Fig. 6From: Interpretable artificial intelligence-based app assists inexperienced radiologists in diagnosing biliary atresia from sonographic gallbladder imagesChanges in diagnostic confidence between initial diagnosis made by radiologists alone and final diagnosis made by radiologists with smartphone app’s assistance. a Changes in junior radiologists in the cohort with true label biliary atresia. b Changes in senior radiologists in the cohort with true label biliary atresia. c Changes in experienced pediatric radiologists in the cohort with true label biliary atresia. d Changes in junior radiologists in the cohort with true label non-biliary atresia. e Changes in senior radiologists in the cohort with true label non-biliary atresia. f Changes in experienced pediatric radiologists in the cohort with true label non-biliary atresia. * “Initial” represents independent diagnosis by radiologists, and “Final” represents the radiologist’s diagnosis with the assistance of the model. “1” to “4” represent “definitely non-BA,” “probably non-BA,” “probably BA,” and “definitely BA,” respectivelyBack to article page