A combined deep learning and radiomics model predicts Ki-67 expression in HCC using MRI and multifrequency MR elastography
Xumei Hu1, Ruokun Li2, Jiahao Zhou2, Jing Guo3, Ingolf Sack3, Weibo Chen4, He Wang5, Fuhua Yan2, and Chengyan Wang1
1Human Phenome Institute,Fudan University, Shanghai, China, 2Department of Radiology, Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China, 3Department of Radiology, Charité–Universitätsmedizin Berlin, Berlin, Germany, 4Philips Healthcare, Shanghai, China, 5Institute of Science and Technology for Brain-inspired Intelligence, Fudan University, Shanghai, China
In this study, a combined deep learning and radiomics (DLR) approach using six different network architectures was tested and compared for the prediction of high Ki-67 expressions in patients with hepatocellular carcinoma (HCC). The model was based primarily on data from MRI and tomoelastography, a multifrequency MR elastography technique. Xception delivered the best performance and recognized seven prominent features among which four were obtained from tomoelastography. Our findings demonstrated that biomechanical properties, especially viscosity and the fluid behavior of the tumor, are crucial imaging features that are important for imaging-based cancer diagnostics.
This abstract and the presentation materials are available to members only;
a login is required.