To What Extent Do Token-Level Representations from Pathology Foundation Models Improve Dense Prediction?

TL;DR

This paper introduces PFM-DenseBench, a comprehensive benchmark evaluating 17 pathology foundation models across 18 datasets to understand their performance and stability in dense tissue segmentation tasks.

Contribution

It provides a large-scale, systematic evaluation framework for PFMs in dense pathology prediction, including reproducible protocols and practical insights.

Findings

PFMs vary significantly in performance across datasets

Fine-tuning strategies impact model stability and accuracy

Reproducible evaluation tools are provided for real-world application

Abstract

Pathology foundation models (PFMs) have rapidly advanced and are becoming a common backbone for downstream clinical tasks, offering strong transferability across tissues and institutions. However, for dense prediction (e.g., segmentation), practical deployment still lacks a clear, reproducible understanding of how different PFMs behave across datasets and how adaptation choices affect performance and stability. We present PFM-DenseBench, a large-scale benchmark for dense pathology prediction, evaluating 17 PFMs across 18 public segmentation datasets. Under a unified protocol, we systematically assess PFMs with multiple adaptation and fine-tuning strategies, and derive insightful, practice-oriented findings on when and why different PFMs and tuning choices succeed or fail across heterogeneous datasets. We release containers, configs, and dataset cards to enable reproducible evaluation and informed PFM selection for real-world dense pathology tasks. Project Website: https://m4a1tastegood.github.io/PFM-DenseBench