Pathology Foundation Models are Scanner Sensitive: Benchmark and Mitigation with Contrastive ScanGen Loss

TL;DR

This paper demonstrates that pathology foundation models are sensitive to scanner bias in WSIs and introduces ScanGen, a contrastive loss during fine-tuning, to mitigate this bias and improve model robustness across different scanners.

Contribution

The paper benchmarks foundation models on multi-scanner datasets and proposes ScanGen, a contrastive loss method, to reduce scanner bias in pathology models.

Findings

ScanGen improves scanner robustness in pathology models.

Foundation models still suffer from scanner bias without mitigation.

ScanGen maintains or improves mutation prediction accuracy.

Abstract

Computational pathology (CPath) has shown great potential in mining actionable insights from Whole Slide Images (WSIs). Deep Learning (DL) has been at the center of modern CPath, and while it delivers unprecedented performance, it is also known that DL may be affected by irrelevant details, such as those introduced during scanning by different commercially available scanners. This may lead to scanner bias, where the model outputs for the same tissue acquired by different scanners may vary. In turn, it hinders the trust of clinicians in CPath-based tools and their deployment in real-world clinical practices. Recent pathology Foundation Models (FMs) promise to provide better domain generalization capabilities. In this paper, we benchmark FMs using a multi-scanner dataset and show that FMs still suffer from scanner bias. Following this observation, we propose ScanGen, a contrastive loss function applied during task-specific fine-tuning that mitigates scanner bias, thereby enhancing the models' robustness to scanner variations. Our approach is applied to the Multiple Instance Learning task of Epidermal Growth Factor Receptor (EGFR) mutation prediction from H\&E-stained WSIs in lung cancer. We observe that ScanGen notably enhances the ability to generalize across scanners, while retaining or improving the performance of EGFR mutation prediction.