Power Pitch 1: Whole-heart MR microscopy with ultrahigh gradient strength clinical MRI scanner

Despite centuries of research, questions remain about the link between microstructural cardiomyocyte arrangement and dynamics and cardiac function in the human heart. Clinical cardiac MR (CMR) diagnoses cardiovascular disease at whole-heart scale (1-100mm) but remains inadequate for the assessment of underlying cardiac microstructure. In contrast, optical microscopy depicts cardiac microstructure at a cellular level (10-100 𝜇m) but is limited to small samples of tissue so lacks the volumetric coverage to study the complex three-dimensional microstructure of the heart.

Ex vivo CMR can bridge this gap [1], probing the cardiac microstructure at the mesoscale level (100 𝜇m to 1mm) through high resolution anatomical imaging and diffusion tensor imaging (DT-CMR) in high performance systems. In this abstract we introduce an approach for whole-heart CMR microscopy of large (human-scale) mammalian samples in an ultrahigh gradient strength clinical MRI scanner. 





Methods:

A fixed healthy porcine heart (6 weeks immersed in 10% formalin) was scanned at 3T (MAGNETOM Cima.X, Siemens Healthineers AG, Forchheim, Germany, gradient strength 200 mT/m at 200 T/m/s slew rate, 18 channel Tx/Rx knee coil) using a 3D turbo spin echo sequence (TSE) and 3D gradient echo sequence (GRE) with 160𝜇m isotropic acquired resolution reconstructed to 80x80x160𝜇m. DT-CMR images were acquired with a multislice monopolar spin echo EPI sequence, 1.5mm isotropic resolution reconstructed to 0.8x0.8x1.5mm, b=0,150,750 s/mm², 30 directions, 40 averages. All images were acquired in a short axis view.

To assess the ability of these images to depict the known cardiac microstructure, DT-CMR images were analyzed using conventional diffusion tensor fitting, while anatomical TSE and GRE images were analyzed using a structure tensor approach [2]. For both diffusion tensor and structure tensor, primary and secondary eigenvectors were projected onto the local ventricular wall to obtain helix angle (HA) and absolute secondary eigenvector angle (|E2A|), which reflect cardiomyocytes primary and secondary organization [3].





Results:

Fig 1 shows example orthogonal views from TSE and GRE images. TSE images visually exhibit higher contrast resulting in a more detailed depiction of the heart microstructure despite being acquired with the same spatial resolution as GRE images. A visual comparison of tensor maps obtained from DT-CMR and high-resolution TSE and GRE images can be seen in Fig 2, showing that TSE and DT-CMR result in a comparable depiction of HA and |E2A|, albeit at higher resolution.

Finally, example images showing the smooth transmural rotation of the cardiomyocyte orientation from the from endocardium to epicardium (Fig 3) demonstrate the potential of high-resolution TSE imaging to provide a whole-heart non-destructive microscopy-like depiction of cardiac microstructure.



Conclusion:

High-resolution 3D TSE data from ultrahigh gradient strength clinical MRI scanners delivers non-destructive whole-heart imaging at mesoscopic resolutions, with each voxel containing potentially a few 10s of cells. These advances pave the way for non-destructive whole-heart MR microscopy in large hearts.