A deep learning-based multiscale integration of spatial omics with tumor morphology
Benoît Schmauch, Loïc Herpin, Antoine Olivier, Thomas Duboudin, Rémy Dubois, Lucie Gillet, Alexandre Filiot, Jean-Baptiste Schiratti, Valentina Di Proietto, Delphine Le Corre, A Bourgoin, Julien Taı̈eb, Jean‐François Emile, Wolf H. Fridman, Elodie Pronier, Pierre Laurent‐Puig, Eric Y. Durand
Abstract
Spatial Transcriptomics (spTx) offers unprecedented insights into the spatial arrangement of the tumor microenvironment, tumor initiation/progression and identification of new therapeutic target candidates. However, spTx remains unlikely to be routinely used in the near future. Hematoxylin and eosin (H&E) stained histological slides, on the other hand, are routinely generated for a large fraction of cancer patients. Here, we present a deep learning-based approach for multiscale integration of spTx with tumor morphology (MISO). We train MISO to predict spTx from H&E and validate it on a dataset of 72 10X Genomics Visium samples. We further validate our approach on 348 samples from five indications from the MOSAIC consortium and show that MISO significantly outperforms competing methods in extensive benchmarks. We demonstrate that MISO enables near single-cell-resolution, spatially-resolved gene expression prediction. Spatial transcriptomics technologies are still too restrictive for widespread clinical use, and methods that have been designed to bridge them with histopathology carry important limitations. Here, the authors develop MISO, a deep learning framework that allows inferring tissue spatial organisation and gene expression with near single-cell resolution from histopathology images.