Biophysical Models of the Cardiovascular System - Systems Biology: Networks, Models, and Applications - page 281

4. Reconstruction time is greatly reduced (tens of hours versus
weeks to months) relative to that for histological methods.

Preparation of hearts for DTMR imaging is relatively straightfor-
ward, and proceeds as follows. Hearts are first excised and fixed in a
3% formalin mixture. Following fixation, excised hearts are placed in
an acrylic container filled with Fomblin, a perfluoropolyether used to
increase contrast and eliminate unwanted susceptibility effects near
boundaries of the heart. Images presented here were acquired with a
3-D Fast Spin Echo Diffusion Tensor (3D FSE DT) sequence on a 1.5 T
GE CV/I MRI Scanner (GE, Medical System, Wausheka, WI) using
an enhanced gradient system with 40 mT/m maximum gradient
amplitude and a 150 T/m/s slew rate. Depending on heart size, the MR
parameters were varied to minimize the number of slices acquired.
Generally, the FOV was 9 cm with an image size of 256 x 256 yielding
an in-plane resolution of ~350 µ m. The volume was imaged with

130-140 slices at ~800 m thickness. Diffusion gradients were applied

in sixteen noncollinear directions with a maximum b value of
900 s/mm 2 . With an echo train length of 2 and a TR of ~700 ms, the total
scan time was ~64 hours.
Recent advances in our lab have enabled automation of DTMRI data
acquisition and analysis for ventricular reconstruction. Briefly, once
image data are acquired, software written in the MatLab programming
language is used to estimate epicardial and endocardial boundaries in
each short-axis section of the image volume using either the method
of region growing or the method of parametric active contours [84].
Diffusion tensor eigenvalues and eigenvectors are computed from the
DTMRI data sets at those image voxels corresponding to myocardial
points, and fiber orientation at each image voxel is computed as the
primary eigenvector of the diffusion tensor.
Representative results from DTMR imaging are shown in figure 9.7.
Figure 9.7a shows a finite-element model (see following section)
describing the epicardial surface of the canine ventricles derived from
DTMRI data. The line segments on the surface of this heart show
epicardial fiber orientation measured using DTMRI. Figure 9.7b shows
the volume-rendered DTRMI data, with voxels color-coded according
to fiber inclination angle. Inclination angle (also known as fiber or
helix angle) is known to vary continuously as a function of transmural
depth. This variation is shown in figure 9.7b. Figure 9.7c shows data
similar to that of figure 9.7b, the difference being that a single short-
axis section showing fiber inclination angle estimated using the
DTMRI data is rendered with an MR intensity image of this same
heart. Inclination angle is again color-coded. In addition, diffusion
tensors are visualized using a family of objects called glyphs [85]. This
visualization technique facilitates voxel-based display of tensor data.

µ