Flexible estimation in cure survival models using Bayesian P-splines

In the analysis of survival data, it is usually assumed that any unit will experience the event of interest if it is observed for a sufficient long time. However, one can explicitly assume that an unknown proportion of the population under study will never experience the monitored event. The promotion time model, which has a biological motivation, is one of the survival models taking this feature into account. The promotion time model assumes that the failure time of each subject is generated by the minimum of N latent event times which are independent with a common distribution independent of N . We propose an extension which allows the covariates to influence simultaneously the probability of being cured and the latent distribution. We estimate the latent distribution using a flexible Cox proportional hazard model where the logarithm of the baseline hazard function is specified using Bayesian P-splines. Introducing covariates in the latent distribution implies that the population hazard function might not have a proportional hazard structure. However, the use of the P-splines provides a smooth estimation of the population hazard ratio over time. We propose a restricted use of the model when the follow up of the study is not sufficiently long. A simulation study evaluating the accuracy of our methodology is presented. The proposed model is illustrated on data from the phase III Melanoma e1684 clinical trial.