Assessing Foundation Models for Computational Pathology in Endometrial Cancer
Year: 2026
Type: article
Abstract: Computational pathology leverages deep learning to extract clinically relevant information from digitized tumor slides, predicting histopathological subtypes, molecular alterations, and patient outcomes. Recent pipelines increasingly rely on foundation models trained on large pan-cancer data sets to generate generalizable features. In endometrial cancer (EC), their comparative performance for clinical diagnostic tasks remains unexplored. This study evaluates the performance of 7 state-of-the-art foundation models across morphological, molecular, and prognostic tasks using a large EC data set of 3293 patients from randomized trials and clinical cohorts. In addition, their performance was compared with 2 versions of an EC-specific feature extractor (EsVIT) exclusively trained on EC. The foundation models H-OPTIMUS-0, CONCH, and VIRCHOW2 achieved the highest mean performance, but the best-performing foundation model varied by task. The top-performing foundation model outperformed EsVIT across all tasks, which highlights the superiority of foundation models over the domain-specific feature extractor EsVIT in EC. Selecting the optimal foundation model for novel tasks remains challenging due to performance plateaus and limited information on the training data sets, requiring rigorous benchmarking and domain insights to reach maximum potential.
Source: Modern Pathology
Authors: Nikki J. van den Berg, Sarah Volinsky-Fremond, Jurriaan Barkey Wolf, Lydia A. Schoenpflug, Sonali Andani, Gitte Ørtoft, Jan J. Jobsen, Ludy C.H.W. Lutgens, Melanie E. Powell, Linda R. Mileshkin, Helen Mackay, Alexandra Leary, Rubina R. Razack, Marco de Bruyn, Stephanie M. de Boer, Remi A. Nout, Vincent T.H.B.M. Smit, Carien L. Creutzberg, Viktor H. Koelzer, Tjalling Bosse, Nanda Horeweg
Institutions: Leiden University Medical Center, University of Basel, University Hospital of Basel, SIB Swiss Institute of Bioinformatics, ETH Zurich, Copenhagen University Hospital, Rigshospitalet, Medisch Spectrum Twente, National Health Service, Barts Health NHS Trust, Peter MacCallum Cancer Centre, Sunnybrook Health Science Centre, Sunnybrook Hospital, Odette Cancer Centre, Institut Gustave Roussy, Tygerberg Hospital, University Medical Center Groningen, University of Groningen, Erasmus MC, Erasmus MC Cancer Institute
Language: English
Topic: AI in cancer detection
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $4,480
Funder: Hanarth Fonds
OpenAlex ID: W7203733508