Disentangled Multi-modal Learning of Histology and Transcriptomics for Cancer Characterization

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Yupei, Wang, Xiaofei, Liu, Anran, Yu, Lequan, Li, Chao
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908856426168320
author Zhang, Yupei
Wang, Xiaofei
Liu, Anran
Yu, Lequan
Li, Chao
author_facet Zhang, Yupei
Wang, Xiaofei
Liu, Anran
Yu, Lequan
Li, Chao
contents Histopathology remains the gold standard for cancer diagnosis and prognosis. With the advent of transcriptome profiling, multi-modal learning combining transcriptomics with histology offers more comprehensive information. However, existing multi-modal approaches are challenged by intrinsic multi-modal heterogeneity, insufficient multi-scale integration, and reliance on paired data, restricting clinical applicability. To address these challenges, we propose a disentangled multi-modal framework with four contributions: 1) To mitigate multi-modal heterogeneity, we decompose WSIs and transcriptomes into tumor and microenvironment subspaces using a disentangled multi-modal fusion module, and introduce a confidence-guided gradient coordination strategy to balance subspace optimization. 2) To enhance multi-scale integration, we propose an inter-magnification gene-expression consistency strategy that aligns transcriptomic signals across WSI magnifications. 3) To reduce dependency on paired data, we propose a subspace knowledge distillation strategy enabling transcriptome-agnostic inference through a WSI-only student model. 4) To improve inference efficiency, we propose an informative token aggregation module that suppresses WSI redundancy while preserving subspace semantics. Extensive experiments on cancer diagnosis, prognosis, and survival prediction demonstrate our superiority over state-of-the-art methods across multiple settings. Code is available at https://github.com/helenypzhang/Disentangled-Multimodal-Learning.
format Preprint
id arxiv_https___arxiv_org_abs_2508_16479
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Disentangled Multi-modal Learning of Histology and Transcriptomics for Cancer Characterization
Zhang, Yupei
Wang, Xiaofei
Liu, Anran
Yu, Lequan
Li, Chao
Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
Histopathology remains the gold standard for cancer diagnosis and prognosis. With the advent of transcriptome profiling, multi-modal learning combining transcriptomics with histology offers more comprehensive information. However, existing multi-modal approaches are challenged by intrinsic multi-modal heterogeneity, insufficient multi-scale integration, and reliance on paired data, restricting clinical applicability. To address these challenges, we propose a disentangled multi-modal framework with four contributions: 1) To mitigate multi-modal heterogeneity, we decompose WSIs and transcriptomes into tumor and microenvironment subspaces using a disentangled multi-modal fusion module, and introduce a confidence-guided gradient coordination strategy to balance subspace optimization. 2) To enhance multi-scale integration, we propose an inter-magnification gene-expression consistency strategy that aligns transcriptomic signals across WSI magnifications. 3) To reduce dependency on paired data, we propose a subspace knowledge distillation strategy enabling transcriptome-agnostic inference through a WSI-only student model. 4) To improve inference efficiency, we propose an informative token aggregation module that suppresses WSI redundancy while preserving subspace semantics. Extensive experiments on cancer diagnosis, prognosis, and survival prediction demonstrate our superiority over state-of-the-art methods across multiple settings. Code is available at https://github.com/helenypzhang/Disentangled-Multimodal-Learning.
title Disentangled Multi-modal Learning of Histology and Transcriptomics for Cancer Characterization
topic Image and Video Processing
Artificial Intelligence
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.16479