Performance of smoking duration-based lung cancer screening eligibility criteria: a comparative modeling study

Background Replacing the pack-year criterion for lung cancer screening with smoking duration has been proposed to reduce racial disparities in eligibility. However, no studies have evaluated the potential long-term benefits and harms of duration-based criteria. We conducted a comparative modeling study to evaluate the effectiveness of duration-based eligibility vs that from current guidelines and other alternative criteria.
Methods We used 3 Cancer Intervention and Surveillance Modeling Network models to evaluate the performance of duration-based screening with various smoking duration thresholds (15-40 years in 5-year increments) for the 1960 and 1970 US birth cohorts. We first examined age-specific eligibility patterns of duration-based strategies vs pack-year and risk-based criteria. We then evaluated the performance of different strategies, comparing the resulting number of screens, lung cancer deaths averted, life-years gained, false-positive screens, and overdiagnosed cases. We compared the strategies' efficiency using lung cancer deaths averted and life-years gained per screen and the benefit-to-harm ratios using lung cancer deaths averted and life-years gained per overdiagnosed case.
Results Risk-based criteria resulted in the most lung cancer deaths averted and life-years gained but also in more overdiagnosed cases. Duration-based strategies with a 35-year cutoff achieved comparable lung cancer deaths averted and life-years gained to current US guidelines and resulted in similar false-positive and overdiagnosed cases per screen. Duration-based scenarios with a 20-year cutoff required substantially more screenings but yielded only modest additional lung cancer deaths averted and life-years gained, resulting in lower benefits per screen than current guidelines.
Conclusion Duration-based screening may be as efficient as current US guidelines. Given their potential to reduce disparities in eligibility shown in recent studies and simpler implementation, duration-based criteria warrant consideration.