The Cost of Binarizing Survival Outcomes in Clinical Prognostic Modeling

2026-08-04

AI Summary

In clinical prognostic modeling, dichotomizing survival outcomes leads to the exclusion of censored samples, loss of temporal information, and omission of critical prognostic features. To address this, this work proposes a survival-aware Bayesian network that integrates time-to-event data directly within the Bayesian network framework by replacing conventional binary scoring with the Cox partial likelihood function during feature selection. Evaluated on datasets from head and neck, breast, colorectal, and kidney cancers, the method successfully recovers prognostic features overlooked by dichotomization. Ablation studies confirm that the performance gains stem from effective utilization of time-to-event information, demonstrating strong generalizability across cancer types.

Abstract

Survival analysis is an established framework for analyzing time-to-event data, yet many clinical machine learning studies still binarize the outcome before model training. This practice excludes censored patients, collapses temporal information into a single threshold, and can affect which features are selected as prognostically relevant. We examine the cost of this binarization in the context of Bayesian network (BN) feature selection, using two recent publications as case studies: one that applies BN-based feature selection to a head-and-neck cancer cohort and a second surgical cohort study that, while not BN-based, likewise binarizes its survival endpoint. We replace the binary scoring function with the Cox partial log-likelihood for feature-to-outcome edges, a modification we call the Survival-Aware Bayesian network, and recover prognostic features that binarization misses. Our ablation experiment confirms that the improvement is driven by the time-to-event scoring formulation rather than by retaining more patients. The results generalize across five endpoint-cohort combinations in head-and-neck cancer and extend to three further cancer types (breast, colorectal, and kidney). We propose that clinical studies with survival outcomes should use time-to-event methods by default, as binarization discards the prognostic signal retained by survival analysis.

Problem

• survival analysis
• binarization
• clinical prognostic modeling
• time-to-event data
• censored data

Innovation

• Survival-Aware Bayesian Network
• Cox partial log-likelihood
• feature selection
• time-to-event data
• binarization bias