aceA survey on computational pathology foundation models: datasets, adaptation strategies, and evaluation tasks
Abstract Computational pathology foundation models (CPathFMs) have emerged as a powerful approach for analyzing histopathological data, leveraging self-supervised learning to extract robust feature representations from unlabeled whole-slide images. These models, categorized into uni-modal and multi-modal frameworks, have demonstrated promise in automating complex pathology tasks such as segmentation, classification, and biomarker discovery. However, the development of CPathFMs presents significant challenges, su…
"Lung - Miliary tuberculosis" by Pulmonary Pathology is licensed under CC BY-SA 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/2.0/.
As of the July 2 2026 publication date, this peer-reviewed survey summarized the state of computational pathology foundation models that use self-supervised learning on unlabeled whole-slide images to support pathology tasks. It reported that these uni-modal and multi-modal models have shown promise for segmentation, classification, and biomarker discovery, while focusing its review on datasets, adaptation strategies, and evaluation tasks.
The clinical relevance lies in moving AI pathology tools toward robust and clinically applicable use, but the source itself highlights unresolved barriers that limit that translation. It explicitly lists limited data accessibility, dataset variability, need for domain-specific adaptation, and lack of standardized benchmarks, indicating that standardization and accessibility remain open issues for future work.
- CPathFMs categorized into uni-modal and multi-modal frameworks for histopathological data
- Pre-training relies on self-supervised learning to extract robust feature representations from unlabeled whole-slide images
- Key techniques analyzed include contrastive learning, masked image modeling and multi-modal integration
- Survey focuses on pre-training datasets, adaptation strategies, and evaluation tasks
Development of computational pathology foundation models is constrained by limited data accessibility, high variability across datasets, need for domain-specific adaptation, and lack of standardized evaluation benchmarks
The rundown
The survey defines CPathFMs as leveraging self-supervised learning to extract robust feature representations from unlabeled whole-slide images, and organizes them into uni-modal and multi-modal frameworks.
It reviews pre-training datasets, adaptation strategies, and evaluation tasks, analyzing techniques such as contrastive learning, masked image modeling and multi-modal integration, and notes gaps and future directions from four perspectives toward clinically applicable solutions.
Sources
- Peer-reviewedKnowledge and Information Systems2026-07-02
How should this claim be treated?
The debate