A presmoothing approach for estimation in semiparametric mixture cure models

TL;DR

This paper introduces a presmoothing estimation method for semiparametric mixture cure models in survival analysis, improving accuracy over traditional methods and demonstrated through simulations and melanoma data applications.

Contribution

A novel presmoothing estimation procedure for cure rate models that is independent of the latency model and shows improved accuracy.

Findings

Presmoothing improves estimation accuracy in logistic/Cox cure models.

The method performs well in simulations compared to maximum likelihood.

Application to melanoma data demonstrates practical utility.

Abstract

A challenge when dealing with survival analysis data is accounting for a cure fraction, meaning that some subjects will never experience the event of interest. Mixture cure models have been frequently used to estimate both the probability of being cured and the time to event for the susceptible subjects, by usually assuming a parametric (logistic) form of the incidence. We propose a new estimation procedure for a parametric cure rate that relies on a preliminary smooth estimator and is independent of the model assumed for the latency. We investigate the theoretical properties of the estimators and show through simulations that, in the logistic/Cox model, presmoothing leads to more accurate results compared to the maximum likelihood estimator. To illustrate the practical use, we apply the new estimation procedure to two studies of melanoma survival data.