Empirical Modelling of Cancer Incidence, Mortality and Mortality-to-Incidence Patterns in Iran: A GLOBOCAN 2020 Analysis

Abstract

Background: Site-specific incidence and mortality estimates permit descriptive comparisons of cancer burden, but their association must be interpreted in light of the limitations of ecological and modelled data. Objective: To characterise the association between estimates of cancer incidence and mortality across 35 cancer sites in Iran in 2020, and to describe site-level mortality-to-incidence and rank-based patterns. Methods: Published GLOBOCAN 2020 estimates for 35 specified cancer sites were examined using a through-origin linear regression, ordinary least-squares regression of natural-log-transformed counts, and descriptive rank analyses. Mortality-to-incidence ratios (MIRs) were calculated as deaths divided by incident cases in the same calendar year. Results: GLOBOCAN estimated 131,191 incident cancer cases and 79,136 deaths in Iran in 2020 (national deaths-to-cases ratio, 60.32%). The 35 listed sites accounted for 115,441 cases and 71,290 deaths. The through-origin incidence–mortality slope was 0.6029 (95% confidence interval [CI], 0.508–0.697). The log–log slope was 1.0470 (95% CI, 0.953–1.141; r = 0.9695). The unweighted mean of 35 site-specific MIRs was 56.74%; pancreatic (96.59%), liver (93.39%) and oesophageal (90.67%) cancers had the highest ratios among the selected sites. Conclusion: The estimates exhibit a positive cross-sectional association between site-specific incidence and mortality counts, with substantial variation in MIR across cancer sites. These in-sample empirical relationships are descriptive, are affected by the estimation of the underlying inputs, and should not be interpreted as patient survival probabilities, causal explanations or validated forecasts.