Automated identification of interstitial lung abnormalities in lung screening using quantitative CT.

OBJECTIVES: Interstitial lung abnormalities (ILAs) are an important incidental finding in lung screening. Quantitative CT (qCT) offers a promising approach to standardising ILA assessment; however, its adoption into routine clinical practice remains limited, largely due to the need for further clinical validation. SETTING: We evaluated the performance of e-Lung (Brainomix), a commercially available qCT tool in assessing ILAs for participants attending a lung cancer screening programme. PARTICIPANTS: All participants invited to attend the West London lung cancer screening pilot programme between 2018 and 2020. PRIMARY OUTCOME MEASURES: To define the optimal qCT biomarker thresholds that identified ILA as defined by expert thoracic radiologists. RESULTS: e-Lung qCT biomarkers had an area under the curve of between 0.82 and 0.88, and between 0.84 and 0.87 for visually quantified ILA extent thresholds of at least 5% and >10%, respectively. CONCLUSION: e-Lung is a qCT tool with potential utility in the automated identification of ILA in participants undergoing lung screening.