This pioneering book explores how machine learning and other AI techniques impact millions of cancer patients who benefit from ionizing radiation. It features contributions from global researchers and clinicians, focusing on the clinical applications of machine learning for medical physics.
AI and machine learning have attracted much recent attention and are being increasingly adopted in medicine, with many clinical components and commercial software including aspects of machine learning integration. General principles and important techniques in machine learning are introduced, followed by discussion of clinical applications, particularly in radiomics, outcome prediction, registration and segmentation, treatment planning, quality assurance, image processing, and clinical decision-making. Finally, a futuristic look at the role of AI in radiation oncology is provided.
This book brings medical physicists and radiation oncologists up to date with the most novel applications of machine learning to medical physics. Practitioners will appreciate the insightful discussions and detailed descriptions in each chapter. Its emphasis on clinical applications reaches a wide audience within the medical physics profession.
Author(s): Gilmer Valdes, Lei Xing
Series: Imaging in Medical Diagnosis and Therapy
Publisher: CRC Press
Year: 2023
Language: English
Pages: 184
City: Boca Raton
Cover
Half Title
Series Page
Title Page
Copyright Page
Dedication
Table of Contents
Authors
Contributors
1. AI Applications in Radiation Therapy and Medical Physics
1.1 Why Artificial Intelligence for Radiotherapy?
1.1.1 Automation Applications of AI in Radiotherapy
1.2 AI in Radiotherapy
1.3 Sample Applications of AI in Radiotherapy
1.3.1 AI for Radiotherapy Automation
1.3.1.1 Autocontouring (Segmentation)
1.3.1.2 Knowledge-Based Planning
1.3.1.3 Quality Assurance and Error Detection
1.3.2 AI for Radiotherapy Predictive Analytics
1.3.2.1 Outcome Modelling
1.3.2.2 Knowledge-Based Adaptive Radiotherapy
1.4 Challenges for AI in Radiotherapy and Recommendations
1.4.1 Dataset Requirements
1.4.2 Validation of AI Models
1.4.3 Interpretability and Explainability
1.4.4 Quality Assurance
1.5 Conclusions
Acknowledgements
References
2. Machine Learning for Image-Based Radiotherapy Outcome Prediction
2.1 Introduction
2.2 Imagining Modalities
2.3 Quantitative Image Biomarkers
2.3.1 Image Standardization
2.3.2 Image Delineation
2.3.3 Biomarker Extraction
2.3.4 Biomarker Repeatability
2.4 RT Outcome Prediction Methodology
2.4.1 Biomarker Selection
2.4.2 Machine Learning Methodology for Image-Based RT Outcome Prediction
2.4.3 Predictor Explainability
2.4.4 Predictor Training Design
2.5 RT Outcomes of Interest Targeted by Machine Learning Predictors
2.5.1 Toxicities
2.5.2 Complete Pathological Response
2.5.3 Cancer Recurrence
2.5.4 Survival
2.6 Discussion
2.7 Conclusion
Acknowledgements
References
3. Metric Predictions for Machine and Patient-Specific Quality Assurance
3.1 Introduction
3.2 Machine Learning Applications to Quality Assurance
3.2.1 Automated Chart Review
3.2.2 Treatment Delivery Systems
3.2.3 Proton QA
3.2.4 Patient-Specific QA
3.3 Future Directions
Conflict of Interest
References
4. Data-Driven Treatment Planning, Plan QA, and Fast Dose Calculation
4.1 Radiation Therapy and Treatment Planning
4.2 Anatomy-Based Dose Distribution Prediction
4.3 Dose Distribution Prediction Based on Prior Geometric, Anatomic, and Dosimetric Properties of the Patients
4.4 Prediction of Machine Delivery Parameters or Fluence Maps for Treatment Planning
4.5 Adjusting Treatment Planning Parameters Directly
4.6 Data-Driven Treatment Plan QA
4.7 Data-Driven Dose Calculation
4.8 Summary
Acknowledgements
References
5. Reinforcement Learning for Radiation Therapy Planning and Image Processing
5.1 Introduction and Overview
5.2 Status
5.2.1 DRL Application in Treatment Planning for RT
5.2.2 DRL Applications in Medical Image Processing for RT
5.3 Current and Future Challenges
5.4 Future Perspective
References
6. Image Registration and Segmentation
6.1 Introduction and Background
6.1.1 Clinical Workflow in Radiation Oncology
6.1.2 Image Registration and Segmentation Applications
6.1.3 Overview of AI in Image Registration and Segmentation
6.2 Image Registration
6.2.1 Traditional Registration Methods
6.2.2 Evaluation of Registration Methods
6.2.3 AI Registration by Anatomical Region
6.2.3.1 Head and Neck
6.2.3.2 Thorax
6.2.3.3 Gastrointestinal
6.2.3.4 Pelvis
6.3 Image Segmentation
6.3.1 Traditional Segmentation Methods
6.3.2 Evaluating Segmentation Performance
6.3.3 AI Segmentation by Anatomical Region
6.3.3.1 Head & Neck
6.3.3.2 Thorax
6.3.3.3 Abdomen
6.3.3.4 Pelvis
6.4 Conclusion
References
7. Motion Management and Image-Guided Radiation Therapy
7.1 Introduction and Background
7.1.1 Motion Management and Image Guidance in Radiotherapy
7.1.2 Clinically Available Solutions
7.1.3 The Role of AI
7.2 Current AI & ML Methods
7.2.1 Inter- and Intra-fractional Imaging
7.2.2 Adaptive Radiotherapy
7.2.3 Real-time Target Position Monitoring
7.2.3.1 Spatial Prediction of Target Positions
7.2.3.2 Temporal Prediction of Target Positions
7.3 Future Directions and Opportunities
Acknowledgements
References
8. Outlook of AI in Medical Physics and Radiation Oncology
8.1 Introduction
8.2 Clinical Assessment
8.3 Simulation and Imaging
8.4 Treatment Planning
8.5 Treatment Delivery
8.6 Integrating AI Into Radiation Oncology Practice
References
Index
