Quick Fire Session
SCMR 22nd Annual Scientific Sessions
Christopher Nguyen, PhD
Instructor
Massachusetts General Hospital / Harvard Medical School
Robin Etzel, MSc
Research Fellow
Mittelhessen University of Applied Science
Timothy Reese, PhD
Assistant Professor
Massachusetts General Hospital / Harvard Medical School
Choukri Mekkaoui, PhD
Assistant Professor
Harvard Medical School - Massachusetts General Hospital
Boris Keil, PhD
Assistant Professor
Mittelhessen University of Applied Science
David Sosnovik, MD
Associate Professor of Medicine
Harvard Medical School - Massachusetts General Hospital
Background:
Tissue from the septum is often acquired during myectomy and myocardial biopsy offering a unique opportunity for histological correlation with diffusion-tensor CMR (DT-CMR). However, the low spatial resolution of current in vivo DT-CMR techniques results in profound volume averaging at the microstructural scale limiting the accuracy of histological correlation. We hypothesized that the use of a tailored 64-element cardiac array coil in combination with a novel DT-CMR sequence, optimized for high-resolution imaging, would allow sub-millimeter in-plane imaging of the ventricular septum to be performed in vivo. Key elements of the sequence included a 2D Radiofrequency pulse to excite a reduced FOV (rFOV), and a second order motion compensated (M2) spin-echo echoplanar DT-CMR acquisition.
Methods:
Six healthy individuals were imaged on a clinical 3T scanner (Siemens Healthcare, Prisma) with a free breathing rFOV M2 DT-CMR acquisition (TR = 12RR, TE=72ms, b = 50, 500 s/mm2, 12 dir, 12 NEX) centered on the septum at submillimeter resolution (0.85x0.85x8mm3) using the custom 64 element coil and a product 32 channel thoracic coil. An additional rFOV M2 DT-CMR acquisition was acquired with the 64-element array at conventional resolution (2.7x 2.7x8mm3). Noise scans (NEX = 20) were performed immediately after each DT-CMR scan by nulling the RF voltage. Mean diffusivity (MD), fractional anisotropy (FA), helix angle (HA) and SNR in the septum were analyzed using a Wilcoxon rank test and intra-class correlation. All images and maps were up-sampled to match the resolution of the submillimeter scans.
Results:
Submillimeter DT-CMR yielded a significant increase in the transmural pixel density compared with reference (16 +/- 4 pixels vs 5 +/- 2 pixels) (Figure 1). Submm-64 yielded significantly higher SNR than submm-32 (15 +/- 4 vs 8 +/- 2). No significant differences were found for MD (1.5 +/- 0.3 um2/ms) and FA (0.27+/- 0.08) between submm-64 and submm-32 (MD: 1.4+/- 0.2 s/mm2 FA: 0.24+/- 0.02) compared with reference (MD: 1.3+/- 0.2 s/mm2 FA: 0.25+/- 0.04) (Figure 2). Submm-64 transmural HA was smoothly continuous and monotonic from epicardium to endocardium for all subjects and revealed a significantly increased sigmoidal (f(x)=120°/(1+exp(-k*(2x-1))-60°) slope (k = 3.4 +/- 0.3) compared with reference (k = 1.8 +/- 0.2). Submm-32 yielded aberrant transmurality for all subjects (k = 1.2 +/- 0.8).
Conclusion:
DT-CMR of the human septum can be performed in vivo with sub-millimeter in plane resolution using a free breathing reduced-FOV M2-compensated DT-CMR sequence and a custom 64-channel cardiac coil. The increase in SNR with the 64-channel coil supported submillimeter resolution, yielding a 3-fold increase in transmural pixel density, and revealing the expected sigmoidal transmurality of HA seen in previous histological validation studies . This has important implications for correlating DT-CMR with histology in cardiomyopathies that rely on biopsy and myectomy.