Survival prediction of heart failure patients using motion-based analysis method.

•We develop a clinical tool based on cardiac motion information for the survival prediction of HF patients. This tool assists the clinicians on the individualized therapy of the HF patients.•We propose the hierarchical spatial-temporal structure and dense optical flow structure to improve the model discrimination and tracking capability.•We propose the multi-modality deep-cox structure to improve the survival probability of HF patients.•Extensive experiments on the cine MR images of 329 subjects show that our method achieves superior performance over other state-of-the-art methods.