and System Components

Fig. 4.3  The fundamental technical differences between two types of DR detectors; indirect conversion detectors (a) and direct conversion detectors (b). See text for further explanation

4.3.1 Indirect Conversion Digital Detectors: Major Components The major technical components of an indirect digital detector are shown in Fig. 4.3a and include an X-ray scintillator coupled to a photodiode with a thin film transistor (TFT) array, and a storage capacitor. The x-ray scintillator is typically made of phosphor that converts x-ray photons into light, which is usually cesium iodide (CsI) or gadolinium oxysulfide (Gd2O2S). The CsI phosphor is deposited in a needle-like fashion and is referred to as a structured phosphor (Fig. 4.3a). This design is intended educe the lateral dispersion of light characteristic of the unstructured scintillator consisting of powdered phosphor (turbid phosphor) design. The purpose of the structured phosphor design is to improve the spatial resolution of the image by reducing the lateral spread of light. The next system component is an a-Si photodiode flat-panel array, and the purpose of which is to convert the light from the scintillator into electrical charges. Essential electrical components in the design of the flat-panel are thin-film transistor (TFT) array, storage capacitors, and associated electronics (not shown in Fig. 4.3) The capacitors collect and store the electrical charge produced in the a-Si photodiode array. The process of conversion of X-ray photons to light photons to electrical charges in this type of digital detector has been referred to as an indirect conversion process.

4.3.2 Direct Conversion Digital Detectors: Major Components Figure 4.3b illustrates the major electrical components of a direct conversion digital detector, which include source of high voltage, top electrode, dielectric layer, photoconductor, collection electrode, TFT, storage capacitor, and the glass substrate.

4.4  Image Processing Considerations

51

It is not within the scope of this book to describe the details of each of these ­components; however, the basics of the photoconductor will be highlighted. As mentioned earlier, this detector directly converts X-ray photons into electrical signals. This is the function of the photoconductor, amorphous selenium (a-Se) although other photoconductors such as lead oxide, lead iodide, thallium bromide, and gadolinium compounds can be used [4]. The use of the photoconductor a-Se has become popular because of its excellent X-ray photon detection properties and it provides images with very high spatial resolution. In a direct conversion digital detector, an electric field is created between the top electrode and the TFT elements. The a-Se layer converts X-ray photons into electrical charges are produced and move toward the TFT elements where they are collected and stored. Subsequently, these charges must now be read out. This task is performed by what is referred to as the readout electronics, consisting of switch lines, data lines, control voltage, and external electronics. As described briefly by Yorkson [5], “all the control contacts of the pixel switching elements along a horizontal line are connected to the same horizontal readout control line, and all of the signal readout connections of the pixels along a vertical column are connected to the same vertical data output line.”

4.4 Image Processing Considerations Digital X-ray imaging systems utilize image processing operations not only to correct raw data from the detector but also to optimize the display of the image that is presented to the observer for interpretation. While the former is intended to reduce artifacts in post processing, the purpose of the latter is to enhance contrast and sharpness of the image while simultaneously reducing image noise [1, 3], using pre-­ processing and post processing operations, respectively. Following this, the raw digital data are pre-processed and presented as a DICOM “for processing” image. This image is then post-processed and is subsequently referred to as a “for presentation” image [6] as illustrated in Fig. 4.4.

4.4.1 Image Pre-processing There is a possibility that some detector elements may not be active or functioning as intended, leading to the creation of artifacts such as dead or bad pixels, or even an entire bad column of pixels. In such cases, the obtained image is termed a flat-­ field image. Pre-processing addresses this issue using a specialized software tool called flat-fielding [6] (Fig. 4.5) and is a mandatory part of system calibration, to ensure detector performance integrity.

52

4  Flat-Panel Digital Radiography: Principles and System Components

Fig. 4.4  Pre-processing and post processing operations. First, the raw digital data is pre-processed and presented as a DICOM “for processing” image. This image is then post-processed, and is subsequently referred to as a “for presentation” image

Fig. 4.5  Flat-fielding is pre-processing software used to correct bad detector elements would lead to the creation of artifacts due to dead or bad pixels, or a bad column of pixels and so on. In this case, the image obtained is called a flat-field image

4.4.2 Image Postprocessing The wide exposure latitude of the digital detector produces an initial image that has poor image contrast. This image is what is referred to as a “for processing” image, and is not presented for viewing and interpretation, but rather is subject to post processing, and is hence referred to as a “for presentation” image, in order to be presented for viewing and interpretation by a radiologist. The steps for getting to the “for presentation” image from the “for processing” image have been described in Chaps. 2 and 3. In summary, these steps include exposure recognition and segmentation. Histogram analysis; histogram scaling and

4.5  Image Quality Considerations

53

finally image processing for contrast enhancement. The displayed image can be post-processed further to enhance image detail or sharpness and to reduce image noise, using the spatial frequencies (high and low frequencies contained in the image). While the high frequencies contain fine details of the image, low frequencies contain the contrast information in the image. While a high-pass filter suppresses low frequencies to sharpen the image, a low-pass filter can be used to suppress the high frequencies to smooth (blur) the image. Additionally, image display can be optimized to enhance diagnostic interpretation [7].

4.5 Image Quality Considerations Image quality descriptors were outlined in Chap. 3 and include spatial resolution, density resolution, noise, detective quantum efficiency (DQE), and image artifacts. The “ability of the detector to produce a high-quality X-ray image” [5] depends on several image quality factors and includes spatial resolution (sharpness), modulation transfer function (MTF), dynamic range, detective quantum efficiency (DQE), image lag, ghosting, and artifacts. It is not within the scope of this book to describe the full details of these factors; however, the following is noteworthy as outlined by Seeram [3]: 1. “Spatial resolution is the ability of the imaging system to resolve fine details present in an object. It also refers to the sharpness of the image. For digital imaging systems, the spatial resolution depends on the size of the pixels in the matrix. Smaller pixels will produce images with better spatial resolution compared with larger pixels.” Spatial resolution is best with the a-Se detector and that structured CsI a-Si TFT detector produces better spatial resolution than the turbid Gd2O2S or CsI digital detector [1, 7]. 2. The detective quantum efficiency (DQE) is a physical concept that describes how a detector converts the radiation falling upon it into a useful image signal, and hence, provides information about the signal-to-noise ratio (SNR). In diagnostic radiography, a high SNR (high signal, low noise) is needed as contrast resolution (the ability to resolve small differences in tissue contrast) is required to enable diagnostic interpretation. In review, the DQE is expressed mathematically as follows: DQE =

SNR 2out SNR 2in

A DQE of 1 indicates a perfect digital detector. In practice, the indirect conversion CsI a-Si TFT flat-panel detectors provides the highest DQE at low frequencies and decreases rapidly as the spatial frequencies increase [6]. Interested readers may check imaging vendors for values of the DQE for their digital detectors.

54

4  Flat-Panel Digital Radiography: Principles and System Components

4.6 Wireless FPDR Detectors Wireless flat-panel digital radiography (FPDR) was introduced commercially as of 2009 [8]. Wireless Fidelity (WiFi), now ubiquitous in various industries, including telecommunications, has found application in digital radiography. In this context, the flat-panel detector is wirelessly connected to radiography unit console via a wireless local area network (LAN). This wireless connection facilitates real-time transmission of images and text between the exposed detector and the digital radiography workstation. Moreover, the detector is equipped with a lithium-ion capacitor (battery power supply) to ensure patient safety by preventing any potential injury [9]. Additional details such as spatial resolution, sensitivity, wireless range, data transfer speed, charging time and charge life of the battery, details of battery use, and software features are available from various vendors. Readers interested in the physical characteristics of these detectors such as the DQE, for example, should refer to an article by Samei et al. [10]. Another study explored the usefulness of wireless digital detectors in an imaging environment, revealing not only improved image quality but also a reduction in patient radiation doses [11].

4.7 Mobile Digital Radiography Systems Mobile digital radiography refers to using a “portable” digital radiographic imaging system that can be transported to the bedside of patients who are too critically ill to travel to the main radiology department for their X-ray examination. These systems have evolved from computed radiography (CR) mobile systems to flat-paned digital radiography (DR) systems [12]. The transitioning from CR to DR systems has been evaluated by Gali et  al. [12] Their results showed that “overall, adoption of DR technology resulted in a drop in time by 27% relative to the use of CR technology.” Transitioning from CR to DR was associated with improved workflow efficiency for radiographic imaging with portable X-ray units. DR mobile imaging systems are now commonplace and are widely used in radiology departments. It is not within the scope of this book to describe the details of mobile digital imaging systems however the interested reader may refer to a publication called “Portable digital radiography systems technical specifications” provided by the World Health Organization (WHO) and the International Atomic Energy Agency (IAEA) [13]. The “core system composition requirement” for example include X-ray generator, X-ray generator stand/frame; X-ray detector; X-ray detector stand/ frame; portable workstation/PC-console and/or portable remote control-station, software/hardware for data management and communication, case/bag for packing and transportation, and accessories, including radiation protection devices [13]. Of these components, it is noteworthy to highlight major features of the X-ray generator, X-ray tube, and the digital detector.

References

55

• The X-ray generator is a high frequency generator such that the “voltage range must include the range from 50 kV to not less than 90 kV (better with minimum range at least: from 50  kV up to 110  kV), preferably digitally displayed … current-­time range must include the range from 0.5 to 2.5 mAs (better with minimum range at least: from 0.3 up to 100 mAs) and preferably digitally displayed. Maximum current: at least 5 mA @ 90 kV (or calculated on the product voltage available) (better: up to at least 20 mA)” [13]. • The X-ray tube can be either a “stationary or rotating (better) anode with focal spot size less than 1.3 mm; Heat storage capacity of the anode at least 10,000 HU (preferably higher)” [13]. • The digital detector “active detector area not less that 35 × 43 cm” and the main features as noted in by WHO [13]: –– –– –– ––

“Time to display image after exposure no longer than 10 s. Preferably, pixel pitch not greater than approximately 150 μm. Spatial resolution not less than 3 lp/mm (better: at least 3.5 lp/mm). Preferably, DQE, detective quantum efficiency @ RQA5 at least 70% (better: at least 80%). –– Dynamic range of A/D converter at least 14 bit (preferably 16 bit) or at least 10 pixels resolution. –– Exposure capacity when fully charged (battery autonomy time) at least 100 chest X-ray @ 90 KV (or calculated on the product voltage available). –– Detector connectivity to workstation capabilities (wireless feature/option preferably included)” [13]. Commercially available detectors now commonly include gadolinium (GOS) scintillator detectors for general radiographic imaging and cesium (CsI) scintillator for imaging the pediatric patient.

References 1. Bushberg JT, Seibert JA, Leidholdt EM Jr, Boone JM. The essential physics of medical imaging. Philadelphia: Wolters Kluwer; 2021. 2. Yorkston J. Digital radiographic technology. In: Samei E, Flynn M, editors. Advances in digital radiography. RSNA categorical course in diagnostic radiology physics. Oak Brook: RSNA; 2003. p. 23–36. 3. Seeram E.  Digital radiography: physical principles and quality control. 2nd ed. Singapore: Springer Nature; 2019. 4. Korner M, et al. Advances in digital radiography: physical principles and system overview. Radiographics. 2007;27:675–86. 5. Yorkston J.  Flat-panel DR detectors for radiography and fluoroscopy. In: Goldman L, Yester MV, editors. Specifications, performance, and quality assurance of radiographic and fluoroscopic systems in the digital era, AAPM monograph 30. College Park: AAPM; 2004. p. 177–229. 6. Seibert JA. Computed radiography/digital radiography: adult. In: Frush DP, Huda W, editors. From invisible to visible-the science and practice of X-ray imaging and radiation dose optimization: RSNA categorical course in diagnostic radiology physics. Oak Brook: RSNA; 2006. p. 57–71.

56

4  Flat-Panel Digital Radiography: Principles and System Components

7. Bushong S. Radiologic science for technologists. 12th ed. St Louis: Elsevier-Mosby; 2021. 8. Lanca L, Silva A. Digital imaging systems for plain radiography. New York: Springer; 2013. 9. Tokuhiro O, Gidou T, Kashino T. Development of a wireless cassette digital radiography detector. Konica Minolta Medical & Graphic, Inc., Development Center, Development Department, Konica Minolta technology report vol 8; 2011. 10. Samei E, Murphy S, Christianson O. DQE of wireless digital detectors: comparative performance with differing filtration schemes. Med Phys. 2013;40(8):081910. 11. Garrido Blazquez M, Agulla Otero M, Rodriguez Reco FJ, Torres Cabrera R, Hernando GI.  Wireless digital radiography detectors in the emergency area: an efficacious solution. Radiologia. 2013;55(3):239–46. 12. Gali RL, Roth CG, Smith E, Dave JK. Does transitioning from computed radiography (CR) to direct radiography (DR) with portable imaging systems affect workflow efficiency? Proceeding of the SPIE 10579, medical imaging 2018: imaging informatics for healthcare, research, and applications, 105790Z; 2018 Mar 6. https://doi.org/10.1117/12.2293754. 13. Portable digital radiography system: technical specifications. Geneva: World Health Organization; 2021. (License: CC BY-NC-SA 3.0 IGO).

Chapter 5

Digital Fluoroscopy: System Components and Principles

5.1 Introduction Chapter 4 provided an in-depth exploration of the technological components of digital radiography imaging systems using flat-panel detectors. The focus of this chapter is to describe the types of digital fluoroscopy (DF) imaging systems and review the major components of each system. These systems include image intensifier-­based digital fluoroscopy, which evolved into digital fluoroscopy with flat-panel detectors and digital subtraction angiography.

5.2 Development of Fluoroscopy Systems at a Glance Fluoroscopy represents a real-time imaging modality that shows images displayed continuously while the X-ray tube is energized [1]. These images not only reveal anatomical structures but also capture the movement of contrast media within blood vessels and organs. The dynamic imaging capability provides functional information to the observer (usually a radiologist) who interprets these dynamic images. The evolution of fluoroscopy is marked by the introduction of several technical innovations intended to enhance the viewing of dynamic images for diagnostic interpretation. Fluoroscopy evolved from the direct viewing of images on a fluoroscope invented by Thomas Edison in 1896, 1 year after the discovery of X-rays by Wilhelm Conrad Roentgen in 1895. The screen of the fluoroscope used the phosphor zinc cadmium sulfide (ZnCdS), which emitted a yellow–green light when struck by X-rays. The limitations of poor image quality (detail and contrast) and the lack of brightness of the image displayed on the fluoroscope provided the motivation to develop newer systems to overcome these limitations and provide advantages, which resulted in lower dose to patients. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_5

57

58

5  Digital Fluoroscopy: System Components and Principles

Subsequent fluoroscopy machines that followed chronologically include image intensifier fluoroscopy, image intensifier-based digital fluoroscopy, and digital fluoroscopy using flat-paned digital detectors (FPDs) used today. The latter systems are considered state-of-the art fluoroscopy, and a major characteristic feature is that they are based on what is popularly referred to as dynamic FPDs. Currently, there are two types: the cesium iodide amorphous-silicon (CsI a-Si) TFT indirect digital detector and the a-selenium TFT direct digital detector. The purpose of an image intensifier is to convert X-ray photons into light photons that are captured by the video camera. The output video signal from the video camera goes to the television display monitor to create dynamic fluoroscopic images.

5.2.1 Image Intensifier-Based Digital Fluoroscopy: Essential Technical Components A typical image intensifier-based digital fluoroscopic imaging chain is shown in Fig. 5.1. The major system components include the X-ray tube and generator, a scattered radiation grid, the image intensifier tube, the optical image distributor, the analog-­ to-­digital converter (ADC), a digital computer. a digital-to-analog converter (DAC), and finally, the television monitor. The essential features of each of these components will be briefly described below. • The X-ray tube and generator provide the appropriate radiation beam required for image intensifier-based digital fluoroscopy. The X-ray generator is a high frequency generator designed to provide high milliampere (mA) required for digital fluoroscopy, and it provides the electrical energy to a high-capacity X-ray tube to produce X-rays. Bushong [1] notes the following points relating to the generation and production of X-rays in digital fluoroscopy: –– The mA is about 100 times higher than that used in earlier fluoroscopic imaging systems (3–5 mA), and for this reason, the generator operates in a pulse

Fig. 5.1  The major system components of a typical image intensifier-based digital fluoroscopic imaging chain. See text for further explanation

5.2  Development of Fluoroscopy Systems at a Glance

59

mode referred to as “pulsed-progressive fluoroscopy” rather continuously, in an effort to reduce the radiation dose to the patient. Furthermore, the X-ray tube operates in the radiographic mode. • The image intensifier (II) tube is responsible for increasing the brightness level of the fluoroscopic image, based on its design and function. The major ­components of the II tube are shown in Fig. 5.2 and include an input screen, a photocathode, an electronic lens system, and an output screen. –– The input screen converts X-ray photons to light photons and is coated with a cesium iodide (CsI) phosphor arranged in a needle-like manner (structured phosphor) to reduce the lateral spread of light (improves spatial resolution of the image). Input screen diameters can range from 13 to 30 cm, and larger diameters (36–57 cm) have become available for imaging larger anatomical regions, such as the abdomen. –– The light photons fall upon the photocathode which is made of antimony cesium (SbCs) and which emits photoelectrons. For example, one X-ray photon falling on the input screen will produce about 200 photoelectrons [2]. Multialkali photocathodes with a combination of potassium, sodium, and cesium will emit about three times more photoelectrons than SbCs photocathodes, making these image intensifiers much more efficient than single alkali photocathodes. –– The electronic lens or electron optics consists of a series of electrodes that accelerate and focus the photoelectrons from the photocathode to the output screen. This requires a voltage of about 25–30 kV applied between the photocathode and the output screen.

Fig. 5.2  The major components of the image intensifier tube include an input screen, a photocathode, an electronic lens system, and an output screen. See text for further information

60

5  Digital Fluoroscopy: System Components and Principles

–– The output screen converts photoelectrons into light photons and is made of ZnCdS phosphor. Additionally, the diameter of the output screen is much smaller that the diameter of the input screen. This minification, together with the acceleration of the photoelectrons which strike the output screen are responsible an extremely bright image, which is too small to look at directly. This increase in brightness is conveniently referred to as the brightness gain. Today the brightness gain concept has been replaced by the term conversion factor which measures the light gain at the output screen. The conversion factor is a ratio of the luminance (candela/square meter = Cd/m2) of the output screen to the exposure rate milliroentgens/second (mR/s) at the input screen and can range between 50 and 300 [1]. The higher the conversion factor means that more efficient performance compared to an image intensifier with a lower conversion factor. –– Magnification fluoroscopy. The image intensifier can be operated in what is referred to as a magnification mode, in an effort to increase provides increased spatial resolution of the displayed image. The consequence of this technique is an increase in radiation dose by about 2.2 times when the image intensifier operates in the full-field mode [1]. A detailed description of this concept can be found in Bushong [1]. • The image distributor is an optical system designed to direct the light from the output screen to the video camera. • The video camera is a charged couple device (CCD) and has replaced television camera tubes for image intensifier-based digital fluoroscopy. The CCD has extremely high sensitivity and low readout noise level. Images can be acquired at 60/s compared with 30/s for television tubes. • The analog-to-digital converter (ADC) digitizes the analog signal from the CCD for input into a digital computer. A 10-bit ADC for example will divide the signal into 1024 (210) parts. The higher the number of bits, the more accurate is the ADC. • The digital computer for digital fluoroscopy is a mid-range (previously called minicomputer) computer capable of processing the data set received from the ADC for image display on the television monitor. These images can be post-­ processed using a number of different image processing operations, such as the last image hold, grayscale image manipulation, and edge enhancement [3].

5.3 Digital Fluoroscopy with Flat-Panel Digital Detectors The image intensifier-based digital fluoroscopy (II-based DF) system has been replaced by digital fluoroscopy systems using flat-panel digital detectors (FPDs) due to the shortcomings of II-based DF systems that are related to artifacts, such as veiling glare, vignetting, image lag, and pincushion and “S” distortions. While veiling glare refers to scattering of light in the intensifier tube, vignetting refers to the

5.3  Digital Fluoroscopy with Flat-Panel Digital Detectors

61

Fig. 5.3  Two examples of in artifacts in image intensifier-based digital fluoroscopy are pincushion distortions and “S” distortions illustrated in (a). These distortions appear if there is an electromagnetic field is located close to the intensifier and can be eliminated by a flat-panel DF (FPDF) system, as illustrated in (b)

brightness loss at the periphery of the image, while the image is sharper and much brighter in the central portion of the screen. Image lag on the other hand continued emission of light from the screen when the radiation beam has been turned off. Other artifacts such as pincushion distortions and “S” distortions are illustrated in Fig. 5.3a. These distortions appear if there is an electromagnetic field located close to the intensifier. These distortions are eliminated by a flat-panel DF (FPDF) system, as illustrated in Fig. 5.3b. The fundamental characteristics of a FPDF system will be described further.

5.3.1 System Configuration The overall equipment configuration of a FPD digital fluoroscopy imaging system (real-time imaging) is shown in Fig. 5.4. The major system component in this design is the FPD, which has replaced the image intensifier in the II-based DF system. These detectors are used in radiographic imaging and produce static images, and hence, they are referred to as static FPDs. There are two major the types of static FPDs: indirect FPD and direct FPD. The major components of each are illustrated in Fig. 5.5. An important and significant technical feature of FPDs used for FPDF system is that it must be capable of producing dynamic images that can be displayed and viewed in real time. For this reason, these detectors are sometimes referred to as dynamic FPDs.

62

5  Digital Fluoroscopy: System Components and Principles

Fig. 5.4  The overall equipment configuration of a flat-panel digital fluoroscopy imaging system (real-time imaging). See text for further explanation

Fig. 5.5  The major technical components of the indirect FPD and direct FPD used in flat-panel digital fluoroscopy. See text for further explanation

5.3  Digital Fluoroscopy with Flat-Panel Digital Detectors

63

5.3.2 Types of Dynamic FPDs Digital radiography imaging manufacturers offer two types of dynamic FPDs for digital fluoroscopy, namely, the cesium iodide a-silicon thin film transistor (CsI a-Si TFT) indirect digital detector and the a-selenium (a-Se) TFT direct digital detector. Although these detectors share similarities in design with the static FPDs used for radiographic imaging, there are a few significant differences [4].

5.3.3 Technical Features of Dynamic FPDs The technical features of dynamic FPDs include the detector dimensions, matrix sizes, pixel considerations, and the zoom feature. Typical dimensions of these detectors vary; however, 31 cm × 31 cm, 35 cm × 35 cm, 30 cm × 40 cm, and 41 cm × 41 cm are commonplace. Furthermore, a 43 cm × 43 cm dynamic FPD is now available for both digital fluoroscopy and digital radiography. While the typical matrix sizes include 1024 × 1024, 2304 × 2304, and 2048 × 2048, the pixel size for digital fluoroscopy detectors is larger than the pixel size used in digital radiography detectors. In some systems, it is possible “to adjust the pixel size by binning four pixels into one larger pixel. Such dual-use systems have pixels small enough to be adequate for radiography (e.g., 100–150 μm), but the pixels can be binned to provide a detector useful for fluoroscopy (e.g., 200–300 μm)” [2]. Furthermore, dynamic FPDs offer “zoom” modes of operation. For example, the radiologist may zoom into the survey image and examine the details of smaller structures.

5.3.4 Principles of Operation A significant feature of dynamic real-time fluoroscopy using dynamic FPDs is how the data is read out. This is accomplished by the readout electronics. There are several considerations as follows: • The electronics allow for frame rates of 15–30 frames per seconds (fps) or greater are possible [5, 6]. • Readout speeds of 30–50 milliseconds (ms) [5, 6]. • Two readout modes are available, namely, frame rate mode and pulsed frame rate mode when the X-tube operates in the continuous and pulsed modes respectively. • A single image (completed in at least 33  ms) is obtained following three sequences initialization, integration, and readout [6]. “While initialization prepares the detector electronics for X-ray exposure, integration and readout are intended to collect the detector signal (analog signal) for subsequent digitization and image display” [7].

64

5  Digital Fluoroscopy: System Components and Principles

2 Pixels

4 Pixels

2 2 Binning

Binning

4 Pixels

2 Pixels

Fig. 5.6  Binning is a pixel-based operation used to magnify the image in flat-panel digital fluoroscopy. See text for further explanation

Magnification is yet another consideration in the operation of dynamic FPDs. The purpose of magnification is to improve the spatial resolution of the fluoroscopic image (described above). Dynamic FPDs feature two magnification methods, namely, electronic magnification (zoom) and binning. Electronic magnification is similar to the process used with the image intensifier (with a large field-of-view (FOV) and a smaller central FOV where “collimation is used to select only the central portion of the FPD for imaging” [3]) and there is no increase in spatial resolution. Both original and magnified images have the same signal-to-noise ratio (SNR). Binning on the other hand is illustrated in Fig. 5.6. The shortcoming of this technique is a loss in spatial resolution “because the effective area of each image pixel is four times larger, and it has the advantage of lower data rates and less image mottle than ungrouped pixels” [3]. As noted by Nickoloff [3] “When smaller FOVs are used, the data rate is lower, and binning is no longer required. Unlike image intensifier fluoroscopy systems, the spatial resolution of FPD fluoroscopy systems is the same for all FOVs—if no binning is employed. For those larger FOVs, when binning is employed, the spatial resolution dramatically decreases to 50% of the value without binning. For FPD systems, there is a dramatic, discrete step change in spatial resolution between small and large FOVs.”

5.4 Image Post Processing in Digital Fluoroscopy Image post processing is used in DF (II-based DF and FPDF) for the purpose of supporting the diagnostic interpretation of the fluoroscopic image [7, 8], for both gastrointestinal (GI) tract fluoroscopy and digital subtraction angiography (DSA); however, there are specific image post-processing operations that are applied to each of them. These image post processing algorithms include grayscale image manipulation, road mapping; temporal frame averaging; last image hold, and edge

References

65

enhancement. For example, while grayscale image manipulation is common to GI fluoroscopy, road mapping is an operation commonly used in DSA. It is not within the scope of this chapter to describe these image processing operations in any detail, as this topic is not within the scope of practice of biomedical engineering technologists. The interested reader, however, may refer to Bushberg et al [2], Seeram [7], and Seeram and Seeram [8], for a description of how these operations are used in DF.

References 1. Bushong S. Radiologic science for technologists. 11th ed. Elsevier-Mosby: St Louis; 2021. 2. Bushberg JT, Seibert JA, Leidholdt EM Jr, Boone JM. The essential physics of medical imaging. 4th ed. Philadelphia: Lippincott Williams and Wilkins; 2020. 3. Nickoloff EL. Survey of modern fluoroscopy imaging: flat-panel detectors versus image intensifiers and more. Radiographics. 2011;31:591–602. 4. Seeram E, Brennan P.  Radiation protection in diagnostic X-ray imaging. James & Bartlett Learning: Burlington; 2017. 5. Jones AK, Balter S, Rauch P, Wagner LK. Medical imaging using ionizing radiation: optimization of dose and image quality in fluoroscopy. Med Phys. 2014;41(1):014301. 6. Holmes DR, Laskey WK, et al. Flat-panel detectors in the cardiac catheterization laboratory: revolution or evolution-what are the issues? Catheter Cardiovasc Interv. 2004;63:324–30. 7. Seeram E.  Digital radiography: physical principles and quality control. 2nd ed. Singapore: Springer Nature; 2019. 8. Seeram E, Seeram D. Image postprocessing in digital radiology: a primer for technologists. J Med Imaging Radiat Sci. 2008;39(1):23–43.

Chapter 6

Digital Image Quality Descriptors and Performance Characteristics

6.1 Introduction Computed radiography (CR) and flat-panel digital radiography (FPDR) images are digital images [1]. These digital images are presented as a matrix consisting of several rows and columns [for example, a matrix of 12 × 12 define small square regions referred to as picture elements or pixels (12 × 12 matrix = 144 pixels), as illustrated in Fig. 6.1]. Each pixel will have a numerical value that is expressed as a shade of gray, and therefore, the image matrix of numerical values is converted into a gray scale image (Fig. 6.1). X-ray image quality is a complex subject, especially when using imaging physics to describe the characteristics of the image produced by a digital detector. For example, two such physical concepts include the modulation transfer function (MTF) and the detective quantum efficiency (DQE) both of which play a role in explaining the imaging performance of a detector. Such performance as defined by Yorkson [4] refers to the “ability of the detector to produce a high-quality X-ray image.”

6.2 Image Quality Characteristics The quality of digital images is characterized by several technical parameters describing the sharpness or detail of the image (spatial resolution), the range of gray levels per pixel (bit depth) that define the brightness of the image (density resolution), and the number of photons used to produce a useful image (noise) as illustrated in Fig.  6.2. Additionally, flaws in the various components of the digital detector can lead to image artifacts, which may interfere with the observer’s diagnostic interpretation. Furthermore, measures of the imaging system performance © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_6

67

68

6  Digital Image Quality Descriptors and Performance Characteristics

Fig. 6.1  A digital image matrix consists of small square regions referred to as picture elements or pixels. Each pixel will have a numerical value that is expressed as a shade of gray and therefore the image matrix of numerical values is converted into a gray scale image, for viewing and interpretation

Fig. 6.2  The quality of digital images is characterized by several technical parameters describing the sharpness or detail of the image (spatial resolution), the range of gray levels per pixel (bit depth) that define the brightness of the image (density or contrast resolution), and the number of photons used to produce a useful image (noise). Furthermore, flaws in the various components of the digital detector can lead to image artifacts

with respect to image quality include other more complex parameters such as the modulation transfer function (MTF), the point spread function (PSF), the line spread function (LSF), and the detective quantum efficiency (DQE) [2–4].

6.2  Image Quality Characteristics

69

This chapter will only focus on the essential elements of these image quality descriptors and briefly introduce the system performance parameters.

6.2.1 Spatial Resolution Huda and Abrahams [5] pointed out that “here are several terms used to describe spatial resolution in imaging. Blur and sharpness are good descriptors that are universally understood, and they simply mean that a sharp edge will also appear as sharp (not blurred) in an image obtained with a system that has good resolution performance. Other terms that may be used include resolution, high-contrast resolution, unsharpness, and detail visibility”. This section will only address the bare essentials of spatial resolution in digital radiography detectors. Spatial resolution is the ability of the imaging system to resolve fine details present in an object, It also refers to the sharpness of the image. For digital radiography, the spatial resolution depends on the size of the pixels in the matrix. Smaller pixels will produce images with better spatial resolution compared with larger pixels. The pixel size (PS) can be calculated using the following algebraic expression:



PS 

Field of view  FOV  Matrix size



Where the FOV is the size of the image as illustrated in Fig. 6.3. Thus, for the same FOV, the greater the matrix size, the smaller the pixels and the better the image sharpness, as clearly illustrated in Fig. 6.4. Fig. 6.3  For the same Field-Of-View (FOV), the greater the matrix size, the smaller the pixels and the better the image sharpness. See text for further explanation

70

6  Digital Image Quality Descriptors and Performance Characteristics

Fig. 6.4  The effect of matrix size on picture quality. As the matrix size increases for the same FOV, picture quality improves, that is the image becomes sharper. (Images created by Bruno Jaggi, PEng. Biomedical Engineer; British Columbia Institute of Technology)

The spatial resolution can be measured in line pairs per millimeter (lp/mm). The higher the line pairs/mm, the greater the spatial resolution. The spatial resolution of fluoroscopy, CR, flat-panel digital radiography, and film-screen radiography are 3 lp/mm, 6 lp/mm, 4 lp/mm, and 8 lp/mm, respectively. Using complex measures of spatial resolution (to be introduced in Sect. 6.3), the a-selenium (a-Se detector) shows the best spatial resolution compared with the structured cesium iodide a-­silicon thin film transistor (CsI a-Si TFT) detector. Furthermore, the latter detector produces better spatial resolution than the turbid gandolimium oxysulfide (Gd2O2S) or CsI digital detector [6]. The system components of these detectors are described in Chap. 4.

6.2.2 Contrast Resolution Figure 6.3 shows the contrast of the chest image. Such image contrast shows the density differences or image brightness intensity (black, white, and shades of gray) on various areas of the image. The contrast resolution therefore is the ability of the imaging system to distinguish between differences in image intensity, thus demonstrating the differences in tissue contrast. Such resolution of a digital image is linked to what has been referred to in digital image processing as the bit depth, which is the range of gray levels per pixel. An image with a bit depth of 8 will have 256 (28) shades of gray per pixel. In general, the greater the bit depth, the better the image contrast resolution (grayscale), as shown in Fig. 6.5 (top row). Additionally, one of the processes in digitizing an image is quantization, a process whereby the brightness levels (obtained from sampling the analog signal from the detector) are assigned an integer (zero, a negative, or a positive number) called a gray level. The image is now made up of a range of gray levels. The total number of gray levels is called the gray scale. The greater the quantization, the more accurate the display of gray levels in the image. Figure 6.5 (bottom row) clearly illustrates the effect of the bit depth on image appearance. In general bit depths for digital radiography detectors may range from 10 to 12 bits; however, readers should refer to digital radiography vendors for details of current bit depths used for their detectors.

6.2  Image Quality Characteristics

71

Fig. 6.5  The effect of bit depth on picture quality. As the bit depth increases for the same FOV, picture quality improves, that is the image becomes sharper. (Images created by Bruno Jaggi, PEng. Biomedical Engineer; British Columbia Institute of Technology)

The contrast resolution [size of objects that can be seen on an image (mm at 0.5%)] for nuclear medicine, ultrasound, radiography, computed tomography (CT), and magnetic resonance imaging (MRI) are 20, 10, 10, 4, and 1, respectively [2]. It is clear that MRI offers the best contrast resolution and is able to resolve differences in soft tissue contrast better that all of these imaging modalities.

6.2.3 Noise A third image quality descriptor is noise. There are two kinds of noise, namely, quantum noise and electronic noise. While the latter refers to noise arising from the system (detector) electrical components, the former refers to the photons in the X-ray beam used to produce the image. Quantum noise is important to radiologists and technologists when assessing image quality. On an image, quantum noise appears as mottle or a grainy appearance, compared with a noiseless image as illustrated in Fig. 6.6. Quantum noise is “due to the photon counting statistics, i.e., when more photons are counted in the detector the images will be less noisy” [7] (Fig. 6.6). The number of photons in the beam of X-rays is determined by the exposure technique factors used, namely, the kilovolts (kV) and the milliamperes-seconds (mAs). These photons incident on the patient are attenuated (by the patient) and the photons passing through the patient represent the signal (S, which represents the anatomical structures being imaged) incident on the detector. Compared with high exposure factors, low exposure factors produce fewer photons falling on the detector (less S, and more noise N) thus generating noisy images. Noise is inversely proportional to the square root of the mAs [8]. If the mAs is decreased by a factor of 2, noise will increase by a factor of the square root of 2 or 41%. Increase in

72

6  Digital Image Quality Descriptors and Performance Characteristics

Fig. 6.6  On an image, quantum noise appears as mottle or a grainy appearance, compared with a noiseless image. See text for further explanation

exposure technique factors (more photons) will result in less noisy images, but will result in more radiation dose to the patient. Therefore, it is necessary the operator balance image quality and dose in a radiation protection philosophy of the International Commission on Radiological Protection (ICRP) referred to as low as reasonably achievable (ALARA) [9]. Furthermore, a measure of the signal to noise is a relationship referred to as the signal-to-noise ratio (SNR), a conceptual overview is illustrated in Fig.  6.7. A lower SNR will result in noisy (grainy) images [2, 10].

6.3  Performance Characteristics of Digital Detectors

73

Fig. 6.7  A measure of the signal to noise is a relationship referred to as the signal-to-noise ratio (SNR), a conceptual overview is illustrated in this figure

6.3 Performance Characteristics of Digital Detectors Dr Wolbarst, PhD, a well known medical physicist in writing about measures of image quality and of imaging system capabilities states that “several quantitative parameters have been devised that correlate the abilities of imaging devices to perform clinical tasks… One of the most widely used of these is the modulation transfer function (MTF), a measure of how well the system resolves various levels of detail” [11]. These parameters, including the MTF, the edge spread function (ESF), the line spread function (LSF), and the detective quantum efficiency (DQE), represent more complex measures of the imaging performance of a digital detector, which refers to the “ability of the detector to produce a high-quality X-ray image” [4]. It is not within the scope of this chapter to describe the details of LSF, ESF, and the MTF, and therefore, the interested readers may refer to Yorkston [4], Bushberg et al. [10], and Wolbarst [11]. For biomedical engineering technologists, however, it is important to have a general understanding of the DQE.

6.3.1 Detective Quantum Efficiency The DQE is a metric used to describe the noise characteristics of a digital detector. The overall concept of the DQE for a digital detector is illustrated in Fig. 6.8. First, the X-ray beam falls upon the patient and the transmitted radiation falls upon the

74

6  Digital Image Quality Descriptors and Performance Characteristics

Fig. 6.8  A conceptual overview of the detective quantum efficiency (DQE) for a digital detector is illustrated in this Figure. See text for further explanation

detector, and the incident quanta contains quantum noise. Second, the detector converts the beam into a useful output image. The DQE can be thought of as the efficiency of the digital detector to absorb, utilize, and preserve the X-ray image information. It is the ratio of the signal-to-­ noise out of the detector (image) to the signal-to-noise incident upon the detector (X-ray) expressed algebraically as follows: DQE =

SNR 2out SNR 2in



The DQE for a perfect digital detector is 1 or 100%. This means that there is no loss of information. In general, the higher the DQE for a detector the lower the image noise for a set amount of exposure. Or equivalently, the higher the DQE for a

6.4  Image Artifacts Overview

75

detector the lower the radiation exposure needed to achieve a set amount of image noise [13]. The DQE for digital detectors used in clinical radiology have been evaluated objectively and the results show that indirect conversion CsI a-Si TFT flat-panel detectors offer the highest DQE at low frequencies, however as the spatial frequencies increase, the DQE decreases rapidly [12]. Furthermore, another important fact about DQE for digital detectors used in clinical practice relates to the kV used in the examination. For example, at 70 kV it is about 67% for a-Se and 77% for CsI phosphor. At 120 kV, efficiencies are 37% for a-Se and 52% for CsI. In lower kV applications, such as mammography, a-Se performs better than CsI and has a higher DQE [14].

6.4 Image Artifacts Overview The definitions of an image artifact have been described in detail in the literature. For example, here are two definitions: 1. An artifact is “a distortion or error in an image that is unrelated to the object being imaged” [12]. 2. An artifact is “any false visual feature on a medical image that simulates tissue or obscures tissue” [14]. 3. An artifact is “a feature in an image that masks or mimics a clinical feature” [15]. Artifacts can cause problems for radiologists in their diagnostic interpretation of images. It is imperative therefore that radiologists and technologists be able to not only identify artifacts but also understand how they arise and how they can be reduced or removed from the image. Artifacts can arise from at least three categories of the imaging system, namely, the digital detector, the software, and the object being imaged. While the sources of detector artifacts are dust, dirt, scratches, pixel malfunction and ghost images; software artifacts arise from the histogram, range/scaling, and image compression for example. Object artifacts arise from patient positioning, collimation, and back scatter, for example [2]. For more information on artifacts, readers are encouraged to explore Bushong [2] and Seeram [3]. A pictorial review of computed/digital radiography artifacts and their description is provided in an article by Walz-Flannigan et  al. [16] Artifacts illustrated include detector image lag or ghosting, incorrect detector orientation, that is, upside-­ down cassette, backscatter, stitching artifacts, over exposure, dead pixel artifact, signal dropout, speckled radiopaque spots, detector calibration limitation, failure of detector offset correction, electronic shutter failure, values of interest misread, mid grey clipping, and grid-line suppression failure.

76

6  Digital Image Quality Descriptors and Performance Characteristics

References 1. Lanca L, Silva A. Digital imaging systems for plain radiography. New York: Springer; 2013. 2. Bushong S. Radiologic science for technologists. 12th ed. Saint Louis: Elsevier; 2021. 3. Seeram E.  Digital radiography: physical principles and quality control. 2nd ed. Singapore: Springer Nature; 2019. 4. Yorkston J. Flat-panel DR detectors for radiography and fluoroscopy. In: Goldman L, Yester MV, editors. Specifications, performance, and quality assurance of radiographic and fluoroscopic systems in the digital era, AAPM monograph 30. College Park, MD: AAPM; 2004. p. 177–229. 5. Huda W, Abrahams RB.  X-ray-based medical imaging and resolution. Am J Roentgenol. 2015;204:W393–7. 6. Yorkston J. Senior research scientist, clinical applications research, Carestream Health Inc., New York. Personal communications; 2017. 7. Nett B. X-ray contrast to noise (CNR) illustrated examples of image noise (SNR, quantum mottle) for radiologic technologists. How Radiology Works LLC; 2023. https://howradiologyworks.com/x-­ray-­cnr/ 8. Cody DC, McNitt-Gray MF. CT image quality and patient radiation dose: definitions, methods, and trade-offs. RSNA categorical course in diagnostic radiology physics: from invisible to visible-the science and practice of X-ray imaging and radiation dose optimization; 2006. p. 141–55. 9. Seeram E. Rad Tech’s guide to radiation protection. 2nd ed. Hoboken: Wiley-Blackwell; 2020. 10. Bushberg JT, Seibert JA, Leidholdt EM Jr, Boone JM. The essential physics of medical imaging. 4th ed. Philadelphia: Wolters-Kluwer; 2021. 11. Wolbarst AB. Physics of radiology. 2nd ed. Madison: Medical Physics Publishing; 2005. 12. Seibert JA. Computed radiography/digital radiography: adult. In: Frush DP, Huda W, editors. From invisible to visible-the science and practice of X-ray imaging and radiation dose optimization: RSNA categorical course in diagnostic radiology physics. Oak Brook: RSNA; 2006. p. 57–71. 13. Krugh K, Ph.D., Dept. of Radiology, The Toledo Hospital. Personal Communications; 2019. 14. Spahn M. Flat detectors and their clinical applications. Eur Radiol. 2005;15:1934–147. 15. Willis CE. Computed radiography/digital radiography: pediatric. In: Frush DP, Huda W, editors. From invisible to visible-the science and practice of X-ray imaging and radiation dose optimization: RSNA categorical course in diagnostic radiology physics. Oak Brook: RSNA; 2006. p. 78–83. 16. Walz-Flannigan AI, Brossoit KJ, Magnuson DJ, Schueler BA.  Pictorial review of digital radiography artifacts. Radiographics. 2018 May–Jun;38(3):833–46. https://doi.org/10.1148/ rg.2018170038. Epub 2018 Apr 20. PMID: 29676963.

Chapter 7

Computed Tomography: Basic Physics and Technology

7.1 Introduction Computed tomography (CT) uses the power of X-rays transmitted through the patient to specialized detectors. These detectors transform the radiation beam into digital data, facilitating computer processing. The computer uses sophisticated algorithms called image reconstruction techniques to build up and display images of a patient’s internal anatomy for diagnostic interpretation. These cross-sectional images are planar sections, or slices, that are perpendicular to the long axis of the patient. Computed tomography (CT) is an extraordinary invention made possible through the work of several individuals, most notably Godfrey Newbold Hounsfield and Allan MacLeod Cormack. For their work, Hounsfield and Cormack shared the 1979 Nobel Prize in Medicine or Physiology [1]. The technology used in CT has evolved from scanning a single slice to multiple slices [multislice scanner (MSCT)] in a single breath-hold. Current state-of-the-art CT systems are based on volume data acquisition, in which the X-ray tube and detectors rotate continuously around the patient to gather transmission data from a volume of tissue rather than from 1 slice at a time. Volume-based technology expanded CT use to sophisticated applications in diagnostic CT imaging, along with nuclear medicine and radiation therapy [2, 3]. The increasing use of CT has led to widespread concern about high patient radiation doses from CT examinations relative to other radiography examinations. Several efforts have focused on how to reduce patient dose and operate within the as low as reasonably achievable (ALARA) principle. Sections of the article entitled CT: A Technical Review by Euclid Seeram, (Radiologic Technology, 2018, 89:3) have been reprinted with permission from American Society of Radiologic Technologists for educational purposes. ©2020. All rights reserved.

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 E. Seeram, X-Ray Imaging Systems for Biomedical Engineering Technology, https://doi.org/10.1007/978-3-031-46266-5_7

77

78

7  Computed Tomography: Basic Physics and Technology

The purpose of this chapter is to describe the essential physical principles and technical aspects of CT, including physics related to radiation attenuation and CT numbers along with general technical concepts. Additionally, the chapter will address the major principles of current state-of-the-art multislice CT technology.

7.2 Essential Physical Principles CT equipment acquires images using three primary system components (see Fig. 7.1). These include data acquisition, image reconstruction, and image display, storage, and communications. Data acquisition is the component in which the technology systematically scans a patient. Data acquisition involves an X-ray tube coupled to special electronic detectors. Image reconstruction creates images using sophisticated computer algorithms. Display, storage, and communication of processed images rely on digital technology. Images are stored on magnetic or optical data carriers and can be communicated by electronic means to other locations. Finally, the images can be sent to remote locations using Picture Archiving and Communications System (PACS). An important element of data acquisition is that the X-ray tube and detectors rotate around the patient to systematically collect X-ray attenuation data, which is subsequently transmitted to the computer for the reconstruction of CT images, as illustrated in Fig. 7.2.

Fig. 7.1  The computed tomography (CT) scanner consists of three primary system components for image production: data acquisition, image reconstruction, and image display, recording, storage, and communication systems. (Images courtesy of Siemens Healthcare GmbH)

7.2  Essential Physical Principles

79

Fig. 7.2  CT data collection. The X-ray tube and detectors rotate around the patient to systematically acquire X-ray attenuation data that are subsequently sent to the computer for image reconstruction. (© ASRT 2017)

7.2.1 Radiation Attenuation Radiation Attenuation is the reduction of the radiation beam’s intensity as it passes through an object. This process involves the absorption of some photons by the object, accompanied by scattering of other photons. Attenuation in CT depends on the effective atomic density (atoms/volume), the atomic number of the absorber, and the photon energy. 7.2.1.1 Beer–Lambert Law The attenuation of a homogeneous beam, where all the photons possess the same energy, follows an exponential pattern, which is characterized by the equation shown in Fig. 7.3. During development of the CT scanner, Hounsfield used a gamma radiation source, not an X-ray tube. The gamma radiation emitted a homogeneous beam in his initial experiments because such a beam satisfies the requirements of the Beer–Lambert law [4–6]. This law is an exponential relationship described by the equation:

I  I o e  x

80

7  Computed Tomography: Basic Physics and Technology

Fig. 7.3  Attenuation of a homogeneous beam of radiation is exponential and is described by the Beer–Lambert law. See text for further explanation. (Figure courtesy of the author)

where I is the transmitted intensity, Io is the original intensity, e is Euler constant (2.718), μ is the linear attenuation coefficient, or the fractional reduction in the intensity of a beam of radiation per unit thickness of the medium traversed, and x is the thickness of the object. The fundamental problem in CT is to calculate the linear attenuation coefficient, expressed as μ, which indicates the amount of attenuation that has occurred. The equation I = Ioe−μx can be solved to find the value of μ as follows:



I  Ioe x ln

Io  x I

I 1     · ln o I x

where ln is the natural logarithm. In CT, however, the beam is attenuated by a given amount of tissue with a specific thickness Δx. Hence, the attenuation is expressed as follows:

I  I o e  x

As the CT scanner’s X-ray beam passes through a stack of volume elements, or voxels (see Fig. 7.4) that is part of the slice, the scanner obtains an attenuation measurement called a ray sum [4–6]. The ray sum is the total of all linear attenuation coefficients along the path of a single ray through the patient. In this situation, the transmitted intensity I is represented as follows [5]:   X I  I o e  i1 i n

where

7.2  Essential Physical Principles

81

Fig. 7.4  As a single ray of the X-ray beam passes through a stack of voxels, the CT system obtains an attenuation measurement called a ray sum. A ray sum is the sum of all μ’s along the path of a single ray through the patient. (Figure courtesy of the author)

 i 1i X    1  2  3  n  X n





Taking the natural logarithm, this equation can be written as [5]:



I ln  o  I

n     i 1i X 

Calculating linear attenuation coefficients in CT is challenging because the system must calculate each of the coefficients in the entire slice. The values of transmitted intensity and original intensity are known and are measured by detectors. Furthermore, change in thickness also is known so the system can calculate the linear attenuation coefficient. During image reconstruction process, the system obtain many ray sums for different locations (rotation angles) around the slice. Hounsfield faced a significant problem in his early experiments with the gamma source of radiation. The beam intensity was too low, leading to lengthy scanning times for image production [1]. Hounsfield subsequently switched to a diagnostic

82

7  Computed Tomography: Basic Physics and Technology

X-ray tube. The beam from this tube is heterogeneous in which photons have different energies; therefore, it does not satisfy the Beer–Lambert law. As a result, Hounsfield had to make several assumptions and adjustments to determine the linear attenuation coefficients. It is necessary to make the heterogeneous beam in CT approximate a homogeneous beam to satisfy the equation and solve for linear attenuation coefficient. 7.2.1.2 Attenuation and CT Numbers Each CT imaging slice comprises a matrix of voxels (e.g., 512 × 512 or 1024 × 1024). Figure 7.5 illustrates the attenuation process through a single voxel. The attenuation value is converted into an integer (0, a positive number, a negative number) known as a CT number, or Hounsfield Units (HU) in honor of the inventor Godfrey Hounsfield. The system subsequently normalizes all voxel values to the attenuation of water (μwater). CT numbers are computed using the following algebraic expression: CT Number 

Fig. 7.5 Attenuation through a single voxel. The attenuation value is converted into an integer (0, a positive or negative number) referred to as a CT number. (Figure courtesy of the author)

tissue  water ·K  water



7.2  Essential Physical Principles

83

where K is the scaling factor (contrast factor) of the CT manufacturer. In general, K is equal to 1000. The range of CT numbers produced using this factor is referred to as the Hounsfield (H) scale, which results in 0.1% per CT number, thus expressing linear attenuation coefficients much more accurately compared with the scale used in pioneering CT units. CT numbers are always computed with reference to the attenuation of water. The CT number for water is 0, while it is +1000 for bone and −1000 for air on the Hounsfield scale. Refer to the table for CT number ranges for bone, muscle, white matter, gray matter, blood, tumors, water, fat, lungs, and air. These values can be printed out as a numerical image. However, radiologists interpret grayscale images, which involves converting the numerical image into a grayscale image as shown in Fig. 7.6. CT numbers are referred to as gray levels [1] in digital image processing. Converting the numbers into shades of gray (grayscale) results in higher numbers assigned white, lower numbers black, and gray shades between black and white. This assignment is related to the attenuation characteristics of tissues. Bone attenuates more radiation and therefore is assigned white (the bone’s appearance is the same on a film-screen image as it is on a digital image). Air attenuates very little radiation, appearing black on film-screen and digital images.

Fig. 7.6  The matrix of CT numbers (numerical image = gray levels) must be converted into a gray-scale image for viewing by radiologists and technologists. (Illustration courtesy of the author)

84

7  Computed Tomography: Basic Physics and Technology

The range of CT numbers is defined as the window width and the center of the range is defined as the window level [1]. Finally, the radiologist can manipulate the window width to alter image contrast and the window level to alter image brightness.

7.3 Technology CT technology refers to the basic equipment configuration that consists of three primary systems (Fig. 7.1): data acquisition, image reconstruction, and image display, storage, and communication systems.

7.3.1 Data Acquisition Principles and Components Data acquisition is a systematic collection of attenuation data from the patient through various scanning methods. These methods have evolved from scanning one slice at a time during a single rotation of the X-ray tube and detectors to scanning multiple slices per single rotation of the X-ray tube and detectors around the patient. MSCT scanners are now common and utilize fan-beam geometry to scan tissue volumes. The term geometry, or data acquisition geometry, refers to the size, shape, motion, and path traced by the X-ray beam [4]. The path is created as the patient moves through the gantry aperture during scanning, creating a spiral or helical path. MSCT scanners operate based on spiral/helical beam geometry, as illustrated in

7.3 Technology Fig. 7.7  Multislice CT (MSCT) scanners use fan-beam geometry to scan the volume of tissue. (Illustration © ASRT 2017)

85

X-ray tube

Fan-beam geometry

Patient

Detectors

Spiral/helical beam path Patient movement during scanning

Volume of tissue scanned

Slices selected at arbitrary locations within the scanned volume of tissue

Fig.  7.7. Over time, MSCT scanners have evolved incrementally from 4 to 640 slices per revolution of the X-ray tube. The helical geometry improves the volume coverage speed without compromising image quality. Various components housed in the scanner gantry play roles in data acquisition (Fig. 7.8). The X-ray tube and detectors rotate around the patient to collect multiple attenuation readings. To ensure uniformity at the detector and comply with Beer– Lambert’s law for calculating linear attenuation coefficients, the X-ray beam passes through a specially shaped filter known as bow-tie filter. Collimation directs the beam only through the slice of interest, and detectors measure transmitted photons, which are then converted into digital data by the analog-to-digital converters. Subsequently, this digital data are sent to the computer for image reconstruction.

7.3.2 Image Reconstruction Image reconstruction involves complex mathematics, the reason Godfrey Hounsfield shared his Nobel Prize with physicist and mathematician Allan Cormack. Images are reconstructed by producing a 2D “distribution (usually of some physical

86

7  Computed Tomography: Basic Physics and Technology

Fig. 7.8  Primary components of a CT scanner data acquisition system: the X-ray tube, filter, collimator, patient, detectors, and detector electronics that include the analog-to-digital converter. (Illustration © ASRT 2017. CT scanner image courtesy 2017 Siemens Healthcare GmbH)

property) from estimates of its line integrals along a finite number of lines of known locations” [7]. In CT, the physical property is the linear attenuation coefficient of the tissues, and the line integral refers to the sum of the attenuation along each ray in the beam that passes through the slice [7]. Image reconstruction uses algorithms, or defined rules for solving a problem, to systematically build an image during the scanning process. The algorithms used in CT include the earlier back projection algorithm and newer analytic reconstruction techniques such as the filtered back projection algorithm. Filtered back projection consists of two main components: convolution filtering and subsequent back projection of the filtered profiles (Fig. 7.9) [8, 9]. In the early back projection algorithm, multiple back projections of data obtained at various angles are summed until the complete image is reconstructed. However, this results in a blurry and unusable final reconstructed image. The filtered back projection algorithm was developed to solve image blurring by applying convolution filtering (digital filter) to each set of projection data before back projection. This process enhances sharpness and contrast [8–12]. These convolution filters increase noise in images, especially in low-dose CT examinations, and are thus viewed as a limitation of the filtered back projection algorithm. Other limitations of the filtered back projection algorithm are based on underlying assumptions about scanner geometry, such as the use of pencil X-ray beam geometry, a point source X-ray focal spot (which is an approximation based on the assumption of its infinitely small size), failure to consider the shape and size of detector cells and voxels, and neglecting image noise resulting from Poisson statistical variations of X-ray photons [9, 12].

7.3 Technology

87

Fig. 7.9  Flowcharts showing the primary steps of the filtered back projection algorithm (a) and a generic iterative reconstruction algorithm (b). Diagram courtesy of the author. (CT images in (b) reprinted with permission from Kim M et al. Adaptive iterative dose reduction algorithm in CT: effect on image quality compared with filtered back projection in body phantoms of different sizes. Korean J Radiol. 2014;15 (2):195–204)

7.3.3 Iterative Reconstruction The introduction and widespread use of iterative reconstruction algorithms resulted from shortcomings of the filtered back projection algorithm [8–11]. With the development of MSCT scanning, other algorithms were necessary to address the problems associated with the patient moving through the CT gantry during the scanning process as the X-ray tube and detectors rotate continuously around the patient.

88

7  Computed Tomography: Basic Physics and Technology

Iterative reconstruction algorithms, which first were used in the 1970s but were impractical for clinical use because of a lack of computational power, reemerged and now are commonplace in CT scanners. These algorithms solve the problems of image artifacts and image noise generated by the filtered back projection algorithm, especially in low-dose CT scanning. Iterative reconstruction algorithms model the CT system more accurately and subsequently improve image quality, especially for low-dose CT examinations. Iterative reconstruction algorithms therefore can reduce image noise while preserving image sharpness and contrast, especially at low tube currents. Iterative reconstruction algorithms also help reduce metal artifacts and beam hardening effects caused by photon starvation [9, 10, 12, 13]. Computed tomography manufacturers provide several different iterative reconstruction algorithms with their CT scanners. For example, GE Healthcare offers Adaptive Statistical Iterative Reconstruction (ASiR) and Model-Based Iterative Reconstruction (MBIR)3, Siemens Healthineers offers Iterative Reconstruction in Image Space (IRIS) and Sinogram Affirmed Iterative Reconstruction (SAFIRE). In addition, Philips offers iDose4, and Toshiba Medical Systems offers Adaptive Iterative Dose Reduction (AIDR), along with 3D AIDR. A generalized iterative reconstruction process is illustrated in Fig. 7.9, including the following steps [10, 12]: 1. During scanning, iterative reconstruction acquires measured projection data and then reconstructs the data using the standard filtered back projection algorithm to produce an initial image estimate. 2. Then, the initial image estimate is forward projected to create simulated projection data (artificial raw data) that then are compared with the measured projection data. 3. The algorithm then determines differences between the two sets of data to generate an updated image that is back projected on the current CT image; this minimizes the difference between the current CT image and the measured projection data. 4. The user must evaluate image quality based on a predetermined criterion included in the algorithm. 5. If the image quality criterion is not met, the iteration process repeats several times in an iterative cycle until the difference is considered sufficiently minimal. 6. The final CT image matches quality criteria after the termination of the iterative cycle. A comparison of images reconstructed using filtered back projection and an iterative reconstruction algorithm is shown in Fig. 7.9. Iterative reconstruction algorithms are common in modern-day MSCT scanners, especially for optimizing dose in children.

7.4  Multislice CT (MSCT): Fundamental Principles and Technology

89

7.3.4 Image Display, Storage, and Communication The third main component of CT systems involves displaying the reconstructed image for viewing and interpretation, with postprocessing to suit the needs of the interpreting radiologist. Additionally, images are sent to PACS for storage and communication to virtual data centers for retrospective analysis. For effective electronic communication in CT, a standard protocol is required to facilitate connectivity (networking) among imaging modalities and equipment from various manufacturers. The standard employed for this purpose is Digital Imaging and Communication (DICOM).

7.4 Multislice CT (MSCT): Fundamental Principles and Technology MSCT has evolved from single-slice CT, overcoming the limitations of conventional CT with its stop-and-go, slice-by-slice data acquisition. MSCT achieves this by reducing data acquisition time, thereby increasing volume coverage speed. MSCT offers advantages such as isotropic spatial resolution, efficient X-ray beam utilization, reduced radiation exposure, and improved accuracy in procedures like needle placement for CT fluoroscopy and cardiac CT imaging. The strategy in MSCT involves continuously acquiring data as the patient moves through the gantry, with the X-ray tube and detectors rotating around the patient. This approach necessitates special equipment considerations and various technical approaches such as [1, 4, 14]. • • • • •

Slip-ring technology and X-ray tubes that provide very high X-ray output. Interpolation algorithms. 2D detector arrays Continuous table movement. Mass computer memory buffer.

7.4.1 Slip-Ring Technology Slip-ring technology enables scanning of a patient moving through the gantry, while the X-ray tube and detectors rotate continuously around the patient. Slip rings are electromechanical devices that transmit electrical energy across a rotating interface through circular electrical conductive rings and brushes [18].

90

7  Computed Tomography: Basic Physics and Technology

7.4.2 X-Ray Tube Technology A fundamental problem with conventional X-ray tubes is heat dissipation and slow cooling rates. Furthermore, as gantry rotation times increase, higher milliampere (mA) values are needed to provide the same mA per rotation. As the electrical load (mA and kilovolts, or kV) increases, the X-ray tube requires faster anode cooling rates. MSCT scanners require X-ray tubes that can sustain higher power levels because the tube rotates continuously for a longer period compared with conventional scanners. Several technical advances in component design have resolved these power levels and dealt with the problems of heat generation, storage, and dissipation. For example, the tube envelope, cathode assembly, anode assembly (including anode rotation), and target have been redesigned [15, 16].

7.4.3 Interpolation Algorithms In helical CT data acquisition, movement of the patient through the gantry aperture while the tube rotates around the patient can cause problems, including [1, 4, 6]. • Difficulty localizing a specific slice because there is no defined slice as in conventional CT scanning, as is illustrated in Fig. 7.10. • Slice volume geometry that is somewhat different than in conventional CT scanning (see Fig. 7.10). • Increased effective slice thickness because of the width of the fan beam and the speed of the table.

Fig. 7.10  A main difference between conventional slice-by-slice CT scanning (a) and spiral/helical CT scanning (b) lies in the geometry of the slice. (Illustration © ASRT 2017)

7.4  Multislice CT (MSCT): Fundamental Principles and Technology

91

• Inconsistent projection data because there is no defined slice. This is important because consistent data are needed to satisfy the filtered back-projection algorithm. If inconsistent data are used with this algorithm, the image displays streak artifacts similar to motion artifacts. Image reconstruction steps can overcome these problems. First, the system produces a planar section similar to the slice in conventional CT scanning. The planar section is generated using interpolation algorithms (Fig.  7.10). Interpolation is a mathematical technique to estimate the value of a function from known values on either side of the function. Second, the images are reconstructed using the standard filtered back-projection algorithm. The results are CT images free of motion artifacts [1].

7.4.4 MSCT Detector Technology The detector is one of the most important components in the CT imaging chain because it captures radiation passing through the patient and converts it into electrical signals that subsequently are digitized and sent to the computer for processing and image building. Currently, two categories of detectors capture and convert radiation into digital data. These include scintillation detectors (energy integrating detectors and dual-energy detectors) and photon-counting detectors, as illustrated in Fig. 7.11. The Energy Integrating Detector (EID) is based on the use of scintillation crystals that convert X-ray photons to light photons, which are then converted to electrical signals by photodiodes. Detector electronics called application-specific integrated circuits (ASIC) digitize the signals [17]. Scintillation crystals used in MSCT include cadmium tungstate (CdWO4), ceramic material made of high-purity, rare-earth oxides based on doped rare-earth compounds such as yttria, and gadolinium oxysulfide ultrafast ceramic. GE Healthcare has made use of gemstone spectral imaging. The GE Gemstone is the first garnet scintillator for use in CT. Furthermore, Philips Healthcare uses zinc selenide activated with tellurium in their dual-layer scintillator detectors [17]. The photon-counting detector is an emerging technology that is being tested in prototype scanners such as the Siemens SOMATOM Definition Flash. These detectors use semiconductors such as cadmium telluride (CdTe) and cadmium zinc telluride (CZT) [18] because they can convert x-ray photons directly into electron hole pairs (electric charge) [19]. One of the most important differences between single-slice CT and MSCT is the design of the detector. Single-slice detectors are based on a 1-detector array design, and MSCT detectors are based on a 2D design as illustrated in Fig. 7.12. Whereas 1D detector arrays acquire 1 slice per single rotation of the x-ray tube and detectors, 2D detectors acquire several slices per single rotation of the X-ray tube and detectors during volume scanning. For example, a 2D detector with 64 detector rows acquires 64 slices per single rotation.

92

7  Computed Tomography: Basic Physics and Technology

Fig. 7.11  The main components of two types of scintillation detectors: conventional energy integrating detectors (a) and dual-layer detectors (b). The main components of the photon-counting detector or direct conversion are shown in (c). ASIC application-specific integrated circuit. (Image courtesy of the author)

Fig. 7.12  A main difference between single-slice CT and MSCT is the detector design. Single-­ slice detectors are based on a 1D detector array design (a), and MSCT detectors are based on a 2D design (b). (© ASRT 2017)

7.4  Multislice CT (MSCT): Fundamental Principles and Technology

A

B

X-ray tube

16 1.25 mm

ADC

ADC

ADC

X-ray tube

4 1.5 mm

Matrix array detector

16 Detector Rows

ADC

93

16 1.25 mm

4 1.5 mm

24 Detector Rows

ADC

ADC

ADC

Adaptive array detector

ADC

Fig. 7.13  Two types of 2D detector arrays used in MSCT scanners: the matrix array (a) and the adaptive array (b), and examples of binning the detector channels to obtain different slice thicknesses. ADC analog-to-digital converter. (Images courtesy of the author)

Two types of 2D detector arrays are shown in Fig. 7.13: matrix array and adaptive array. The detector designs are based on preferences of the CT manufacturer. The matrix array detector consists of equal detector elements, while the adaptive array detector features pairs of equal elements. For instance, the two central elements are equal, and the two elements closest to the central elements are also equal to one another. Slice selection using 2D detectors depends on the detector configuration. According to Dalrymple et al., this configuration “describes the number of data collection channels and the effective section thickness determined by the data acquisition system settings” [20]. For example, in Fig. 7.13a, each detector channel in the matrix array detector is 1.25 mm and four cells are activated or grouped (binned) together to produce four separate images, each with a thickness of 1.25 mm, per 360° rotation.

7.4.5 Selectable Scan Parameters The operator can select several MSCT scanning parameters. These include the scan mode, exposure factors (kV, mA, and scan time), gantry rotation time, pitch, scan length, collimation, and slice width. In particular, pitch determines the required image quality and is related to patient dose. As pitch increases, the dose decreases, but there is a trade-off with degradation in image quality, as illustrated in Fig. 7.14. Patient dose is inversely proportional to the pitch. The current universal definition of pitch for MSCT scanner technology is from the International Electrotechnical Commission [21]: the pitch (P) is equal to the

94

7  Computed Tomography: Basic Physics and Technology

Fig. 7.14  The effect of pitch on dose and image quality in MSCT. As the pitch increases, the dose decreases with a corresponding degradation of image quality. (© ASRT 2017)

distance the table travels per rotation (d)/total collimation (W). The total collimation is equal to the number of slices (M) times the collimated slice thickness (S). Algebraically, the pitch is expressed as follows:



P

d d or P  M S W

7.4.6 Isotropic CT Imaging MSCT scanners have evolved from 4 to 640 slices per revolution of the X-ray tube and detectors around the patient. The purpose of this ongoing evolution coupled with other technical developments is to achieve isotropic spatial resolution, in which all dimensions of the voxel (x, y, z) are equal, and therefore, the voxel is a perfect cube. Matrix array detectors are considered isotropic in design (cells equal in all dimensions). If all voxel dimensions are not equal, that is, the slice thickness is not equal to the pixel size, the data set acquired is anisotropic. Figure 7.15 shows the geometry of both isotropic and anisotropic voxels. The overall goal of isotropic imaging in CT is to achieve excellent spatial resolution in all imaging planes, especially when reconstructing multiplanar and 3D images. Technologists should be aware that as the voxel size decreases, radiation dose increases [20].

7.5  MSCT Image Processing

95

Fig. 7.15  The geometry of isotropic and anisotropic voxels. The overall goal of isotropic imaging in CT is to achieve excellent spatial resolution in all imaging planes, especially when reconstructing multiplanar and 3D images. (Figure courtesy of the author)

7.5 MSCT Image Processing Computer architectures for MSCT have evolved from an early pipeline processing system to the use of parallel and distributed processing architectures to accommodate more sophisticated processing and clinical application tasks. The specific architecture chosen for an MSCT system depends on how the computer assigns various tasks, such as preprocessing raw data, image reconstruction, display tasks such as 3D or virtual reality imaging, and purpose, such as CT angiography. Numerous processors in the computer’s electronic circuits complete processing tasks. The MSCT computer rapidly processes the large datasets acquired and the high number of iterations needed to complete the computationally intensive reconstruction operations of CT iterative reconstruction algorithms. The need for rapid processing challenges the capacity of the central processing units of modern-­day computers used in medical imaging and radiation therapy. Therefore, CT computer architecture also includes a graphics processing unit to reduce the processing requirements of the central processing unit. The graphics processing unit now contributes to image reconstruction, image processing, dose calculation and treatment plan optimization, radiation treatment planning, and other applications. Current graphics processing units can construct complex 3D images in milliseconds [22].

96

7  Computed Tomography: Basic Physics and Technology

7.5.1 Image Postprocessing Three classes of software contribute to CT systems. These include image reconstruction, preprocessing, and image postprocessing software. Preprocessing software corrects data collected from the detectors before the data are sent to the reconstruction computer, and image postprocessing software involves manipulating reconstructed images for viewing. The CT processing system is so sophisticated that it has been called visualization and analysis software and consists of basic and advanced image display and analysis tools. Image postprocessing operations are called image enhancement [23, 24]. The purpose of image enhancement is to display an image according to some interpreter preferences and assist with diagnostic interpretation. Image enhancement operations include contrast and brightness enhancement, edge enhancement, spatial and frequency filtering, image combining, and noise reduction. 7.5.1.1 Windowing The most popular of these operations for use in medical imaging is contrast and brightness enhancement, also called windowing. The range of numbers in the image is called the window width (WW) and the center of the range of numbers is called the window level (WL). Window width controls image contrast, and window level controls the image brightness. A wider window width creates an image with less contrast, but a narrow window width provides improved image contrast. As the window level increases and the window width remains constant, the image becomes darker because more of the lower numbers are displayed. Image contrast and brightness changes are illustrated in Figs. 7.16 and 7.17, respectively.

Fig. 7.16  A graphic illustration of the effect of window width (WW) on a CT image. Window width controls image contrast, and window level controls image brightness. A wide window width shows an image with less contrast (a), and a narrow window width improves image contrast (b). (Images courtesy of the author)

7.5  MSCT Image Processing

97

Fig. 7.17  The effect of window level (WL) on image brightness. As the window level increases (keeping the window width constant [e.g., 1500]), the image becomes darker because more of the lower numbers are displayed. A lower WL (e.g., 2500) provides a brighter image because more of the higher CT numbers are displayed. (Figure courtesy of the author)

7.5.2 Three-Dimensional Image Display Techniques The ability to display 3D CT images has applications in evaluating the craniomaxillofacial complex, musculoskeletal system, central nervous system, and cardiovascular, pulmonary, gastrointestinal, and genitourinary systems. Display of CT scans in 3D now is commonplace in CT angiography [25–29]. The use of 3D visualization techniques is a type of digital image processing operation called image synthesis. The purpose of image synthesis is to create images that are useful when the desired image is impossible or impractical to acquire [30]. The 3D display techniques are based on computer graphics principles and technology. At least four major elements are essential to creating 3D images (Fig. 7.18). The elements include CT image acquisition, creation of 3D space, processing for 3D image display, and 3D image display for the interpreting physician or observer. The data collected from scanning the patient are used to create 3D space, which contains all voxel information and is processed for 3D image display. Processing 3D CT images involves preprocessing, visualization, manipulation, and analysis. Surface and volume rendering techniques can transform 3D space into simulated 3D images for display on a 2D computer monitor [25]. Surface rendering is a relatively simple operation in which the surface of an object is created using contour data and shading of pixels to provide the illusion of depth. Surface rendering uses only 10%

98

7  Computed Tomography: Basic Physics and Technology

Fig. 7.18  Creating 3D CT images involves four primary steps: image acquisition, creation of 3D space, processing for 3D image display, and 3D image display for the interpreting physician. MIP maximum intensity projection, MinIP minimum intensity projection

of the data in 3D space and requires little computation. Volume rendering, on the other hand, is much more sophisticated; the technique uses all the data in 3D space and requires more computational power. Maximum and minimum intensity projections, for example, are volume rendering techniques that make use of only those voxels in 3D space that have the maximum (brightness) or minimum values. These values are assigned to the pixels in the displayed maximum and minimum intensity projection images, which are used extensively in CT angiography.

7.6 Image Quality CT image quality, essential for diagnostic interpretation, is characterized by at least five physical parameters including spatial, contrast, temporal resolution, noise, and artifacts [31, 32].

7.6  Image Quality

99

Spatial resolution refers to the fine detail or sharpness of objects seen in an image. More specifically, spatial resolution, also called high-contrast resolution, is a CT scanner’s ability to resolve closely spaced objects that are significantly different from their background, or to show small objects that have high subject contrast. Detail in digital images is influenced by the matrix size, field of view (FOV), and slice thickness. As a result, smaller pixel sizes provide images with greater detail for the same FOV. The pixel size can be calculated as follows:



Pixel size  p  

FOV Matrix size

As slices become thinner, the degree of image sharpness increases. Radiography still demonstrates the best spatial resolution of all imaging modalities because the images can show object sizes of 0.1 mm compared with 0.5 mm for CT and magnetic resonance (MR) images, 2 mm for diagnostic medical sonograms, and 5 mm for nuclear medicine images [33]. Contrast resolution refers to the CT scanner’s ability to display images with small differences in soft tissues expressed in millimeters at 0.5%. CT has much better contrast resolution than radiography, nuclear medicine, or sonography. However, MR imaging has superior contrast resolution to all imaging modalities. Contrast resolution for MR is 1 mm, compared with 20 mm for nuclear medicine, 10 mm for sonography, and 10 mm for radiography [33]. Imaging moving organs, such as a beating heart, has become routine; however, organ movement can cause imaging blurring. Temporal resolution describes the CT techniques used to detect movement over time and freeze the motion of organs. The techniques address the speed of data acquisition and are significant for cardiac CT and other applications that require reducing the effects of motion on images to achieve diagnostic quality [1]. Noise is a random variation of CT numbers in the image of a water phantom. Several factors affect CT image noise, including the number of X-ray photons reaching the detectors, beam energy, voxel size, and slice thickness. For example, larger voxels in anisotropic imaging and thicker slices generate more noise than images with smaller voxels and thinner slices. A noisy image displays a grainy appearance (Fig. 7.15) [1]. Artifacts are distortions or errors in an image that are unrelated to the object being imaged. Any discrepancy between the reconstructed CT image and the true attenuation coefficients of the object also is considered an artifact [34]. As noted by Barrett and Keat [34], the artifacts originate from one of several categories or causes: physics-based artifacts from the physical processes involved in CT data acquisition, patient-based artifacts from factors such as patient movement or metallic materials, scanner-based artifacts from scanner function imperfections, and helical and multisection artifacts, which are produced during image reconstruction.

100

7  Computed Tomography: Basic Physics and Technology

7.7 Dual Energy CT Dual-Energy CT (DECT), also referred to as Spectral CT, has emerged as a tool in CT imaging. In DECT, two sets of data from the same patient are utilized, one imaged at 80 kV (low kV) and the other at 140 kV (high kV) by rapidly switching from 80 kV to 140 kV on a view-by-view basis, thus using the physics of attenuation to image materials with distinct attenuation properties at different energy levels. The scanner employs either two X-ray tubes, each coupled to its own bank of detectors in the gantry to collect low-kV and high-kV datasets [often called Dual-­ Source CT (DSCT)] or a dual-layer detector with the top layer acquiring low-energy photons and bottom layer collecting high-energy photons (often called dual-layer approach). “The first two techniques are source-oriented approach and the last one is a detector-oriented approach. The key is to ensure the mean energies of the two datasets are significantly different so that they provide complimentary information for the reconstruction” [1].

7.8 Photon Counting CT The first photon counting detector CT scanner (PCDT, the Siemens NAEOTOM Alpha) was approved by the Food and Drug Administration (FDA) in September, 2021 [35]. This scanner overcomes the limitations energy integrating detectors such as poor spatial resolution, increase in image noise, and problems with DECT imaging. The advantages of photon counting detectors (PCDs) compared to EIDs are reduction of electronic noise, better spatial resolution (Fig. 7.19) weighting of photons, improved contrast resolution, low-dose imaging, and material specific imaging.

7.8.1 The Major System Components of a PCD The major system components are shown in Fig. 7.20a, compared to that of an EID circuit (Fig. 7.20b). The PCD circuit includes a semiconductor coupled to an electrical processing circuit. Semiconductors such as cadmium telluride (CdTe) or cadmium zinc telluride (CdZnTe or CZT for short) have high Z and high absorption efficiency and are therefore used in PCD CT imaging. Silicon is also a semiconductor and has been evaluated for use as a PCD. When the semiconductor is exposed to x-ray photons, electron-hole pairs (called charge cloud) are produced. The charge cloud is proportional to the energy of the photon striking the semiconductor, and is attracted to the anode, called a pixel electrode. Subsequently these charge clouds produce electrical signals or pulses which are sent to the electrical processing circuit, a very fast counting ASIC which consists of amplifiers, comparator circuits and counters (Seeram [1]).

7.8  Photon Counting CT

101

Fig. 7.19  Comparison of spatial resolution in the phantom images between UHR mode of the PCD CT and EID CT (unpublished our own data). The phantom used for this imaging is Catphan 500 with CTP528 High Resolution Module (Phantom Laboratory Inc., Greenwich, USA). (Figure and legend-From Nakamura Y, Higaki T, Kondo S et al. An introduction to photon-counting detector CT (PCD CT) for radiologists. Jpn J Radiol (2022). https://doi.org/10.1007/s11604-­022-­01350-­6 creativecommons.org/licenses/BY/4.0/)

A

X-ray photon

common electrode

High Voltage hole

+ –

B

X-ray photon

reflector electron

light scintillator

CZT

photodiode (light charge)

pixel electrode E1

Amp. shaping (Charge voltage)

E2 E3 E4 comparator

processing circuit

E1