Cost-effectiveness of a machine learning risk prediction model (LungFlag) in the selection of high-risk individuals for non-small cell lung cancer screening in Spain

Abstract

Objective: The LungFlag risk prediction model uses individualized clinical variables to identify individuals at high-risk of non-small cell lung cancer (NSCLC) for screening with low-dose computed tomography (LDCT). This study evaluates the cost-effectiveness of LungFlag implementation in the Spanish setting for the identification of individuals at high-risk of NSCLC.

Methods: A model combining a decision-tree with a Markov model was adapted to the Spanish setting to calculate health outcomes and costs over a lifetime horizon, comparing two hypothetical scenarios: screening with LungFlag versus non-screening, and screening with LungFlag versus screening the entire population meeting 2013 US Preventive Services Task Force (USPSTF) criteria. Model inputs were obtained from the literature and the clinical practice of a multidisciplinary expert panel. Only direct costs (€of 2023), obtained from local sources, were considered. Deterministic and probabilistic sensitivity analyses were performed to assess the robustness of our results.

Results: A cohort of 3,835,128 individuals meeting 2013 USPSTF criteria would require 2,147,672 LDCTs scans. However, using LungFlag would only require 232,120 LDCTs scans. Cost-effectiveness results showed that LungFlag was dominant versus non-screening scenario, and outperformed the scenario where the entire population were screened since the observed loss of effectiveness (-224,031 life years [LYs] and -97,612 quality-adjusted life years [QALYs]) was largely offset by the significant cost savings provided (€7,053 million). The resulting incremental cost-effectiveness ratio (ICER) for this strategy of screening the whole population versus using LungFlag was €72,000/QALY, showing that LungFlag is cost-effective. Various were described, such as the source of the efficacy or adherence rates, and other limitations inherent to cost-effectiveness analyses.

Conclusions: Using LungFlag for the selection of high-risk individuals for NSCLC screening in Spain would be a cost-effective strategy over screening the entire population meeting USPSTF 2013 criteria and is dominant over non-screening.

Keywords: C52; Cost-effectiveness analysis; I18; LungFlag; lung cancer screening; non-small cell lung cancer; risk prediction model.