DOI: 10.1148/ryct.260087 ISSN: 2638-6135
Comparison of Two Deep Learning Models and Their Combination for 1-year Lung Cancer Detection in Two Screening Cohorts
Anthony P. Reeves, Mohammad Mozafary, Ali Hashemi, Kyle Atlas, Chenyu Zhang, Thomas Atlas, Amir Azimi, Seyed Reza Mirjalili, Rowena Yip, Rozemarijn Vliegenthart, Juan P. de-Torres, J. Alberto Neder, Jessica González, Luis M. Seijo, Claudia Henschke, David Yankelevitz, Morteza NaghaviSybil and a Medical Open Network for Artificial Intelligence (MONAI)–based nodule model demonstrated stable discrimination for 1-year lung cancer detection across geographically distinct screening cohorts. A prespecified combination of Sybil and MONAI-derived nodule features achieved high-sensitivity triage while correctly classifying a substantial proportion of noncancer baseline low-dose CT examinations as negative.