nnMIL: A generalizable multiple instance learning framework for computational pathology

2025-11-18

AI Summary

This work addresses the challenge of reliably aggregating patch-level features into slide-level clinical predictions for whole-slide images (WSIs). We propose a generalizable multi-instance learning (MIL) framework featuring a novel two-stage stochastic sampling mechanism that enables large-batch optimization and task-aware sampling. A lightweight aggregator supports sliding-window ensemble inference and uncertainty quantification, while patch representations are extracted from pathology foundation models. The framework synergistically integrates MIL, sliding-window inference, and ensemble strategies. Extensive experiments across 40,000 WSIs, 35 clinical tasks, and four foundation models demonstrate consistent and significant improvements over state-of-the-art methods in disease diagnosis, histological subtyping, molecular biomarker detection, and cross-cohort pan-cancer prognostic risk stratification. The approach exhibits strong robustness and clinically interpretable predictions.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Multiagent Systems: Other Foundations of Multi Agent Systems
• Computer Vision: Other Foundations of Computer Vision

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Computational pathology holds substantial promise for improving diagnosis and guiding treatment decisions. Recent pathology foundation models enable the extraction of rich patch-level representations from large-scale whole-slide images (WSIs), but current approaches for aggregating these features into slide-level predictions remain constrained by design limitations that hinder generalizability and reliability. Here, we developed nnMIL, a simple yet broadly applicable multiple-instance learning framework that connects patch-level foundation models to robust slide-level clinical inference. nnMIL introduces random sampling at both the patch and feature levels, enabling large-batch optimization, task-aware sampling strategies, and efficient and scalable training across datasets and model architectures. A lightweight aggregator performs sliding-window inference to generate ensemble slide-level predictions and supports principled uncertainty estimation. Across 40,000 WSIs encompassing 35 clinical tasks and four pathology foundation models, nnMIL consistently outperformed existing MIL methods for disease diagnosis, histologic subtyping, molecular biomarker detection, and pan- cancer prognosis prediction. It further demonstrated strong cross-model generalization, reliable uncertainty quantification, and robust survival stratification in multiple external cohorts. In conclusion, nnMIL offers a practical and generalizable solution for translating pathology foundation models into clinically meaningful predictions, advancing the development and deployment of reliable AI systems in real-world settings.

Problem

• Developing a generalizable framework for aggregating patch-level features into slide-level predictions
• Improving reliability and scalability of computational pathology models across clinical tasks
• Enabling robust clinical inference from whole-slide images using multiple instance learning

Innovation

• Random patch and feature sampling for optimization
• Lightweight aggregator with sliding-window ensemble predictions
• Generalizable framework connecting foundation models to clinical inference