A deep learning-based multiscale integration of spatial omics with tumor morphology
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 complex and 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 novel deep learning-based approach for multiscale integration of spTx with tumor morphology (MISO). We trained MISO to predict spTx from H&E on a new unpublished dataset of 72 10X Genomics Visium samples, and derived a novel estimate of the upper bound on the achievable performance. We demonstrate that MISO enables near single-cell-resolution, spatially-resolved gene expression prediction from H&E. In addition, MISO provides an effective patient representation framework that enables downstream predictive tasks such as molecular phenotyping or MSI prediction.
Related
- smiMAD: adaptive multi-view graph learning for paired spatial omics integration
- Interpretable spatial multi-omics data integration and dimensionality reduction with SpaMV
- SubPoE: a subgraph-induced Product-of-Experts framework for spatial multi-omics integration
- SpaMOAL is a deep learning method that enables accurate spatial domain identification from multi-omics data
- BiomiX 3.0: A user-friendly platform for democratized multi-omics integration with graph-based learning
BibTeX
@article{schmauch2026deep,
title = {{A deep learning-based multiscale integration of spatial omics with tumor morphology}},
author = {Benoît Schmauch and Loïc Herpin and Antoine Olivier and Thomas Duboudin and Rémy Dubois and Lucie Gillet and Alexandre Filiot and Jean-Baptiste Schiratti and Valentina Di Proietto and Delphine Le Corre and Alexandre Bourgoin and Julien Taïeb and Jean-François Emile and Wolf H. Fridman and Elodie Pronier and Pierre Laurent-Puig and Eric Y. Durand},
journal = {Nature Communications},
year = {2026},
month = nov,
doi = {10.1038/s41467-025-66691-y},
url = {https://hal.science/hal-05851873v1},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - A deep learning-based multiscale integration of spatial omics with tumor morphology
AU - Benoît Schmauch
AU - Loïc Herpin
AU - Antoine Olivier
AU - Thomas Duboudin
AU - Rémy Dubois
AU - Lucie Gillet
AU - Alexandre Filiot
AU - Jean-Baptiste Schiratti
AU - Valentina Di Proietto
AU - Delphine Le Corre
AU - Alexandre Bourgoin
AU - Julien Taïeb
AU - Jean-François Emile
AU - Wolf H. Fridman
AU - Elodie Pronier
AU - Pierre Laurent-Puig
AU - Eric Y. Durand
PY - 2026
DA - 2026/11/04
JO - Nature Communications
DO - 10.1038/s41467-025-66691-y
UR - https://hal.science/hal-05851873v1
AB - 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 complex and 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 novel deep learning-based approach for multiscale integration of spTx with tumor morphology (MISO). We trained MISO to predict spTx from H&E on a new unpublished dataset of 72 10X Genomics Visium samples, and derived a novel estimate of the upper bound on the achievable performance. We demonstrate that MISO enables near single-cell-resolution, spatially-resolved gene expression prediction from H&E. In addition, MISO provides an effective patient representation framework that enables downstream predictive tasks such as molecular phenotyping or MSI prediction.
SN - 2041-1723
SN - 2041-1723
ER -