A100-01 Trade-offs of Including Smoking Duration as a Criterion for Lung Cancer Screening Eligibility

Abstract Rationale Patient selection is central to the effectiveness of lung cancer screening (LCS). Emerging evidence has raised concerns about the reliance on a pack-year criterion and the restriction based on years since quitting. In particular, recent evidence suggests that a smoking-duration criterion may improve accuracy and fairness of LCS eligibility criteria compared to pack-years. Meanwhile, the adequacy of the age cutoff of 50 has not been adequately investigated. We conducted a comprehensive analysis of a nationally representative population of adults aged 40-80 who have ever smoked to evaluate the performance and tradeoffs of different smoking duration and age thresholds. Methods We analyzed data from the annual National Health Interview Surveys (NHIS) 1997-2018, including all participants age 40-80 years old who reported ever smoking. Criteria evaluated included U.S. Preventive Services Task Force (USPSTF) criteria and alternatives based on smoking duration (20+, 30+, 40+ smoke-years) and age (continuously examined age 40-80). Estimating expected life-years gained from LCS using the Life-Years from Screening-CT (LYFS-CT) model, we assessed each criteria’s sensitivity for selecting high-benefit individuals and specificity for excluding negligible-benefit individuals (I.e., individuals above or below prespecified LYFS-CT thresholds, respectively), accounting for survey weights. Results A 40+ smoke-years threshold was more efficient compared to USPSTF, selecting a smaller population (12,548,288 vs. 14,100,567) while identifying a greater proportion of high-benefit individuals (84% vs 78%) and excluding nearly all negligible-benefit individuals (100% vs 98%). A 30+ smoke-years threshold achieved even higher sensitivity for high-benefit selection (98%) with modestly reduced exclusion of negligible-benefit individuals (93%). By contrast, at 20+ smoke-years, all high-benefit individuals (100%) were captured but only 64% of negligible-benefit individuals were excluded. The selected population nearly doubles at 30+ smoke-years (25,874,488) and triples at 20 + (38,920,750). Age-based analyses identified 1,782 high-benefit individuals under 50 years who were ineligible under USPSTF criteria (Figure). Conclusions This analysis comprehensively demonstrates the clear trade-offs of different smoking duration and age cutoffs. A 20+ smoke-years threshold may be too broad as it includes many with limited benefit, while 30+ offers good targeting even as a single criterion. Ultimately, even optimized criteria should be paired with individualized prediction tools to align screening with patient-specific benefit, support personalized shared decision-making, and to expand screening responsibly. This abstract is funded by: VA ORD ICX002850A