OBJECTIVES: LungFlag^TM is an A.I. risk prediction model that has proved to be effective in the selection of individuals for non-small cell lung cancer (NSCLC) screening. The aim of this analysis was to assess the cost-effectiveness of implementing LungFlag^TM for the selection of individuals at-risk for enrollment in NSCLC screening programs in Spain.

METHODS: A combination of a decision-tree and a Markov model was adapted to the Spanish setting. A multidisciplinary group of experts validated all the parameters and assumptions of the model. A hypothetical cohort of 3.835.128 individuals who met 2013 USPSTF criteria in Spain was defined as the target population. Screening with LungFlag^TM was compared to screening the whole population with no risk model and versus no-screening (current situation). Sensitivity and specificity values for LungFlag^TM were obtained from Gould et al. (2021). Adherence to screening programs was assumed to be the same for both screening scenarios. A lifetime horizon was used, so a 3% discount rate was applied for both cost and effects. Only direct medical costs were considered (€2023). Sensitivity analyses were performed to assess uncertainty.

RESULTS: Using LungFlag^TM to identify subjects at risk would result in 232,120 low-dose computed topographies (LDCTs) performed, compared to 2,1M of LDCTs if the whole target population was screened. Therefore, significant cost savings were obtained over the lifetime horizon (€-7,04M). However, this would mean fewer quality-adjusted life-years (QALYs) gained (-97,612 QALYs) because of fewer people screened. The incremental cost-utility ratio obtained showed that the savings generated by LungFlag^TM offset the loss of QALYs, making it a cost-effective strategy. Both screening scenarios (using LungFlag^TM or screening the whole target population) were dominant compared to no-screening.

CONCLUSIONS: Using LungFlag^TM for the selection of individuals at-risk of NSCLC may be a cost-effective strategy compared to screening all individuals meeting USPSTF 2013 criteria or no-screening.