In vivo fiber tractography of the right and left ventricles using diffusion tensor MRI of the entire human heart
Journal of Cardiovascular Magnetic Resonance volume 16, Article number: P17 (2014)
Diffusion Tensor MRI (DTI) tractography of the human heart in vivo has previously been performed with either large slice gaps or limited coverage [1, 2]. The aim of this study was to investigate the feasibility of performing DTI of the entire human heart in vivo without slice gaps. This, we hypothesized, would enhance the characterization of fiber architecture in the left ventricle (LV), allow myofiber organization in the right ventricle (RV) to be characterized in vivo, and further elucidate microstructural differences in the heart between systole and diastole.
DTI was performed on a clinical 3T scanner (Skyra, Siemens) using a fat-suppressed, zone-selected, diffusion-encoded stimulated echo sequence with 10 diffusion encoding directions, TE/TR 33/80 ms, GRAPPA rate 2, b-value 500 s/mm2, resolution 2.8 × 2.8 × 8 mm3, 8 averages and multiple breath-holds. The entire LV and RV were covered in 13 contiguous short-axis slices. Images were acquired in the systolic and diastolic sweet spots  of the cardiac cycle and were spatiotemporally coregistered . Tractography was performed by numerically integrating the primary eigenvector field into streamlines using an adaptive 5th order Runge-Kutta method. The impact of cumulative image averages (1-8) on the reliability of the fractional anisotropy (FA) and fiber helix angle (HA) indices was assessed.
A composite view of the anterior thorax and heart is shown in Figure 1A. The contrast between the helical pattern of the fibers in the heart (color-coded by HA) and the linear organization of skeletal muscle fibers in the chest wall can be seen. The RV consisted of a bilayer of obliquely-oriented fibers, lacking circumferential fibers ( Figure 1B). In contrast, fibers in the midwall of the LV were circumferential, forming a distinct band between the subendocardial and subepicardial fibers ( Figure 1C). Figure 1D shows the tractogram of the entire heart, depicting the intertwined arrangement of myofibers at the anteroseptal RV-LV junction. Figure 2A and 2B show that FA values in both sweet spots were significantly overestimated when the number of averages used was < 5 (p < 0.01). In contrast, convergence to a stable FA value (LV, diastole: 0.42 ± 0.03, systole: 0.41 ± 0.03; RV, diastole: 0.32 ± 0.11, systole: 0.36 ± 0.02) was observed from 5-8 averages. The transmural gradient in HA exhibited a similar pattern of convergence but required more averages ( Figure 2C, D).
DTI-tractography of the entire human heart can be performed in vivo, without slice gaps, and allows both the LV and the RV to be characterized. A minimum of 5 averages at each slice location is required for accurate quantification. The ability to characterize fiber architecture in the LV and RV in vivo has the potential to provide new insights into a range of diseases affecting both the pulmonary venous and arterial circulations.
R01 HL093038 (D.E.S.), R01HL112831 (D.E.S.), P41RR14075 (Martinos Center).
Toussaint, et al: MEDIA. 2013
Nielles-Vallespin, et al: MRM. 2012
Tseng, et al: MRM. 1999
Mekkaoui, et al: JCMR. 2012
About this article
Cite this article
Mekkaoui, C., Reese, T.G., Jackowski, M.P. et al. In vivo fiber tractography of the right and left ventricles using diffusion tensor MRI of the entire human heart. J Cardiovasc Magn Reson 16 (Suppl 1), P17 (2014). https://doi.org/10.1186/1532-429X-16-S1-P17
- Left Ventricle
- Fractional Anisotropy
- Right Ventricle
- Helix Angle
- Sweet Spot