Can Protein-Derived Knowledge Improve Pathology Foundation Models?

2026-09-27

AI Summary

This study addresses the bottleneck wherein pathology foundation models struggle to leverage independent proteomic data due to the scarcity of paired samples. To overcome this, we propose ProSlide, a three-stage framework that enhances whole-slide image representations by decoupling knowledge acquisition from cross-modal transfer. Specifically, ProSlide introduces a novel virtual spectrum generation strategy to pre-train the protein encoder and employs Prot2Path relational distillation to circumvent strict pairing constraints, achieving cross-modal alignment through multi-view pre-training and hierarchical encoding. Evaluated across twelve cancer-related downstream tasks, ProSlide attains state-of-the-art average accuracy and AUC with minimal labeled data, demonstrating its significant advantages in computational pathology.

Abstract

Molecularly guided pathology foundation models (PFMs) exploit transcriptomic or proteomic information to enrich whole-slide image (WSI) representations, yet effectively leveraging large standalone molecular corpora remains challenging. First, existing molecular foundation models encode protein sequences or single-cell states, not the patient-level bulk expression profiles paired with WSIs. Second, because cross-modal supervision is restricted to paired WSI-omics samples, knowledge from standalone molecular corpora reaches the pathology encoder only indirectly, creating a paired-support bottleneck. To address these challenges, we propose a three-stage framework that decouples proteomic knowledge acquisition from cross-modal transfer, yielding ProSlide, a slide-level hierarchical pathology foundation model. First, to close the modality gap, we pretrain a Proteomic Foundation Encoder (PFE) on 12,695 sample-level bulk protein profiles using virtual profile generation and expression-space multi-view pretraining. Second, we pretrain ProSlide, a patch-region-slide encoder, to predict protein expression from paired WSI-protein samples. Third, to relax the paired-support bottleneck, we introduce Prot2Path, a cross-modal relational distillation objective. For each paired sample, it aligns the similarity distributions of the WSI and its protein profile over a shared, frozen bank of PFE-encoded paired and standalone profiles. We evaluate ProSlide on 12 downstream tasks across breast, lung, and renal cancers. Despite being pretrained with only 2,229 WSIs and 12,695 sample-level protein profiles, ProSlide achieves the highest mean accuracy and AUC within each cancer group.

Problem

• Pathology Foundation Models
• Whole-Slide Image
• Proteomics
• Cross-modal Learning
• Paired-support Bottleneck

Innovation

• Pathology Foundation Model
• Proteomic Foundation Encoder
• Cross-modal Relational Distillation
• Whole-slide Image
• Prot2Path