Modeling Cure from Cancer Accounting for Inevitable Mortality

Area of Science:

• Biostatistics and Survival Analysis
• Cancer Epidemiology and Public Health

Background:

• Cancer is a leading cause of death, with survival rates improving due to treatment advances.
• Cancer patients face competing risks from primary cancer and other diseases, complicating survival analysis.
• Existing models inadequately address cure possibilities within competing risks frameworks.

Purpose of the Study:

• To propose a novel unified competing risks cure model for cancer survival data.
• To incorporate cause-specific cure from primary cancer while ensuring eventual overall survival approaches zero.
• To address limitations in current statistical methods for analyzing cancer patient survival under complex risk scenarios.

Main Methods:

• Developed a unified competing risks cure model based on the cause-specific hazard approach.
• Investigated model identifiability conditions.
• Employed a computationally efficient expectation-maximization algorithm for parameter estimation.

Main Results:

• The proposed model successfully integrates cure possibilities with competing risks.
• The expectation-maximization algorithm provides efficient parameter estimation.
• Simulation studies confirm the model's performance across various settings and competing risks.

Conclusions:

• The novel unified competing risks cure model offers a robust framework for cancer survival analysis.
• The model accurately reflects the complexities of patient outcomes, including cure and competing events.
• Application to breast cancer data from the SEER database demonstrates practical utility.