Explainable AI in Healthcare: Physicians Perspectives and Technical Evaluation of AI-Based Decision Support and Explainability Methods

Sammanfattning

Artificial intelligence (AI) is increasingly integrated into healthcare, particularly inclinical decision support and predictive modeling. However, the limited interpretabilityof many machine learning models remains a major challenge for clinical implementation,motivating growing interest in explainable artificial intelligence (XAI).This thesis investigates XAI in healthcare from both technical and clinical perspectives.The clinical perspective was explored through qualitative interviews withphysicians focusing on AI-based decision support systems and the role of explainableAI in clinical practice. The findings revealed a cautiously optimistic view ofAI-supported decision-making, while emphasizing that clinically useful explanationsshould be concise, intuitive, and seamlessly integrated into existing workflows.The technical part of the study investigated XAI methods for survival predictionin lymphoma patients using both tabular clinical data and medical imaging data.Multiple survival modelling approaches, including Cox regression, DeepSurv, andconvolutional neural network models, were implemented and evaluated using severalpost-hoc explainability methods across the different data modalities. Whileboth modalities demonstrated strong predictive performance, the tabular modelsachieved slightly stronger results with more stable, interpretable explanations. Furthermore,different XAI approaches highlighted complementary but inconsistent patterns,illustrating challenges related to the robustness and reliability of post-hocexplanations.Overall, the findings demonstrated that successful clinical integration of AI dependsas much on providing reliable, clinically meaningful explanations as it does on achievingstrong predictive performance.