Optimizing precision lung cancer screening using willingness-to-pay–anchored risk eligibility thresholds

Abstract Background Despite the proven benefits of low-dose computed tomography (LDCT) in lung cancer management, uptake remains limited due to cost. This study applies a structured decision-analytic framework to optimize LDCT screening through risk-based strategies in Singapore compared with conventional LDCT (based on age and smoking history) and no screening. Methods We developed a lifetime Markov model to simulate costs and outcomes for Singaporean smokers and never-smokers with a family history of lung cancer, stratified by sex. Transition probabilities and costs were informed by local data, with utilities from published sources. Screening strategies varied by start age (50, 55, 60) and frequency (annual, biennial, triennial). Outcomes included costs, quality-adjusted life-years (QALYs), and incremental cost-effectiveness ratios (ICERs), discounted at 3% annually. Using a validated local risk prediction model, reverse threshold analyses identified optimal risk cut-offs for cost-effective screening, supported by deterministic and probabilistic sensitivity analyses across willingness-to-pay thresholds of Singapore dollars (SGD) 45,000–75,000/QALY. Results Among smokers, annual risk-based screening from age 50 at 1-year risk thresholds of ≥ 0.37% for males and ≥ 0.51% for females produced the greatest health gains (ΔQALYs vs. no screening: 0.201 and 0.198, respectively), while the same risk-based annual screening from age 60 was most economically efficient (ΔQALYs: 0.177 and 0.182; incremental ICERs: SGD 40,155/QALY and SGD 42,666/QALY, respectively). Among never-smokers with a family history of lung cancer, conventional LDCT was not cost-effective; annual risk-based screening from age 50 in males at 1-year risk threshold of ≥ 0.16% and females at ≥ 0.21% maximized health gains (ΔQALYs: 0.013 and 0.021, respectively), while less intensive screening from age 60–biennial in males at 1-year risk ≥ 0.20% and triennial in females at ≥ 0.21%–was more economically efficient (ΔQALYs: 0.007 and 0.011; incremental ICERs: SGD 73,548/QALY and SGD 26,953/QALY, respectively). Conclusions Risk-based LDCT screening may improve screening efficiency and health gains while maintaining good value for money, with different screening intensities offering implementation options under varying budget constraints, though further research is needed to support its implementation in practice. These findings provide evidence to inform risk-based screening eligibility and implementation strategies, particularly in Asia with a high burden of lung cancer among never-smokers.