Contrast agent has several limitations in clinical practice, and the diagnostic performance of non-enhanced MR for lesion characterization should be thoroughly exploited. Inspired by the work of cross-modal learning framework，we propose a deeply supervised cross-modal transfer learning method to remarkably improve the malignancy characterization of HCC in non-enhanced MR, in which the cross-modal relationship between the non-enhanced modal and contrast-enhanced modal is explicitly learned and subsequently transferred to another CNN model for improving the characterization performance of non-enhanced MR. The visualization method Grad-CAM is also applied to verify the effectiveness of the proposed cross-modal transfer learning model.
This abstract and the presentation materials are available to members only; a login is required.