Why the study?
Most cardiac MRI strain estimation methods either rely on sparse annotated frame pairs—reducing usable data—or train without segmentation supervision, and rarely use more than two phases.
A novel semi-supervised deep learning algorithm using distance maps improves motion flow and myocardial strain estimation in cardiac MRI compared to existing methods.
Enhances precision of myocardial strain quantification via cardiac MRI; extends semi-supervised learning to reduce labeling needs in cardiovascular imaging.
Myocardial strain plays a crucial role in diagnosing heart failure and myocardial infarction. Its computation relies on assessing heart muscle motion throughout the cardiac cycle. This assessment can be performed by following key points on each frame of a cine Magnetic Resonance Imaging (MRI) sequence. The use of segmentation labels yields more accurate motion estimation near heart muscle boundaries. However, since few frames in a cardiac sequence usually have segmentation labels, most methods either rely on annotated pairs of frames/volumes, greatly reducing available data, or use all frames of the cardiac cycle without segmentation supervision. Moreover, these techniques rarely utilize more than two phases during training. In this work, a new semi-supervised motion estimation algorithm using all frames of the cardiac sequence is presented. The distance map generated from the end-diastolic segmentation label is used to weight loss functions. The method is tested on an in-house dataset containing 271 patients. Several deep learning image registration and tracking algorithms were retrained on our dataset and compared to our approach. The proposed approach achieves an average End Point Error (EPE) of 1 . 02 mm , against 1 . 19 mm for RAFT (Recurrent All-Pairs Field Transforms). Using the end-diastolic distance map further improves this metric to 0 . 95 mm compared to 0.91 for the fully supervised version. Correlations in systolic peak were 0.83 and 0.90 for the left ventricular global radial and circumferential strain respectively, and 0.91 for the right ventricular circumferential strain. • Iterative motion aggregation algorithm with memory encoder. • Distance maps used to weight loss functions. • Results reported on a multi-vendor/multi-center dataset of 271 patients. • Unsupervised, semi-supervised and fully supervised versions compared. • Superior performance to RAFT and Voxelmorph on point tracking and strain estimation.
No takes yet. Share an insight, caveat, or question.
Portal et al. (2025) studied this question.
Synapse has enriched 2 closely related papers on similar clinical questions. Consider them for comparative context: