Estimating Cure Fraction in Population-Based Survival Colon Cancer Data

In recent decades, cancer prognosis has notably improved, leading to a higher rate of patient cure. Monitoring the proportion of cured individuals is valuable for patients and provides insight into survival trends for treatable diseases. It is crucial to estimate cure fraction in a diseased population when other competing moralities are present, and Relative Survival offers a method to isolate disease-related mortality when cause of death data is unavailable or unreliable. In the context of cancer survival, relative survival often exhibits a plateau, indicating statistical cure. In this paper we use a mixture cure model with Weibull distribution to model relative survival of cancer patients and estimate the proportion of cured cases. We apply our methodology to a cohort of patients diagnosed with one of the most common cancers, colon cancer, and offer insights into their survival. The data is sourced from a long-established Italian population-based cancer registry.