Three-dimensional models of individual cardiac histoanatomy: tools and challenges

Abstract

There is a need for, and utility in, the acquisition of data sets of cardiac histoanatomy, with the vision of reconstructing individual hearts on the basis of noninvasive imaging, such as MRI, enriched by reference to detailed atlases of serial histology obtained from representative samples. These data sets would be useful not only as a repository of knowledge regarding the specifics of cardiac histoanatomy, but could form the basis for generation of individualized high-resolution cardiac structure-function models. The current article presents a step in this general direction: it illustrates how whole-heart noninvasive imaging can be combined with whole-heart histology in an approach to achieve automated construction of histoanatomically detailed models of cardiac 3D structure and function at hitherto unprecedented resolution and accuracy (based on 26.4 x 26.4 x 24.4 microm MRI voxel size, and enriched by histological detail). It provides an overview of the tools used in this quest and outlines challenges posed by the approach in the light of applications that may benefit from the availability of such data and tools.

FIGURE 1

Histological appearance of left ventricular tissue (Trichrome stained, 5 × objective), revealing extensive detail on structures such as aortic valve (AO), myocardium (M), and fibrous tissue (F) adjacent to the annulus fibrosus (AF). Scale bar: 250 μm. (see online version for color image.)

FIGURE 2

(A) Mosaic montage image of a longitudinal cardiac section (Trichrome stained), consisting of ~250 individual frames. LA: left auricular appendage; LV: left ventricle; P: papillary muscle. Fiber orientation, cleavage planes, myocardial sheets (or laminae), coronary vasculature, and myocyte/nonmyocyte distribution can be clearly identified. (B) Schematic representation of regional fiber orientation extraction (see text for detail). (See online version for color image.)

FIGURE 3

MR image planes obtained from contractured (A), slack (B), and volume-loaded (C) hearts (scale bar: 5 mm). In (A) and (B), voxel size is 26.5 × 26.5 × 24.5 μm, and in (C) 32 × 32 × 44 μm (a larger bore NMR tube was required for the volume-loaded heart). RV: right ventricle; LV: left ventricle. A full movie of the data in Figure 3B, progressing through transversal MRI sections from aorta to apex, can be downloaded from: http://mef.physiol.ox.ac.uk/MRI/Aorta_to_apex.avi.

FIGURE 4

Longitudinal section of the data set in Figure 3B, reconstructed from multiple transverse 2D MR images showing structural detail such as aortic (AO) and right atrioventricular (AV) valves, free-running Purkinje network (PN), and ventricular cleavage planes (CP). LV: left ventricle, RA: right atrium.

FIGURE 5

Three-dimensional reconstruction of a rabbit heart from serial histological sections that were performed approximately along the mediolateral axis of the heart (i.e., roughly parallel to the ventricular septum, see outline in (A). Slices were registered using an elastic registration with maximum rigidity, and results are visualized in two planes that are perpendicular to the original slicing plane, before (B and D) and after (C and E) registration. (See online version for color image.)

FIGURE 6

Results of vascular segmentation from the data set shown in Figure 3B. (A) Inverted maximum intensity projection. (B) Volume rendering of segmented voxels representing the coronary tree. (C and D) Gray-level coded vessel radius distribution of coronary vasculature from 340 μm (dark) to 30 μm (light). (A and C) horizontal projections; (B and D) lateral projections.

FIGURE 7

Left-ventricular wedge with papillaries and trabeculae. (A) 3D visualization of segmented tissue sub-stack. (B) adaptive FEM mesh. (C) partially cut wedge exposing histology-based microstructural detail. (D) bidomain simulation of effects of fiber direction on electrical impulse propagation. Scale bar illustrates voltage distribution in mV. (See online version for color figure.)

FIGURE 8

Generation of individualized high-resolution computational meshes from combined MRI and histological image data. Segmentation of MRI data results in a contiguous 3D data set. Segmented histological images have to be registered, slice by slice, and transformed to create a 3D histological stack. Both 3D data sets are then fused together. Such a histoanatomical global data set will enable the detailed segmentation of myocardial fibers, interstitial clefts, vessels, and connective tissue. In the process of segmentation, each voxel is classified and indexed (for instance 0 = myocardium, 1 = connective tissue, 2 = vessel, etc.), to account for the different properties of cells and regions that comprise the computational mesh. On this basis, provided the registration of histological data is sufficiently accurate, fiber orientation can be computed even without resorting to DTMRI.