Complete spatiotemporal quantification of cardiac motion in mice through multi-view magnetic resonance imaging and super-resolution reconstruction.

BACKGROUND

Structural indices of cardiac diseases estimated via cardiac magnetic resonance imaging (CMR) have shown promise as early-stage markers. Despite the growing popularity of CMR-based myocardial strain calculations, measures of complete spatiotemporal strains (i.e., three-dimensional strains over the cardiac cycle) remain elusive, especially in mice. The high metabolic rates and rapid cardiac motion affect high-resolution imaging, thus compromising strain accuracy. We hypothesize that a super-resolution reconstruction (SRR) framework that combines low-resolution scans at multiple orientations will enhance the reliability of complete spatiotemporal strains in mice.

METHODS

Multi-view cine CMR comprising short- and long-axis (SA and LA) fast low angle shot scans were obtained in a cohort of wild-type-mice (n = 5) and a diabetic mouse (n = 1). The "SRR in CMR" approach, consisting of tissue-class -specific scattered data interpolation, was used to generate full four-dimensional (4D) images of high spatial resolution. Image registration using the diffeomorphic demons algorithm was applied to quantify complete spatiotemporal motion in terms of 4D myocardial strains. The effects of SRR on CMR quality were verified in all mice through image metrics, namely, root mean squared error (MSE) and structural similarity index. Strain calculations were validated against an in silico heart model phantom through MSE analysis, followed by investigations of strain accuracy and reproducibility for all mice using MSE and coefficient of variation analyses. Results: SRR-derived strains were validated against a kinematic benchmark established through the in-silico heart model phantom. Improvements in global strain accuracy were confirmed in both in-plane (radial and circumferential) and through-plane (longitudinal) strains. Mouse-specific SRR provided near isotropic spatial resolution, high structural similarity, and minimal loss of contrast, which led to overall improvements in strain reproducibility and intra-cohort homogeneity in wild-type mice, with global longitudinal strain lying of ≈-14%.

CONCLUSIONS

A comprehensive methodology was presented to quantify complete and reproducible myocardial deformation, aiding in the much-needed standardization of complete spatiotemporal strain analysis in small animals.