An Ensembled Deep Learning Model Outperforms Human Experts in Diagnosing Biliary Atresia from Sonographic Gallbladder Images

AbstractIt is still difficult to make accurate diagnosis of biliary atresia (BA) by sonographic gallbladder images particularly in rural area lacking relevant expertise. To provide an artificial intelligence solution to help diagnose BA based on sonographic gallbladder images, an ensembled deep learning model was developed based on a small set of sonographic images. The model yielded a patient-level sensitivity 93.1% and specificity 93.9% (with AUROC 0.956) on the multi-center external validation dataset, superior to that of human experts. With the help of the model, the performance of human experts with various levels would be improved further. Moreover, the diagnosis based on smartphone photos of sonographic gallbladder images through a smartphone app and based on video sequences by the model still yielded expert-level performance. Our study provides a deep learning solution to help radiologists improve BA diagnosis in various clinical application scenarios, particularly in rural and undeveloped regions with limited expertise.