Risk of Relapse and Death from Colorectal Cancer and its Related Factors Using Non-Markovian Multi-State Model.

Introduction: The objective of this investigation was to employ multi - state models for the purpose of modeling the likelihood of local relapse and mortality following the initial treatment for colorectal cancer, along with its associated determinants. In the realm of cancer research, the modeling of disease progression with respect to various events experienced by patients holds significant significance. Multi - state models, designed to account for intermediate events while treating death as the ultimate endpoint, have been advanced for this purpose. Methods: Conducted as a historical cohort study, this research focused on 235 individuals diagnosed with colorectal cancer, who sought medical attention at Omid Hospital in Mashhad between 2006 and 2011. The participants were monitored continuously until 2017. Transition probabilities to death resulting from metastasis, both with and without the occurrence of local relapse, along with pertinent variables, were ascertained using a non - Markovian multi - state model. The study delineated three distinct states of disease, encompassing local relapse and death. Results: The likelihood of not encountering any events, experiencing only relapse, or facing death within the initial 5 years were 0.45, 0.09, and 0.46, respectively. In cases where patients remained event - free during the first year of treatment, the probabilities of relapse and death before the fifth year were 0.04 and 0.33, respectively. Conversely, if relapse occurred during this period, the probability of death by the fifth year escalated to 0.62. The cancer stage exhibited associations with both relapse and death, while factors such as ethnicity and a history of addiction were linked to death without relapse. Additionally, BMI demonstrated a significant correlation with death following relapse (p < 0.05). Conclusions: Risk of death in patients with colorectal cancer depends on local relapse and the time between them. [ABSTRACT FROM AUTHOR]