Diffusion MRI tractography of the human heart In Vivo at end-diastole and end-systole
Journal of Cardiovascular Magnetic Resonance volume 14, Article number: O49 (2012)
Diffusion Tensor MRI (DTI) of the human heart in vivo has to date been performed in 2D and at a single phase of the cardiac cycle. Here we perform 3D tractography of the human heart in vivo at both end diastole and end systole. We show that fiber orientation in the subepicardium becomes more oblique during systole, and that scalar indices of diffusion (mean diffusivity and fractional anisotropy) decrease during systole. Our data suggest that myocardial fiber architecture is dynamic and is a function of both chamber geometry and myocardial contraction.
Diffusion Tensor Imaging (DTI) of the human heart in vivo has been described [1, 2] but has only been performed in 2D and at a single phase of the cardiac cycle. The impact of myocardial contraction on 3D fiber architecture in vivo thus remains poorly defined. Here we use a recently developed diffusion-weighted stimulated echo single shot EPI sequence , to address this question with a 3D tractographic approach . The purpose of this study was thus to perform in vivo DTI of normal human hearts at end-diastole and end-systole and quantify changes in myofiber organization as the myocardium contracts and relaxes.
DTI of three normal volunteers was performed on a 3T scanner (Skyra, Siemens) with the following parameters: 6 diffusion-encoding directions, b=350s/mm2, TR/TE=1100/23ms, BW=2442Hz/pixel, spatial resolution=2.7x2.7x8mm3, 3 slices, 6-8 averages, multiple breathholds. Fiber tracts were constructed by integrating the primary eigenvector field from the dyadic tensor into streamlines using a 4th order Runge-Kutta approach. Myofiber tracts were color-coded by their median helix angle . Mean diffusivity (MD), fractional anisotropy (FA) and individual eigenvalue maps were averaged in 12 sectors of the anterior, lateral, inferior and septal walls of the left ventricle (LV).
Tractography showed that myofibers in the subepicardium of the LV assumed a more oblique orientation at end-systole versus end-diastole (Figure 1). MD and FA were significantly (p<0.05, Mann-Whitney) higher at end-diastole than end-systole (Figure 2). Likewise, all three eigenvalues were higher at end-diastole than end-systole (p<0.05), although the change in the principal eigenvalues (λ1) was greatest.
Here, for the first time, we perform DTI tractography of the human heart in vivo at two phases of the cardiac cycle. The decrease in MD and FA at end-systole is likely due to myocyte thickening and compression of the extracellular space during systole. The increase in fiber obliquity in the subepicardium in systole could play an important role in base-apex shortening. Further study will be needed to determine the relationship between myocardial strain and the diffusion tensor. However, our data suggest that myofiber architecture in the human heart is dynamic and is a function of both chamber geometry and LV contraction.
R01 HL093038 (Sosnovik), NCRR P41RR14075 (Martinos Center) and MGH-ECOR (Mekkaoui).
Wu, et al: Circulation. 2006
Gamper U, et al: MRM. 2007
Nielles-Vallespin S, et al: ISMRM. 2011
Mekkaoui C, et al: ISMRM. 2011
About this article
Cite this article
Mekkaoui, C., Nielles-Vallespin, S., Gatehouse, P.D. et al. Diffusion MRI tractography of the human heart In Vivo at end-diastole and end-systole. J Cardiovasc Magn Reson 14 (Suppl 1), O49 (2012). https://doi.org/10.1186/1532-429X-14-S1-O49