Computational Pathology for Prediction of Isocitrate Dehydrogenase Gene Mutation from Whole-Slide Images in Adult Patients with Diffuse Glioma.

Researchers

Journal

Modalities

Models

Abstract

Isocitrate dehydrogenase gene (IDH) mutation is one of the most important molecular markers of glioma. Accurate detection of IDH status is a crucial step for integrated diagnosis of adult-type diffuse gliomas. A clustering-based hybrid of a convolutional neural network (CNN) and a vision transformer (ViT) deep learning model was developed to detect IDH mutation status from annotation-free hematoxylin & eosin-stained whole-slide pathological images (WSIs) of 2275 adult patients with diffuse gliomas. For comparison, we also assessed a pure CNN, a pure ViT, and a classical multiple-instance learning model. The hybrid model achieved an AUC of 0.973 in the validation set and 0.953 in the external test set, outperforming the other models. We further assessed the hybrid model’s ability in IDH detection between difficult subgroups with different IDH status but shared histological features, achieving AUCs ranging from 0.850 to 0.985 in validation and test sets. Our data suggests that the proposed hybrid model has a potential to be used as a computational pathology tool for preliminary rapid detection of IDH mutation from WSI in adult patients with diffuse gliomas.Copyright © 2024. Published by Elsevier Inc.