CenTime: Event-Conditional Modelling of Censoring in Survival Analysis

TL;DR

CenTime introduces a novel survival analysis method that directly estimates event times using an event-conditional censoring mechanism, improving accuracy especially with scarce uncensored data, and integrates seamlessly with deep learning models.

Contribution

The paper presents CenTime, a new survival analysis approach that robustly estimates event times and handles censored data effectively, even with limited uncensored samples, outperforming existing methods.

Findings

CenTime achieves state-of-the-art time-to-event prediction accuracy.

The method maintains competitive ranking performance.

CenTime is compatible with deep learning frameworks and requires no restrictions on batch size or uncensored data.

Abstract

Survival analysis is a valuable tool for estimating the time until specific events, such as death or cancer recurrence, based on baseline observations. This is particularly useful in healthcare to prognostically predict clinically important events based on patient data. However, existing approaches often have limitations; some focus only on ranking patients by survivability, neglecting to estimate the actual event time, while others treat the problem as a classification task, ignoring the inherent time-ordered structure of the events. Furthermore, the effective utilization of censored samples - training data points where the exact event time is unknown - is essential for improving the predictive accuracy of the model. In this paper, we introduce CenTime, a novel approach to survival analysis that directly estimates the time to event. Our method features an innovative event-conditional censoring mechanism that performs robustly even when uncensored data is scarce. We demonstrate that our approach forms a consistent estimator for the event model parameters, even in the absence of uncensored data. Furthermore, CenTime is easily integrated with deep learning models with no restrictions on batch size or the number of uncensored samples. We compare our approach with standard survival analysis methods, including the Cox proportional-hazard model and DeepHit. Our results indicate that CenTime offers state-of-the-art performance in predicting time-to-death while maintaining comparable ranking performance. Our implementation is publicly available at https://github.com/ahmedhshahin/CenTime.