Parametric mixture model for analysing relative survival of patients with multiple cancers.

Abstract

Background: This study aimed to provide a parametric mixture model for analysing relative survival and cure rates of patients with multiple cancers.

Methods: A new model was introduced by extending a parametric mixture model on relative survival of patients with a single cancer. The model was applied to empirical data on lung cancer as a first and second tumour after a first localised colorectal cancer.

Results: Survival and cure from subsequent cancer are estimatable with the parametric mixture model on relative survival of patients with multiple cancers. Survival from lung cancer does not differ between the first and second tumour, but there is suggestive evidence that survival from second lung cancer is higher than that from the first lung cancer.

Discussion: Relative survival analysis of patients with multiple cancers can be extended to incorporate cure rates and excess hazards related to the individual cancers. Prognostic factors can be included in the model.