The Observed Sex Difference: Observational data from lung cancer screening programs have consistently suggested that women appear to derive greater mortality reduction from screening compared to men. However, the reasons for this difference have remained incompletely understood.
Why It Matters: If women genuinely benefit more from screening, this could influence screening eligibility criteria, frequency recommendations, and resource allocation. Alternatively, if the difference is explained by confounders such as competing risks and comorbidities, the apparent advantage would be spurious.
NLST as the Data Source: The National Lung Screening Trial enrolled 53,452 high-risk smokers (31,530 men and 21,922 women) and randomized them to low-dose CT or chest X-ray screening. With 6-7 years of follow-up, NLST provides the largest dataset to rigorously analyze sex differences in screening effectiveness.
Study Hypothesis: The authors hypothesized that women's apparent better outcomes might be explained by differences in smoking history, lung cancer histology, comorbidities, and competing causes of death - rather than a true biological advantage in CT screening benefit.
Cause-Specific Hazard vs. Subdistribution Hazard: Standard survival analysis can be misleading when comparing sex differences because men have higher rates of competing non-cancer mortality (cardiovascular disease, COPD). The sub-distribution hazard ratio (sHR) from Fine-Gray regression accounts for competing risks, providing a more accurate estimate of screening's net effect on lung cancer-specific mortality.
Comorbidity Adjustment: The analysis systematically adjusted for baseline comorbidities including COPD severity, cardiovascular disease history, and overall health status using the Charlson Comorbidity Index. Men in the NLST had significantly higher comorbidity burdens, which increase non-lung cancer mortality.
Smoking History Matching: Despite similar smoking eligibility criteria, men had higher average pack-year exposures. Analyses were stratified and adjusted by detailed smoking history to isolate sex-specific effects from smoking intensity differences.
Cancer-Specific Outcomes: Primary outcomes included lung cancer-specific mortality (using competing risks analysis) and overall mortality. Secondary analyses examined screen-detected cancer characteristics, lead time, and overdiagnosis rates by sex.
Women's Apparent Advantage: In unadjusted analysis, women undergoing CT screening had a sub-distribution hazard ratio of 0.74 for lung cancer mortality (a 26% relative reduction), while men had an sHR of 0.91 (a 9% relative reduction that was not statistically significant). This raw comparison suggested women benefit substantially more.
Competing Risks Explain Much of the Difference: After adjusting for higher comorbidity burdens and competing non-cancer mortality in men, the sex difference in CT screening benefit became substantially attenuated. Men's higher background mortality reduces the relative observable benefit of screening on lung cancer death.
Absolute vs. Relative Benefit: When absolute risk reduction rather than relative hazard ratios was examined, the sex difference in benefit was smaller, because women had lower baseline lung cancer mortality rates. This highlights the importance of absolute risk framing in screening discussions.
Histological Differences: Women in NLST had a higher proportion of adenocarcinoma (which has slower growth and higher CT detection sensitivity) relative to squamous cell carcinoma and SCLC (which are harder to detect early). This histological distribution independently contributed to women's higher screen-detected early-stage proportions.
COPD and Cardiovascular Disease: Male NLST participants had significantly higher rates of severe COPD and established cardiovascular disease. These conditions are major competing causes of death that reduce the proportion of deaths attributable to lung cancer, thereby reducing the observable benefit of lung cancer-specific screening.
Overall Survival vs. Lung Cancer-Specific Survival: The sex difference in overall mortality reduction from CT screening was smaller than the difference in lung cancer-specific mortality reduction. This is consistent with men dying from other causes even when lung cancer mortality is reduced, diluting the all-cause survival benefit.
Smoking Cessation Patterns: Former versus current smoking status at enrollment differed by sex, with men more likely to be current smokers. Smoking cessation reduces lung cancer incidence and shifts the risk-benefit profile of screening over time in ways that differed between men and women.
Implications for Individualized Screening Decisions: These findings suggest that screening benefit predictions should be individualized based on comorbidity burden rather than sex alone. Men with low comorbidity burdens may derive benefit comparable to average women, while highly comorbid individuals of either sex may gain less from screening.
Against Sex-Specific Eligibility Criteria: These findings do not support creating sex-differentiated lung cancer screening eligibility criteria. The apparent sex difference in benefit is largely a statistical artifact of competing risk differences rather than a true biological screening advantage for women.
Comorbidity-Based Screening Decisions: Rather than sex-stratified thresholds, individualized risk assessment tools that incorporate comorbidity status and expected non-lung cancer mortality could provide more equitable and accurate predictions of who benefits from screening.
Communication with Patients: Clinicians should discuss screening benefit in terms of absolute risk reduction rather than relative hazard ratios, which can be misleading. For men with heavy comorbidities, framing screening benefit realistically helps informed decision-making.
Future Screening Guideline Updates: As guidelines evolve to include younger patients or lighter smokers, sex-specific analyses should be repeated in these expanded eligibility populations, where comorbidity distributions and histological patterns may differ from the NLST cohort.
NLST Population Characteristics: NLST enrolled only heavy smokers aged 55-74. Analyses cannot be generalized to the expanded USPSTF 2021 eligibility criteria (50+ years, 20+ pack-years), where sex and comorbidity distributions may differ and screening benefit may vary accordingly.
Residual Confounding: Despite extensive covariate adjustment, unmeasured confounders such as healthcare-seeking behavior, self-reported smoking accuracy, and socioeconomic factors may still contribute to observed sex differences in outcomes.
Overdiagnosis Question: This study did not fully address whether women experience higher rates of screening-related overdiagnosis of indolent adenocarcinomas that would not cause symptoms in their lifetime. Overdiagnosis is a critical harm of screening that disproportionately affects groups with more indolent tumor biology.
Integration with Risk Models: Future research should assess whether incorporating competing risk and comorbidity parameters into existing lung cancer risk prediction models (PLCOm2012, LCRAT) improves screening benefit predictions across sex and age strata in prospective screening settings.