Integrating Pathology Foundation Models and Spatial Transcriptomics for Cellular Decomposition from Histology Images

2025-07-09

AI Summary

This study addresses the high cost and limited clinical adoption of spatial transcriptomics by proposing a novel method that accurately predicts cellular composition directly from routine H&E-stained histopathological images. Methodologically, it leverages features extracted from a pre-trained histopathology foundation model and trains a lightweight MLP regressor in an end-to-end fashion, using cell-type abundances inferred by cell2location as supervision. To our knowledge, this is the first approach to effectively integrate histopathology foundation models with spatial transcriptomic prior knowledge. Evaluated across multiple independent datasets, the method achieves prediction accuracy comparable to the state-of-the-art Hist2Cell, while reducing model parameters by over 90% and significantly lowering computational overhead. It demonstrates strong generalizability across diverse tissue types and scanners, and exhibits high potential for clinical deployment due to its efficiency, interpretability, and reliance solely on widely available H&E slides.

Technology Category

• Machine Learning: Other Foundations of Machine Learning
• Computer Vision: Other Foundations of Computer Vision
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

The rapid development of digital pathology and modern deep learning has facilitated the emergence of pathology foundation models that are expected to solve general pathology problems under various disease conditions in one unified model, with or without fine-tuning. In parallel, spatial transcriptomics has emerged as a transformative technology that enables the profiling of gene expression on hematoxylin and eosin (H&E) stained histology images. Spatial transcriptomics unlocks the unprecedented opportunity to dive into existing histology images at a more granular, cellular level. In this work, we propose a lightweight and training-efficient approach to predict cellular composition directly from H&E-stained histology images by leveraging information-enriched feature embeddings extracted from pre-trained pathology foundation models. By training a lightweight multi-layer perceptron (MLP) regressor on cell-type abundances derived via cell2location, our method efficiently distills knowledge from pathology foundation models and demonstrates the ability to accurately predict cell-type compositions from histology images, without physically performing the costly spatial transcriptomics. Our method demonstrates competitive performance compared to existing methods such as Hist2Cell, while significantly reducing computational complexity.

Problem

• Predict cellular composition from H&E histology images
• Leverage pathology foundation models for spatial transcriptomics
• Reduce computational cost of cell-type analysis

Innovation

• Leveraging pathology foundation models for feature embeddings
• Using lightweight MLP regressor for cell-type prediction
• Predicting cell composition without spatial transcriptomics