Multimodal Survival Analysis with Locally Deployable Large Language Models
This work addresses survival prediction for medical institutions with tight computational and privacy needs, though it is incremental as it adapts existing methods to a specific deployment setting.
The paper tackled multimodal survival analysis by integrating clinical text, tabular data, and genomic profiles using locally deployable large language models to address computational and privacy constraints, achieving improved performance over standard baselines on a TCGA cohort while reducing risks like hallucinations and miscalibration.
We study multimodal survival analysis integrating clinical text, tabular covariates, and genomic profiles using locally deployable large language models (LLMs). As many institutions face tight computational and privacy constraints, this setting motivates the use of lightweight, on-premises models. Our approach jointly estimates calibrated survival probabilities and generates concise, evidence-grounded prognosis text via teacher-student distillation and principled multimodal fusion. On a TCGA cohort, it outperforms standard baselines, avoids reliance on cloud services and associated privacy concerns, and reduces the risk of hallucinated or miscalibrated estimates that can be observed in base LLMs.