Deep learning on histological slides accurately predicts Consensus Molecular Subtypes and spatial heterogeneity in colon cancer.

Fiche publication

Date publication

août 2025

Journal

Modern pathology : an official journal of the United States and Canadian Academy of Pathology, Inc

Auteurs

Membres identifiés du Cancéropôle Est :
Pr BOUCHE Olivier , Pr LEPAGE Côme , Pr BIBEAU Frédéric

Tous les auteurs :
LE Douget JE, Jacob P, Lepage C, Gallois C, Sroussi M, Reynies ADE, Cazelles A, Gonzalez D, Maussion C, Bouche O, Levache CB, Jary M, Mineur L, Jegou S, Saillard C, Morel M, Taïeb J, Bibeau F, Emile JF, Laurent-Puig P

Lien Pubmed

https://www.ncbi.nlm.nih.gov/pubmed/40865918

Résumé

Colon cancer (CC) is the third most prevalent cancer type. It is highly heterogeneous, particularly in terms of molecular profiles, which have both prognostic and predictive impacts on the treatment efficacy. However, CC treatment in adjuvant situations is currently guided solely by T and N staging. In this context, Consensus Molecular Subtypes (CMS) was introduced to stratify CC patients based on molecular profiles. Recent studies have shown that CMS can be heterogeneous in CC, leading to a worse prognosis. This study focuses on predicting CMS and its heterogeneity in CC using deep learning on digitized haematoxylin-eosin ± saffron-stained Whole Slide Images (WSIs). Data and WSI of 1,996 patients from the PETACC-8, TCGA-COAD, and PRODIGE-13 cohorts were used. The model is trained to predict a 4-dimensional CMS vector, reflecting intra-tumor heterogeneity (ITH). It comprises a self-supervised model for embedding image patches into vectors and a weakly-supervised model predicting CMS calls. Ground-truth CMS scores are obtained with the CMSclassifier package. Interpretability analyses are performed at the slide and patch levels. For homogeneous tumors, the model trained on PETACC-8 achieves 93.0% (±1.4%) macro-average AUC in internal cross-validation (CV) and 94.4% macro-average AUC in external validation over PRODIGE-13, while the TCGA-COAD model reaches 85.4% (±3.0%) in CV and 92.4% over PRODIGE-13. The trained models also provide spatial distributions of CMS across tumor slides and associate specific histological features to each CMS. Finally, the models are able to predict ITH. The results show that a deep learning model trained on routine histology slides is capable of providing an efficient and robust method for predicting CMS and characterizing a patient's ITH, paving the way for the routine consideration of CMS/ITH in clinical decision-making in the adjuvant setting.