CACSurv: Concordance-Aligned Comparative Learning with Large Language Models for Cancer Survival Prediction

Tianqi Xiang; Qixiang Zhang; Xinpeng Ding; Yi Li; Xiaomeng Li

cs.AIAug 17, 2026

CACSurv: Concordance-Aligned Comparative Learning with Large Language Models for Cancer Survival Prediction

Authors: Tianqi Xiang, Qixiang Zhang, Xinpeng Ding, Yi Li, Xiaomeng Li

Organizations: The Hong Kong University of Science and Technology

Abstract

Cancer survival prediction supports treatment planning, risk stratification, and follow-up management. Existing methods use structured clinical variables, whole-slide images, genomic profiles, or multimodal inputs, while patient reports remain underexplored. We study report-centric survival prediction using reports that organize pathological, clinical, and molecular evidence. Large language models (LLMs) can reason over such reports, but case-wise time regression introduces two mismatches. First, a formulation mismatch arises because survival evaluation depends on ordering comparable patients, whereas independent time predictions do not enforce ranking consistency. Second, a supervision mismatch arises because a censored patient's observed time indicates survival beyond that point and cannot serve as an exact regression target, although it still implies orderings relative to patients who died earlier. To address these mismatches, we propose CACSurv, a Concordance-Aligned Comparative framework for report-centric survival prediction. CACSurv reformulates survival modeling as mini-cohort comparative reasoning, where an LLM predicts relative prognostic orderings. We introduce concordance-aligned rewards derived from comparable relations under right censoring, enabling censored outcomes to provide ranking supervision without exact event-time targets. At inference, Monte Carlo Reference Aggregation compares each patient with sampled references and aggregates positions into a cohort-level ranking. We establish TCGA-SurvReport, a benchmark covering six TCGA cancer cohorts. CACSurv achieves the highest C-index on all six cohorts and an average C-index of 0.722, outperforming the strongest published survival model by 6.5 percentage points and the strongest LLM time-regression baseline by 4.2 percentage points. Our code, models, and dataset will be available at https://github.com/xmed-lab/CACSurv.

Figures & tables

Figure 1: Comparison of survival prediction formulations. (a) Conventional WSI/multimodal models predict a case-wise risk score. (b) A straightforward report-based LLM performs case-wise time regression, but patients who remain alive at the last follow-up lack exact-time targets. (c) CACSurv directly ranks patients within a mini-cohort with reliable relative comparisons.

Figure 2: Overview of CACSurv. During training, CACSurv learns survival orderings over pair and triplet mini-cohorts using concordance-aligned rewards and GRPO. During inference, Monte Carlo Reference Aggregation constructs multiple mini-cohorts for each evaluation patient using randomly sampled reference patients, then averages its normalized rank positions to obtain the final cohort-level survival ordering.

WSI/Multimodal Survival Models
Method	BLCA	BRCA	COADREAD	HNSC	LUAD	STAD	AVG@6
TransMIL	$0.571_{\pm 0.083}$	$0.640_{\pm 0.056}$	$0.642_{\pm 0.084}$	$0.568_{\pm 0.068}$	$0.548_{\pm 0.027}$	$0.599_{\pm 0.052}$	$0.595_{\pm 0.039}$
TITAN	$0.625_{\pm 0.065}$	$0.666_{\pm 0.033}$	$0.649_{\pm 0.060}$	$0.586_{\pm 0.033}$	$0.607_{\pm 0.048}$	$0.548_{\pm 0.074}$	$0.613_{\pm 0.043}$
PAMoE	$0.646_{\pm 0.054}$	$0.704_{\pm 0.038}$	$0.702_{\pm 0.050}$	$0.620_{\pm 0.032}$	$0.646_{\pm 0.035}$	$0.612_{\pm 0.070}$	$0.655_{\pm 0.036}$
DPSurv	$0.611_{\pm 0.064}$	$0.600_{\pm 0.055}$	$0.585_{\pm 0.089}$	$0.531_{\pm 0.074}$	$0.570_{\pm 0.020}$	$0.536_{\pm 0.041}$	$0.572_{\pm 0.033}$
MCAT	$0.620_{\pm 0.012}$	$0.564_{\pm 0.026}$	$0.612_{\pm 0.052}$	$0.518_{\pm 0.024}$	$0.571_{\pm 0.031}$	$0.528_{\pm 0.075}$	$0.569_{\pm 0.042}$

Table 1: Main results from 5-fold cross-validation on TCGA-SurvReport (C-index, mean

\pm

std). Time-ZS and Time-SFT denote direct survival-time prediction using zero-shot inference and supervised fine-tuning, respectively. Pair-only uses only pair mini-cohorts during both training and inference. Best results are shown in bold , and second-best results are underlined .

Figure 3: Cohort statistics of TCGA-SurvReport. For each cancer cohort, we report the number of patients, the proportions of alive/censored and dead/event cases, and their observed-time distributions. Observed time denotes the last follow-up time for censored patients and the event time for deceased patients. Each curve shows the percentage of patients within the corresponding status group in each time interval.

Setting	Eval.Pair	Eval.Mix
Train.Pair	0.702	0.703
Train.Mix	0.715	0.722

Table 2: Ablation on pair and triplet mini-cohort instantiations during training and inference. Mix means Pair+Triplet.

Optimization	BLCA	COADREAD	HNSC	AVG@6
Rank-SFT	0.656	0.740	0.653	0.682
Ours	0.685	0.806	0.689	0.722

Table 3: Ablation of ranking-based SFT and our concordance -aligned reward optimization under the same comparative formulation.

Setting	$K=10$	$K=20$	$K=40$	$K=80$
Train.Pair, Eval.Pair	0.692	0.700	0.702	0.704
Train.Mix, Eval.Pair	0.701	0.707	0.715	0.718
Train.Mix, Eval.Mix	0.706	0.715	0.722	0.723

Table 4: Sensitivity to the number of reference comparisons

K

during inference. Mix means Pair+Triplet.

Figure 4: Example CACSurv rollout on a pair mini-cohort. In the <think> section, the model identifies patient-specific prognostic factors, compares discriminative evidence across patients, and derives a ranking decision. The predicted survival ordering is then returned in the <answer> section.

Cancer Survival Prediction

Cancer survival prediction estimates patient outcomes and time to events from clinical, molecular, and medical-imaging data.

See all 89 papers

May 25, 2026

Towards end-to-end LLM-based censoring-aware survival analysis

Objective: Survival analysis is central to medical prediction, yet large language models (LLMs) are rarely used as end-to-end survival models because censoring prevents straightforward supervised fine-tuning. Here we present LLMSurvival, a framework that enables censoring-aware survival analysis with unmodified LLMs operating directly on tabular clinical data. Materials and Methods: LLMSurvival reformulates time-to-event prediction as pairwise ranking among comparable subjects, and derives test-time risk by aggregating comparisons against anchor individuals from the training cohort. Results: Across two clinical tasks (ICU mortality prediction in MIMIC-IV and fragility fracture prediction in a NewYork-Presbyterian/Weill Cornell Medicine cohort), LLMSurvival improves overall concordance over Cox proportional hazards modeling by 3.1% for ICU mortality and 0.5% for fracture risk, 2.1% on average for ICU mortality and 2.8% for fracture risk over three established deep learning survival models. Discussion: The results show that survival modeling with censoring can be made compatible with LLM fine-tuning through comparison-based reformulation. The framework demonstrates high portability and superior performance over expert curated scores like SAPS-II and FRAX scores across diverse clinical context. Furthermore, the framework supports local deployment, as compact, publicly available base models provide sufficient performance. Conclusion: The LLMSurvival framework serves as a proof of concept for an integrated, censoring-conscious approach to survival analysis via LLMs.

Yishu Wei, Hexin Dong, Yi Lin +3

Department of Population Health Science, Weill Cornell Medicine, New York, 10065, US · Department of Medicine, Weill Cornell Medicine, New York, 10065, US

Survival Analysis Mortality Prediction Medical LLMs

Ask PDF

Mar 23, 2026

Multimodal Survival Analysis with Locally Deployable Large Language Models

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.

Moritz Gögl, Christopher Yau

University of Oxford

Multimodal Survival Analysis Cancer Survival Prediction Medical LLMs

Ask PDF

Jul 23, 2026

A Multi-Cohort Validation of Censoring-Aware Conformal Lower Predictive Bounds for Pathology Survival Models

Whole-slide survival models commonly provide risk rankings without calibrated statements about individual event times. We evaluate fixed-cutoff drcosarc, a post-hoc conformal wrapper for discrete-time multiple-instance learning survival heads using frozen UNI2-h representations, in an internal 18-configuration sweep across five TCGA cohorts and an external five-configuration evaluation across three CPTAC cohorts. We distinguish configuration--fold--split summaries of the inverse-probability-of-censoring-weighted (IPCW) estimate and median lower predictive bound (LPB) from a hierarchy-aware patient-ensemble estimand of the mean drcosarc--naive LPB difference. At

α=0.1

, the drcosarc IPCW estimate was nearest 0.90 in KIRC, LUAD, and STAD. Patient-ensemble drcosarc--naive intervals excluded zero in KIRC, KIRP, STAD, UCEC, and CPTAC-CCRCC, but included zero in internal LUAD, CPTAC-LUAD, CPTAC-UCEC, and the internal LUSC extension. In a 20-replicate low-censoring semi-synthetic setting with known event times, drcosarc empirical coverage was 0.9129 [0.9053, 0.9207]. An exploratory analysis supported a head-error-by-censoring interaction within that data-generating process. In a two-cohort ABMIL sensitivity analysis, increasing the hazard grid to

K=16

raised localized marginal IPCW estimates above the prespecified 0.87 threshold and yielded positive paired LPB differences, although worst-group estimates remained below 0.87. Overall, performance was cohort dependent, and its interpretation changed with the patient-level unit, estimand, and censoring assumptions.

Mingi Hong

Independent Researcher

Computational Pathology Cancer Survival Prediction Survival Analysis Conformal Prediction Oncology Regression Prediction Intervals

Ask PDF

Abstract

Cancer Survival Prediction

Explore similar work

Towards end-to-end LLM-based censoring-aware survival analysis

Multimodal Survival Analysis with Locally Deployable Large Language Models

A Multi-Cohort Validation of Censoring-Aware Conformal Lower Predictive Bounds for Pathology Survival Models