Factors affecting survival in bone marrow transplantation using mixture cure model

Abstract

Background & Aims: Bone marrow transplantation (BMT) is a curative treatment for various hematological malignancies. In standard survival models, the possibility of a cure has not been considered. Mixture cure models, which account for the possibility of a cure, can provide valuable insights into patient outcomes. The purpose of this study was to apply a smooth semi-nonparametric analysis for the mixture cure model to determine risk factors for survival and effective factors for the cure in bone marrow transplant patients. Materials & Methods: Data from BMT patients treated at Taleghani Hospital in Tehran were analyzed using a Weibull mixture cure model and an accelerated failure time mixture cure (AFTMC) model with an exponential kernel. The goodness-of-fit of each model was assessed using Akaike's information criterion (AIC). Results: The Weibull mixture cure model indicated that non-Hodgkin's lymphoma and acute leukemia were significantly associated with time to death. Age, recurrence after transplant, and hemoglobin levels were associated with the cure probability. The AFTMC model confirmed the prognostic effects of age, non-Hodgkin's lymphoma, and acute leukemia on time to death and further revealed that age and recurrence after transplant also influenced the cure probability. Conclusion: The smooth semi-nonparametric approach to mixture cure models provides a comprehensive analysis of BMT patient outcomes, identifying both prognostic and curative factors. This information can guide treatment decisions and improve patient survival.

Keywords

• accelerated failure time model
• bone marrow transplant
• cure models
• semi-nonparametric density
• survival analysis