Explaining Uncertainty in Multiple Sclerosis Lesion Segmentation Beyond Prediction Errors
This work addresses the need for trustworthy AI in healthcare by explaining uncertainty sources in medical image segmentation, though it is incremental as it builds on existing uncertainty quantification methods.
This study tackled the problem of understanding the clinical informativeness of predictive uncertainty in cortical lesion segmentation for multiple sclerosis, revealing that uncertainty is strongly related to lesion size, shape, and cortical involvement, with evaluations on 206 patients and nearly 2000 lesions confirming utility under in-domain and distribution-shift conditions.
Trustworthy artificial intelligence (AI) is essential in healthcare, particularly for high-stakes tasks like medical image segmentation. Explainable AI and uncertainty quantification significantly enhance AI reliability by addressing key attributes such as robustness, usability, and explainability. Despite extensive technical advances in uncertainty quantification for medical imaging, understanding the clinical informativeness and interpretability of uncertainty remains limited. This study introduces a novel framework to explain the potential sources of predictive uncertainty, specifically in cortical lesion segmentation in multiple sclerosis using deep ensembles. The proposed analysis shifts the focus from the uncertainty-error relationship towards relevant medical and engineering factors. Our findings reveal that instance-wise uncertainty is strongly related to lesion size, shape, and cortical involvement. Expert rater feedback confirms that similar factors impede annotator confidence. Evaluations conducted on two datasets (206 patients, almost 2000 lesions) under both in-domain and distribution-shift conditions highlight the utility of the framework in different scenarios.