Deep Learning for Detecting BRCA Mutations in High-Grade Ovarian Cancer based on an Innovative Tumor Segmentation Method from Whole-Slide Images
Résumé
BRCA1/2 genes play a crucial role in repairing DNA double-strand breaks through homologous recombination. Their mutations represent a significant proportion of homologous recombination deficiency and are a reliable effective predictor of sensitivity of high-grade ovarian cancer (HGOC) to poly(ADP-ribose) polymerase inhibitors. However, their testing by next-generation sequencing is costly, time-consuming, and can be affected by various preanalytical factors. In this study, we present a deep learning classifier for BRCA mutational status prediction from HES-stained whole-slide images (WSI) of HGOC. We constituted the OvarIA cohort composed of 867 HGOC patients with known BRCA somatic mutational status coming from two different pathology departments. We first developed a tumor segmentation model according to dynamic sampling and then trained a visual representation encoder with momentum contrastive learning on the predicted tumor tiles. We finally trained a BRCA classifier on over a million tumor tiles in multiple-instance learning with an attention-based mechanism. The tumor segmentation model trained on 8 WSI obtained a dice score of 0.915 and an intersection-over-union of 0.847 on a test set of 50 WSI while the BRCA classifier achieved the state-of-the-art AUC of 0.739 in 5-fold cross-validation and 0.681 on the testing set. An additional multiscale approach indicates that the relevant information for predicting BRCA mutations is more located in the tumor context than in the cell morphology. Our results suggest that BRCA somatic mutations have a discernible phenotypic effect which could be detected by deep learning and could be used as a pre-screening tool in the future.
Fichier principal
Bourgade et al 2023-Deep Learning for Detecting BRCA Mutations in High-Grade Ovarian Cancer.pdf (2.37 Mo)
Télécharger le fichier
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|---|
| Licence |
