AI-driven mortality analysis of three major neoplasms in Mexico.

13

Background: Mexico faces a growing cancer burden with limited comparative mortality data across major cancer types. We analyzed 13-year mortality trends for the three leading cancer killers using artificial intelligence-based statistical modeling to inform public health priorities. Methods: We analyzed 271,630 death certificates from Mexico's National Health Information System (SINAIS/INEGI, 2012-2024) for breast cancer (BC), colorectal cancer (CRC), and lung cancer (LC). AI-assisted analysis was implemented using Python 3.x and Claude Sonnet 4.5 on Linux Ubuntu 24 to optimize processing time, handle large-scale data, and generate robust projections. Age-standardized mortality rates were calculated using direct standardization (Mexico 2020 standard population). Second-degree polynomial regression models and Prophet algorithm were developed to project mortality through 2030. Model performance was assessed using R² and mean absolute percentage error (MAPE). 95% confidence intervals were calculated using normal distribution. Results: Analysis through 2024 showed concordance with National Institute of Statistics and Geography (INEGI) official data, validating our approach. Total deaths 2012-2024: BC 98,947; CRC 86,016; LC 86,667. Divergent trends emerged: BC showed stabilization after initial growth (2012-2020: +10.7%; 2020-2024: -4.8%). CRC demonstrated sustained acceleration (+5.43% annually; 2024 surge: +11.4%), surpassing LC for the first time in 2024. LC exhibited unique trend reversal with post-2016 decline (-10.2%). Mean age at death: BC 58.6 years; CRC 65.5 years; LC 69.4 years. All models achieved excellent predictive accuracy (R² >0.92, MAPE <2%). 2030 projections: BC 10,268 deaths (+17.8% vs 2024); CRC 10,362 (+17.5%); LC 6,907 (+9.3%). CRC is projected to become the second leading cancer mortality cause by 2030, despite being detectable early. Conclusions: CRC emerges as Mexico's critical epidemiological challenge, showing accelerated growth despite high prevention potential, indicating systematic failures in screening and risk factor control. Divergent trajectories across cancer types validate the need for differentiated policy strategies. The limited window of opportunity demands immediate implementation of national CRC screening programs and risk factor regulation to prevent an avoidable crisis by 2030.

Model performance and 2030 mortality projections.