Prediction of gene expression from H&E slides

Context

Predicting gene expression and genomic aberrations in tumors from whole slide images (WSIs) would greatly facilitate patient diagnosis, prediction of response to treatment, and survival outcome.

‍

Methods

Owkin developed a model called HE2RNA able to extract molecular features from histology images alone, providing a powerful tool to enhance clinical research and practice — from diagnosis to treatment optimization.

‍

Results

HE2RNA predicts the transcriptome without the need for RNA-Seq. For each predicted gene, it can map the virtual spatial representation of its expression.

‍

Impact

Identify novel biomarkers in images of specific tumoral regions that are important to better understand disease evolution and differentiated outcomes.

‍

Published in Nature Communications: Schmauch  et al. 2020

‍