Abstract

Objectives: The aim was to compare, in a cohort of asbestos-exposed workers, the sensitivity and the specificity of low radiation helical chest CT scan with chest radiograph for the biennial screening of bronchopulmonary cancer, according to the size of detected nodules.

Material and methods: The screening procedure consisted in biennial chest radiograph and monodetector chest CT scan, proposed to 972 individuals having been highly exposed to asbestos. A total of 2,555 screening procedures were performed. The study focuses on the 1,230 screening procedures for which a 2-year follow-up period was available.

Results: Twenty-four cases of bronchopulmonary cancer were diagnosed. CT scan detected 20 cancers, 12 of which had not been detected by chest radiograph. Sensitivity of chest radiograph and CT scan, was respectively 33% and 83%, lesions measuring over 2mm in diameter being considered as suspect. The specificity of chest radiograph and CT scan was respectively 95% and 78%.

Calculation of the differential FP/TP ratio (false positive/true positive) and the “ROC curve”, performed for both chest radiograph and CT scan, facilitated the determination of the best possible compromise between specificity and sensitivity, according to the diameter threshold applied for considering a nodule as suspect.

Conclusions: Although this study confirms the superior sensitivity of chest CT scan compared to conventional chest radiograph, the associated loss in specificity leads to recommend a diameter of 5mm as the threshold for considering non-calcified lesions as "suspect", for the surveillance of asbestos-exposed individuals.