Direct myocardial strain assessment from frequency estimation in tagging MRI

We propose a new method to analyse deformation of the cardiac left ventricular wall from tagging magnetic resonance images. The method exploits the fact that the time-dependent frequency covector field representing the tag pattern is tightly coupled to the myocardial deformation and not affected by tag fading. Deformation and strain tensor fields can be retrieved from local frequency estimates given at least n (independent) tagging sequences, where n denotes spatial dimension. Our method does not require knowledge of material motion or tag line extraction. We consider the conventional case of two tag directions, as well as the overdetermined case of four tag directions, which improves robustness. Additional scan time can be prevented by using one or two grid patterns consisting of multiple, simultaneously acquired tag directions. This concept is demonstrated on patient data. Tracking errors obtained for phantom data are smaller than those obtained by HARP, 0.32±0.14 px versus 0.53±0.07 px. Strain results for volunteers are compared with corresponding linearised strain fields derived from HARP