xMIL: Insightful Explanations for Multiple Instance Learning in Histopathology

2024-06-06

arXiv.org

AI Summary

To address the poor interpretability of Multiple Instance Learning (MIL) models in digital pathology—hindering pathologists’ understanding of biomarker prediction rationales—this paper proposes xMIL, the first framework to systematically integrate Layer-wise Relevance Propagation (LRP) into MIL. Unlike conventional approaches relying on the small-bag assumption and instance independence, xMIL explicitly models inter-instance interactions, thereby enhancing biological plausibility and clinical interpretability. Evaluated on four real-world histopathological datasets, xMIL achieves significantly higher explanation faithfulness than state-of-the-art methods. Its open-source implementation demonstrates practical utility in clinical knowledge discovery and model debugging. The core contribution is the establishment of an MIL-specific LRP interpretability paradigm, uniquely reconciling weak supervision constraints with domain-specific histopathological semantics.

Technology Category

• Machine Learning: Transparent, Interpretable, Explainable ML
• Humans and AI: Explainable AI (XAI) for Human Understanding
• Computer Vision: Interpretability, Explainability, and Transparency

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Multiple instance learning (MIL) is an effective and widely used approach for weakly supervised machine learning. In histopathology, MIL models have achieved remarkable success in tasks like tumor detection, biomarker prediction, and outcome prognostication. However, MIL explanation methods are still lagging behind, as they are limited to small bag sizes or disregard instance interactions. We revisit MIL through the lens of explainable AI (XAI) and introduce xMIL, a refined framework with more general assumptions. We demonstrate how to obtain improved MIL explanations using layer-wise relevance propagation (LRP) and conduct extensive evaluation experiments on three toy settings and four real-world histopathology datasets. Our approach consistently outperforms previous explanation attempts with particularly improved faithfulness scores on challenging biomarker prediction tasks. Finally, we showcase how xMIL explanations enable pathologists to extract insights from MIL models, representing a significant advance for knowledge discovery and model debugging in digital histopathology. Codes are available at: https://github.com/bifold-pathomics/xMIL.

Problem

• Multi-instance Learning
• Pathology
• Interpretability

Innovation

• xMIL
• Interpretable AI
• Multi-Instance Learning in Pathology