Modeling the complex smoking exposure history in assessment of pan-cancer risk

This study used UK Biobank data from ∼500 000 participants to examine how smoking intensity and duration jointly shape the risk of 12 smoking-related cancers. We modeled excess relative risks (ERRs) per pack-year, separating the effect of smoking intensity while holding total pack-years constant to compare low-intensity/long-duration with high-intensity/short-duration patterns. For the same cumulative exposure, cancer risk increases differed across these smoking patterns, indicating that pack-years alone do not fully capture carcinogenic risk. We further evaluated time since cessation (TSC) and found that longer TSC was associated with lower ERR/pack-year for most cancers. Lifestyle factors modified these associations: participants with favorable behaviors (e.g., regular physical activity and healthier diet) generally had lower ERRs/pack-year than those with unfavorable behaviors. Overall, our findings show that smoking pattern and cessation history meaningfully influence pan-cancer risk beyond cumulative exposure, supporting the incorporation of detailed smoking histories-intensity, duration, and cessation-into risk stratification, screening eligibility, and prevention strategies.