Tabular Foundation Models for Clinical Survival Analysis via Survival-Aware Adaptation

2026-06-10

AI Summary

This study addresses the challenge of right-censored time-to-event prediction in clinical settings, where labeled data are scarce and conventional models require task-specific training. It introduces a novel approach that integrates tabular foundation models—such as TabPFN, TabDPT, and TabICL—with survival analysis by attaching a lightweight multi-task logistic regression (MTLR) head, without fine-tuning the pre-trained backbone. Evaluated on real-world ICU cohorts from MIMIC-IV and eICU, the method achieves significant performance gains: a C-index of 0.856 on MIMIC-IV, surpassing the best non-foundation and zero-shot baselines by 1.4% and 6.7%, respectively, and 0.797 on eICU, with corresponding improvements of 1.7% and 6.4%. This work effectively overcomes a key bottleneck in applying tabular foundation models to continuous-time survival prediction.

Abstract

Predicting time-to-event outcomes such as mortality is a fundamental task in clinical decision-making, commonly addressed through survival analysis. While classical statistical and deep learning approaches have been widely studied, they typically require task-specific training and sufficient labeled data. Recent advances in tabular foundation models offer a new paradigm by learning general-purpose representations for structured data. However, their applicability to censored time-to-event prediction in clinical settings remains underexplored, as typical applications are restricted to discrete classification rather than survival analysis tasks. In this work, we propose a lightweight adaptation approach for applying tabular foundation models to clinical survival analysis by directly training a survival-aware head on top of the pretrained representations. We study representative architectures, including TabPFN, TabDPT, and TabICL, and adapt them using a multi-task logistic regression (MTLR) head to model right-censored time-to-event outcomes. We evaluate this approach on a diverse set of public survival benchmarks and two large-scale ICU cohorts, MIMIC-IV and eICU. Our results show that this transfer learning approach achieves competitive or superior performance compared to strong baselines. On MIMIC-IV, TabDPT-FT-MTLR reaches a C-index of 0.856, corresponding to a relative improvement of +1.4% over the best non-FM baseline (DeepSurv, 0.844) and +6.7% over the best zero-shot model (0.802). On eICU, TabICL-FT-MTLR achieves 0.797, yielding gains of +1.7% (DeepSurv, 0.784) and +6.4% (0.749), respectively. These findings highlight the importance of combining pretrained tabular representations with survival-aware objectives and suggest that tabular foundation models provide a practical and effective alternative for clinical survival prediction.

Problem

• tabular foundation models
• clinical survival analysis
• time-to-event prediction
• right-censoring
• transfer learning

Innovation

• tabular foundation models
• survival analysis
• MTLR
• transfer learning
• clinical prediction