Investigating the Impact of Histopathological Foundation Models on Regressive Prediction of Homologous Recombination Deficiency

TL;DR

This study evaluates the effectiveness of histopathological foundation models in predicting homologous recombination deficiency scores from whole slide images, demonstrating improved accuracy and generalization over traditional methods.

Contribution

It systematically compares multiple foundation models for regression tasks in pathology, introduces a distribution-based upsampling strategy, and provides insights into model selection and data handling for biomarker prediction.

Findings

Foundation model features outperform contrastive learning features in HRD prediction.

The proposed upsampling strategy improves recall and balanced accuracy for rare cases.

Large-scale pretraining enhances the transferability and precision of biomarker prediction.

Abstract

Foundation models pretrained on large-scale histopathology data have found great success in various fields of computational pathology, but their impact on regressive biomarker prediction remains underexplored. In this work, we systematically evaluate histopathological foundation models for regression-based tasks, demonstrated through the prediction of homologous recombination deficiency (HRD) score - a critical biomarker for personalized cancer treatment. Within multiple instance learning frameworks, we extract patch-level features from whole slide images (WSI) using five state-of-the-art foundation models, and evaluate their impact compared to contrastive learning-based features. Models are trained to predict continuous HRD scores based on these extracted features across breast, endometrial, and lung cancer cohorts from two public medical data collections. Extensive experiments demonstrate that models trained on foundation model features consistently outperform the baseline in terms of predictive accuracy and generalization capabilities while exhibiting systematic differences among the foundation models. Additionally, we propose a distribution-based upsampling strategy to mitigate target imbalance in these datasets, significantly improving the recall and balanced accuracy for underrepresented but clinically important patient populations. Furthermore, we investigate the impact of different sampling strategies and instance bagsizes by ablation studies. Our results highlight the benefits of large-scale histopathological pretraining for more precise and transferable regressive biomarker prediction, showcasing its potential to advance AI-driven precision oncology.