Complete Table of Contents

Complete Table of Contents (Original pp. 6–11)

Page numbers in parentheses refer to the original document pages.

• Approval Page of the Examination Committee – APPROVAL PAGE (p. 2)
• Dedication (p. 3)
• Acknowledgements (p. 4)
• Our Published Papers Related to Research – OUR PUBLISHED PAPERS RELATED TO RESEARCH (p. 5)
• Table of Contents (p. 6)
• List of Tables (p. 12)
• List of Figures and Diagrams (p. 15)
• Abstract – ABSTRACT (p. 21)

Chapter 1: Introduction (p. 22)

Chapter 2: Keratoconus (p. 24)

• 2.1 Introduction to Corneal Anatomy, Embryology, and Physiology (p. 25)
• 2.1.1 Embryological Overview (p. 25)
• 2.1.2 Anatomical and Physiological Overview (p. 26)
• 2.2 The Optical Function of the Cornea (p. 31)
• 2.3 Corneal Biomechanics (p. 32)
• 2.4 Ectatic Corneal Diseases and Keratoconus (p. 34)
• 2.4.1 Keratoconus (p. 34)
• 2.4.2 Pellucid Marginal Degeneration (p. 34)
• 2.4.3 Keratoglobus (p. 35)
• 2.4.4 Terrien Marginal Dystrophy (p. 36)
• 2.4.5 Post-Refractive Surgery Ectasia (p. 37)
• 2.5 Keratoconus (p. 39)
• 2.5.1 Definition (p. 39)
• 2.5.2 Clinical Course (p. 39)
• 2.5.3 Clinical Signs (Mazen M. Sinjab, 2011b) (p. 39)
• 2.5.4 Associated Conditions in Keratoconus (Coster, 2002b; Kanski & Bowling, 2015) (p. 42)
• 2.5.5 Inheritance (p. 42)
• 2.5.6 Classification (p. 42)
• 2.5.7 Forme Fruste Keratoconus and Subclinical Keratoconus, and Keratoconus-Suspect Cornea (Suspect KC) (p. 43)
• 2.5.8 Pellucid-Like Keratoconus (Pellucid-Like KC) (p. 43)
• 2.6 Topographers and Tomographers (p. 44)
• 2.6.1 Keratometer and Photokeratometer (p. 44)
• 2.6.2 Computer-Assisted Videokeratography (p. 45)
• 2.6.3 Orbscan Topographer (p. 45)
• 2.6.4 Pentacam (p. 46)
• 2.6.5 Sirius (p. 47)
• 2.6.6 MS-39 (as an Example of AS-OCT) (p. 48)
• 2.7 Basic Indices and Maps of the SIRIUS Device (p. 50)
• 2.7.1 Basic Maps (p. 50)
• 2.7.2 Basic Indices (p. 57)
• 2.8 Topographic Features of Keratoconus (p. 60)
• 2.8.1 Curvature Maps (p. 60)
• 2.8.2 Elevation Maps (p. 61)

Chapter 3: Artificial Intelligence (p. 63)

• 3.1 Introduction to Artificial Intelligence (p. 64)
• 3.1.1 Introduction (p. 64)
• 3.1.2 Definition (Russell & Norvig, 2002a) (p. 64)
• 3.1.3 Historical Overview (Mijwel, 2015; Russell & Norvig, 2002a) (p. 65)
• 3.2 Overview of Techniques Used in Artificial Intelligence (p. 67)
• 3.2.1 Search Algorithms (p. 67)
• 3.2.2 Expert Systems (p. 67)
• 3.2.3 Machine Learning (p. 68)
• 3.2.4 Evolutionary and Genetic Algorithms (p. 69)
• 3.3 Artificial Neural Networks (p. 70)
• 3.3.1 Definition (p. 70)
• 3.3.2 Training Process (p. 71)
• 3.3.3 Advantages and Disadvantages of Artificial Neural Networks (p. 73)
• 3.4 Deep Learning and Computer Vision (p. 74)
• 3.4.1 Deep Learning (p. 74)
• 3.4.2 Computer Vision and Convolutional Neural Networks (p. 75)
• 3.5 Challenges and Solutions in the Training Process (p. 78)
• 3.5.1 Overfitting (p. 78)
• 3.5.2 Small Dataset (p. 79)
• 3.5.3 Class Imbalance (p. 80)

Chapter 4: Applications of Artificial Intelligence in Medicine (p. 82)

• 4.1 Introduction (p. 83)
• 4.2 Examples of AI Applications in Clinical Practice (p. 84)
• 4.2.1 As a Screening Tool (p. 84)
• 4.2.2 As a Prognostic Tool (p. 84)
• 4.2.3 As Treatment Support (p. 85)
• 4.2.4 As a Substitute for a Healthcare Provider (p. 85)
• 4.2.5 As an Aid to a Healthcare Provider (p. 85)
• 4.3 Examples of AI Applications in Ophthalmology (p. 86)
• 4.3.1 Diabetic Retinopathy (p. 87)
• 4.3.2 Glaucoma (p. 88)
• 4.3.3 Age-Related Macular Degeneration (p. 89)
• 4.3.4 Cataract (p. 89)
• 4.3.5 Various Other Applications (p. 89)
• 4.4 Examples of AI Applications for Keratoconus Detection (p. 91)
• 4.4.1 Keratoconus Diagnosis Using Corneal Biomechanics and Regression Algorithms (p. 91)
• 4.4.2 Keratoconus Diagnosis Using a Support Vector Machine (p. 93)
• 4.4.3 Keratoconus Diagnosis Using Convolutional Neural Networks (p. 93)

Chapter 5: Study Design and Methods (p. 97)

• 5.1 Study Design (p. 98)
• 5.2 Study Population (p. 98)
• 5.3 Sample Size (p. 98)
• 5.4 Inclusion and Exclusion Criteria and Technical Image Properties (p. 98)
• 5.4.1 Inclusion Criteria (p. 98)
• 5.4.2 Exclusion Criteria (p. 99)
• 5.4.3 Technical Image Properties (p. 99)
• 5.5 Characteristics of the Study Groups (p. 100)
• 5.5.1 Characteristics of the Training Group (p. 100)
• 5.5.2 Characteristics of the Test Group (p. 105)
• 5.6 Study Methods (p. 110)
• 5.6.1 First Part (p. 110)
• 5.6.2 Second Part (p. 110)
• 5.6.3 Training Process (p. 111)
• 5.6.4 Primary Study Outcomes (p. 112)
• 5.6.5 Statistical Data Analysis (p. 113)

Chapter 6: Results (p. 114)

• 6.1 Neural Network Results for the Anterior Sagittal Curvature Map (p. 115)
• 6.1.1 Training Group (p. 115)
• 6.1.2 Validation Group (p. 116)
• 6.1.3 Test Group (p. 117)
• 6.2 Neural Network Results for the Posterior Sagittal Curvature Map (p. 118)
• 6.2.1 Training Group (p. 118)
• 6.2.2 Validation Group (p. 119)
• 6.2.3 Test Group (p. 119)
• 6.3 Neural Network Results for the Anterior Tangential Curvature Map (p. 121)
• 6.3.1 Training Group (p. 121)
• 6.3.2 Validation Group (p. 122)
• 6.3.3 Test Group (p. 122)
• 6.4 Neural Network Results for the Posterior Tangential Curvature Map (p. 124)
• 6.4.1 Training Group (p. 124)
• 6.4.2 Validation Group (p. 125)
• 6.4.3 Test Group (p. 125)
• 6.5 Neural Network Results for the Pachymetry Map (p. 127)
• 6.5.1 Training Group (p. 127)
• 6.5.2 Validation Group (p. 128)
• 6.5.3 Test Group (p. 128)
• 6.6 Neural Network Results for the Anterior Elevation Map (p. 130)
• 6.6.1 Training Group (p. 130)
• 6.6.2 Validation Group (p. 131)
• 6.6.3 Test Group (p. 131)
• 6.7 Neural Network Results for the Posterior Elevation Map (p. 133)
• 6.7.1 Training Group (p. 133)
• 6.7.2 Validation Group (p. 134)
• 6.7.3 Test Group (p. 134)
• 6.8 Neural Network Results for the Anterior Refractive Power Map (p. 136)
• 6.8.1 Training Group (p. 136)
• 6.8.2 Validation Group (p. 137)
• 6.8.3 Test Group (p. 137)
• 6.9 Neural Network Results for the Posterior Refractive Power Map (p. 139)
• 6.9.1 Training Group (p. 139)
• 6.9.2 Validation Group (p. 140)
• 6.9.3 Test Group (p. 140)
• 6.10 Neural Network Results for the Equivalent Refractive Power Map (p. 142)
• 6.10.1 Training Group (p. 142)
• 6.10.2 Validation Group (p. 143)
• 6.10.3 Test Group (p. 143)
• 6.11 Overall Accuracy Results of the AI System (p. 145)
• 6.11.1 Training Group (p. 145)
• 6.11.2 Validation Group (p. 146)
• 6.11.3 Test Group (p. 147)
• 6.12 Results of the Software Accompanying the SIRIUS Device (p. 148)
• 6.13 Results of Physician Assessment without AI Support (p. 149)
• 6.14 Results of Physician Assessment with AI Support (p. 150)
• 6.15 Results of Physician Assessment with SIRIUS Device Support (p. 151)
• 6.16 Application of the McNemar Test to Compare the Above Models (p. 152)
• 6.16.1 Comparison between AI System and SIRIUS Device (p. 153)
• 6.16.2 Comparison between AI System and Physician without Support (p. 153)
• 6.16.3 Comparison between AI System and Physician with AI Support (p. 154)
• 6.16.4 Comparison between AI System and Physician with SIRIUS Support (p. 154)
• 6.16.5 Comparison between SIRIUS Result and Physician without Support (p. 155)
• 6.16.6 Comparison between SIRIUS Result and Physician with AI Support (p. 155)
• 6.16.7 Comparison between SIRIUS Result and Physician with SIRIUS Support (p. 156)
• 6.16.8 Comparison between Physician without Support and Physician with AI Support (p. 156)
• 6.16.9 Comparison between Physician without Support and Physician with SIRIUS Support (p. 157)
• 6.16.9 (numbered 6.16.9 again in the original) Comparison between Physician with AI Support and Physician with SIRIUS Support (p. 157)

Chapter 7: Discussion (p. 158)

• 7.1 Demographic Information (p. 159)
• 7.1.1 Prevalence (p. 159)
• 7.1.2 Age Distribution by Diagnosis (p. 159)
• 7.1.3 Gender Distribution by Diagnosis (p. 159)
• 7.2 Topographic Maps (p. 160)
• 7.2.1 Discussion of Neural Network Accuracy Indices in Their Ability to Distinguish Definite Keratoconic Corneas from Normal and Suspect Corneas (p. 160)
• 7.2.2 Discussion of Neural Network Accuracy Indices in Their Ability to Distinguish Normal Corneas from Keratoconic and Suspect Corneas (p. 162)
• 7.2.3 Discussion of Neural Network Accuracy Indices in Their Ability to Distinguish Suspect Corneas from Keratoconic and Normal Corneas (p. 163)
• 7.3 The AI System and Comparison with Similar Studies (p. 165)
• 7.4 Comparison with Other Models (p. 166)
• 7.5 Discriminative Features and Heatmaps (p. 167)
• 7.5.1 Anterior Sagittal Curvature Map (p. 167)
• 7.5.2 Posterior Sagittal Curvature Map (p. 167)
• 7.5.3 Anterior Tangential Curvature Map (p. 168)
• 7.5.4 Posterior Tangential Curvature Map (p. 168)
• 7.5.5 Anterior Elevation Map (p. 169)
• 7.5.6 Posterior Elevation Map (p. 170)
• 7.5.7 Pachymetry Map (p. 170)
• 7.5.8 Equivalent Refractive Power Map (p. 171)
• 7.5.9 Anterior Refractive Power Map (p. 171)
• 7.5.10 Posterior Refractive Power Map (p. 172)
• 7.6 Examination of Selected Cases (p. 173)
• 7.6.1 Case 1 (p. 173)
• 7.6.2 Case 2 (p. 173)
• 7.6.3 Case 3 (p. 173)
• 7.6.4 Case 4 (p. 177)
• 7.6.5 Case 5 (p. 177)
• 7.6.6 Case 6 (p. 177)
• 7.6.7 Case 7 (p. 181)
• 7.6.8 Case 8 (p. 181)
• 7.6.9 Case 9 (p. 181)
• 7.7 Strengths and Limitations of Our Study (p. 185)

Chapter 8: Summary and Conclusions (p. 186)

Chapter 9: Recommendations (p. 187)

Chapter 10: Concluding Remarks (p. 188)

References (p. 190)

Abstract (Arabic) (p. 195)

All 58 tablesAll 114 figures