Recently, Prof. Zhangsheng Yu from the School of Life Sciences and Biotechnology (SLSB), Shanghai Jiao Tong University, published a paper in Journal of Hematology & Oncology in collaboration with researchers from Xinhua Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Renji Hospital Affiliated to Shanghai Jiao Tong University School of Medicine and Hangzhou First People’s Hospital Affiliated to Zhejiang University School of Medicine, titled Deep learning for differential diagnosis of malignant hepatic tumors based on multi-phase contrast-enhanced CT and clinical data. In this paper, the researchers built an automatic diagnostic model for differentiating malignant hepatic tumors based on patients’ multimodal medical data including multi-phase contrast-enhanced CT and clinical features.

Fig. 1 The architecture of the automatic diagnostic model

Hepatic malignancies, including hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC) and metastatic liver cancer, are common malignant tumors and have poor prognosis. The treatment regimen for the different subtypes of hepatic tumors is all distinct, and multi-phase CECT has become the primary tool for diagnosis of hepatic tumors before surgery. However, the differential diagnosis of malignant hepatic tumors is challenging, and misdiagnosis prior to surgery can mislead the treatment decision. An automated diagnostic model is desirable to be developed, which can assist doctors in hepatic tumors diagnosis, reduce observer variations and improve diagnostic efficiency.

In this study, the researchers proposed a deep learning-based model, which take the advantage of deep convolutional neural network (CNN) and gated recurrent neural network (RNN) to effectively extract and integrate the diagnosis-related radiological and clinical features of patients. The model achieved an accuracy of 72.6% on the test set for differential diagnosis of malignant hepatic tumors, comparable with the diagnostic level of doctors’ consensus (70.8%). With the assistance of the diagnostic model, doctors achieved better performance than doctors’ consensus diagnosis, with an increase of 8.3% in accuracy and 26.9% in sensitivity for ICC diagnosis on average. On the external test set from center 2, the diagnostic model achieved an accuracy of 82.9%, which verify the model’s generalization ability. The model established by Prof. Yu’s Lab can assist doctors to achieve better diagnostic performance, which is expected to serve as an AI assistance system and promote the precise treatment of liver cancer.

This work is collaborated with Prof. Jinyang Gu at Xinhua Hospital, Prof. Yingbin Liu at Renji Hospital, and Prof. Xiao Xu at Hangzhou First People’s Hospital. Two PhD students in the School of Life Sciences and Biotechnology at SJTU, Ruitian Gao and Ting Wei, are the first author and co-first author of this paper, respectively. The authors acknowledge the grants from National Natural Science Foundation of China, Shanghai Science and Technology Development Fund, and Medical Engineering Cross Fund of Shanghai Jiao Tong University.

Link: https://jhoonline.biomedcentral.com/articles/10.1186/s13045-021-01167-2