Explainable AI for Transparent and Trustworthy Medical Decision Support equips readers with a comprehensive and timely resource that presents the principles, methodologies, and real-world applications of explainable AI (XAI) within the medical context. Covering a wide range of use cases—from radiology and pathology to genomics and clinical decision support systems—the book provides in-depth discussions on how XAI techniques can enhance interpretability, improve clinician trust, meet regulatory requirements, and ultimately lead to better patient outcomes. The book demystifies the workings of machine learning models and highlights techniques that make them interpretable.
It is designed to empower not only AI researchers and developers but also healthcare administrators and policymakers with the knowledge needed to evaluate, adopt, and trust AI solutions in critical medical applications. The book's authors bring together theory, implementation strategies, ethical implications, and case studies under one cover, offering a multidisciplinary perspective that aligns computer science with medical practice and healthcare policy.
1. 1. Introduction to explainable artificial intelligence (XAI) - p. 3 2. 2. The need for transparency in medical AI systems - p. 23 3. 3. Ethical and legal dimensions of AI in healthcare - p. 41 4. Section II. Techniques and tools for medical XAI - p. S2 5. 4. Trust, accountability, and human-in-the-loop decision making - p. 61 6. 5. Attention mechanisms and feature importance in deep learning - p. 81 7. 6. Explainable AI in radiology and medical imaging - p. 99 8. Section III. Applications in clinical domains - p. S3 9. 7. XAI for predictive modelling in electronic health records (EHRs) - p. 117 10. 8. Transparent artificial intelligence for disease diagnosis and prognosis - p. 137 11. 9. The future of human–AI collaboration in medical practice - p. 157 12. 10. Interpretable versus explainable models: A practical overview - p. 173 13. 11. A unified evaluation framework for model-agnostic XAI: Comparative analysis of LIME, SHAP, anchors, and counter factual - p. 189 14. Section IV. Ethics, evaluation, and future directions - p. S4 15. 12. Trustworthy artificial intelligence in COVID-19 diagnosis and cancer detection: Case studies in clinical implementation - p. 207 16. 13. Evaluating explanation effectiveness in healthcare: A comprehensive framework for interpretable AI systems with application to machine learning-based Parkinson's disease detection - p. 223 17. 14. Personalized explanations and adaptive decision support systems for Parkinson's disease: Tailoring AI insights to clinical user characteristics - p. 247 18. 15. An improved dual-input XAI model to reduce false negatives in predicting SARS_Pneumonia_COVID-19 disease using CT images - p. 271 19. 16. Regulatory frameworks and standards for explainable artificial intelligence in medicine: A comprehensive analysis of governance approaches - p. 289 20. 17. Explainable AI in medical imaging: A comparative study of model-agnostic methods and gradient-based class activation mapping techniques for classification and segmentation - p. 311 21. Index - p. 331
Explainable AI for Transparent and Trustworthy Medical Decision Support equips readers with a comprehensive and timely resource that presents the principles, methodologies, and real-world applications of explainable AI (XAI) within the medical context. Covering a wide range of use cases—from radiology and pathology to genomics and clinical decision support systems—the book provides in-depth discussions on how XAI techniques can enhance interpretability, improve clinician trust, meet regulatory requirements, and ultimately lead to better patient outcomes. The book demystifies the workings of machine learning models and highlights techniques that make them interpretable.
It is designed to empower not only AI researchers and developers but also healthcare administrators and policymakers with the knowledge needed to evaluate, adopt, and trust AI solutions in critical medical applications. The book's authors bring together theory, implementation strategies, ethical implications, and case studies under one cover, offering a multidisciplinary perspective that aligns computer science with medical practice and healthcare policy.
1. 1. Introduction to explainable artificial intelligence (XAI) - p. 3 2. 2. The need for transparency in medical AI systems - p. 23 3. 3. Ethical and legal dimensions of AI in healthcare - p. 41 4. Section II. Techniques and tools for medical XAI - p. S2 5. 4. Trust, accountability, and human-in-the-loop decision making - p. 61 6. 5. Attention mechanisms and feature importance in deep learning - p. 81 7. 6. Explainable AI in radiology and medical imaging - p. 99 8. Section III. Applications in clinical domains - p. S3 9. 7. XAI for predictive modelling in electronic health records (EHRs) - p. 117 10. 8. Transparent artificial intelligence for disease diagnosis and prognosis - p. 137 11. 9. The future of human–AI collaboration in medical practice - p. 157 12. 10. Interpretable versus explainable models: A practical overview - p. 173 13. 11. A unified evaluation framework for model-agnostic XAI: Comparative analysis of LIME, SHAP, anchors, and counter factual - p. 189 14. Section IV. Ethics, evaluation, and future directions - p. S4 15. 12. Trustworthy artificial intelligence in COVID-19 diagnosis and cancer detection: Case studies in clinical implementation - p. 207 16. 13. Evaluating explanation effectiveness in healthcare: A comprehensive framework for interpretable AI systems with application to machine learning-based Parkinson's disease detection - p. 223 17. 14. Personalized explanations and adaptive decision support systems for Parkinson's disease: Tailoring AI insights to clinical user characteristics - p. 247 18. 15. An improved dual-input XAI model to reduce false negatives in predicting SARS_Pneumonia_COVID-19 disease using CT images - p. 271 19. 16. Regulatory frameworks and standards for explainable artificial intelligence in medicine: A comprehensive analysis of governance approaches - p. 289 20. 17. Explainable AI in medical imaging: A comparative study of model-agnostic methods and gradient-based class activation mapping techniques for classification and segmentation - p. 311 21. Index - p. 331
