Cost-effectiveness of combining AI and cfDNA with LDCT for lung cancer screening in China: a modelling study

Background: Lung cancer remains one of the leading causes of cancer-related mortality in China, where early detection through screening is critical for improving survival rates. Low-dose computed tomography (LDCT) has proven effective for early lung cancer screening, but its high false-positive rate increases the economic burden and psychological stress on patients. Previous studies have shown that combining artificial intelligence (AI) and cell-free DNA methylation (cfDNA) can potentially reduce the false-positive rates of LDCT. However, the cost-effectiveness of using AI and cfDNA in combination with LDCT for lung cancer screening in the Chinese population remains unclear. Furthermore, current guidelines vary in the risk factors and thresholds. The impact of them on the cost-effectiveness of screening strategies remains underexplored. This study aims to evaluate the cost-effectiveness of combining AI and cfDNA with LDCT for lung cancer screening in China. Additionally, we assess the impact of varying smoking exposure thresholds (20 pack-years vs. 30 pack-years) and environmental or occupational risks on screening outcomes. Methods: We simulated a cohort of 100,000 individuals aged 45-74, stratified by screening methods (LDCT alone, LDCT+AI+cfDNA, and no screening), risk factors (20 pack-years, 30 pack-years, and environmental/occupational exposures), and screening intervals (annual, biennial, and one-time screening). A Markov state transition model with a lifetime horizon was used to simulate lung cancer progression and related health outcomes. The model was validated against lung cancer-specific mortality data from the Global Burden of Disease study. Primary outcomes were incremental cost-effectiveness ratios (ICERs), life years (LYs), and quality-adjusted life years (QALYs). Sensitivity analyses were performed to test the robustness of results under different parameter assumptions. Value-based pricing analysis was performed to evaluate the maximum cost of AI+cfDNA at the current willingness-to-pay threshold. Findings: For individuals aged 45-49, the one-time LDCT screening strategy was the most cost-effective, with an incremental cost-effectiveness ratio (ICER) of 5,458 USD/QALY. For those aged 50-74, annual LDCT screening for individuals with a smoking history of 20 pack-years and environmental or occupational exposures was the most cost-effective (ICER range: 4,382-33,204 USD/QALY). At current pricing, AI+cfDNA combined with LDCT was not cost-effective. The value-based pricing analysis revealed that AI+cfDNA combined with LDCT would become cost-effective if the AI+cfDNA cost was reduced to a range of $232-$340. Interpretation: Annual LDCT screening for individuals aged 50-74 with a smoking history of 20 pack-years and environmental or occupational risks is the most cost-effective strategy in China. For younger individuals (aged 45-49), a one-time LDCT screening is cost-effective. While the combination of AI and cfDNA offers the potential to reduce false positives, significant cost reductions are necessary for it to become a viable screening option in China’s healthcare system. Funding: No external funding was received.