Commonly Applied Selection Criteria for Lung Cancer Screening May Have Strongly Varying Diagnostic Performance in Different Countries

SIMPLE SUMMARY: Smoking causes the majority of lung cancers. Smoking history is thus used to select individuals among whom screening for lung cancer could be the most beneficial. The aim of our study was to estimate sensitivity and specificity of pre-selection by heavy smoking in individual European countries. Due to differences in smoking histories across the countries and sexes within the countries, the sensitivities were found to be between 33 and 80% for men and between 9 and 79% for women. Corresponding specificities of heavy smoking varied between 48 and 90% (men) and 70 and 99% (women). Our results may inform the design of lung cancer screening programs in European countries and serve as benchmarks for novel alternative or complementary tests for selecting people at high risk for computed tomography-based lung cancer screening. ABSTRACT: Lung cancer (LC) screening often focuses heavy smokers as a target for screening group. Heavy smoking can thus be regarded as an LC pre-screening test with sensitivities and specificities being different in various populations due to the differences in smoking histories. We derive here expected sensitivities and specificities of various criteria to preselect individuals for LC screening in 27 European countries with diverse smoking prevalences. Sensitivities of various heavy-smoking-based pre-screening criteria were estimated by combining sex-specific proportions of people meeting these criteria in the target population for screening with associations of heavy smoking with LC risk. Expected specificities were approximated by the proportion of individuals not meeting the heavy smoking definition. Estimated sensitivities and specificities varied widely across countries, with sensitivities being generally higher among men (range: 33–80%) than among women (range: 9–79%), and specificities being generally lower among men (range: 48–90%) than among women (range: 70–99%). Major variation in sensitivities and specificities was also seen across different pre-selection criteria for LC screening within individual countries. Our results may inform the design of LC screening programs in European countries and serve as benchmarks for novel alternative or complementary tests for selecting people at high risk for CT-based LC screening.