Explainable AI for computational pathology identifies model limitations and tissue biomarkers
Jakub Roman Kaczmarzyk, Kim, Chanwoo, Gadgil, Soham, Savant, Deepika, Zhao, Zhen, Saltz, Joel H., Lee, Su-In, Koo, Peter K. · PubMed · 2024
Introduction: Deep learning models hold great promise for digital pathology, but their opaque decision-making processes undermine trust and hinder clinical adoption. Explainable AI methods are essential to enhance model transparency and reliability. Methods: mutation classification in gliomas. In computational experiments, HIPPO was compared against traditional metrics and attention-based approaches to assess its ability to identify key tissue elements driving model predictions. Results: mutation classification, HIPPO more robustly identified the pathology regions responsible for false negatives compared to attention, suggesting its potential to outperform attention in explaining model decisions. Conclusions: HIPPO expands the explainable AI toolkit for computational pathology by enabling deeper insights into model behavior. This framework supports the trustworthy development, deployment, and regulation of weakly-supervised models in clinical and research settings, promoting their broader adoption in digital pathology.