MUST: Modality-Specific Representation-Aware Transformer for Diffusion-Enhanced Survival Prediction with Missing Modality
This addresses a critical challenge in precision oncology by enabling accurate survival predictions even when medical data modalities are missing, which is common in clinical settings.
The paper tackles the problem of survival prediction from multimodal medical data with frequently incomplete modalities by proposing MUST, a framework that decomposes modality representations into specific and cross-modal components and uses diffusion models to generate missing information, achieving state-of-the-art performance on five TCGA cancer datasets with robust predictions in missing modality conditions.
Accurate survival prediction from multimodal medical data is essential for precision oncology, yet clinical deployment faces a persistent challenge: modalities are frequently incomplete due to cost constraints, technical limitations, or retrospective data availability. While recent methods attempt to address missing modalities through feature alignment or joint distribution learning, they fundamentally lack explicit modeling of the unique contributions of each modality as opposed to the information derivable from other modalities. We propose MUST (Modality-Specific representation-aware Transformer), a novel framework that explicitly decomposes each modality's representation into modality-specific and cross-modal contextualized components through algebraic constraints in a learned low-rank shared subspace. This decomposition enables precise identification of what information is lost when a modality is absent. For the truly modality-specific information that cannot be inferred from available modalities, we employ conditional latent diffusion models to generate high-quality representations conditioned on recovered shared information and learned structural priors. Extensive experiments on five TCGA cancer datasets demonstrate that MUST achieves state-of-the-art performance with complete data while maintaining robust predictions in both missing pathology and missing genomics conditions, with clinically acceptable inference latency.