Kinematic analysis of left ventricular deformation in myocardial infarction using magnetic resonance cardiac tagging.

The Magnetic Resonance (MR) tagging technique provides detailed information about 2D motion in the plane of observation. Interpretation of this information as a reflection of the 3D motion of the entire cardiac wall is a major problem. In finite element models of the mechanics of the infarcted heart, an infarcted region causes motional asymmetry, extending far beyond the infarct boundary. Here we present a method to quantify such asymmetry in amplitude and orientation. For this purpose images of a short-axis cross-section of the ejecting left ventricle were acquired from 9 healthy volunteers and 5 patients with myocardial infarction. MR-tags were applied in a 5 mm grid at end-diastole. The tags were tracked by video-image analysis. Tag motion was fitted to a kinematic model of cardiac motion. For the volunteers and the patients the center of the cavity displaced by about the same amount (p = 0.11) during the ejection phase: 3.8 +/- 1.4 and 3.0 +/- 0.9 mm (mean +/- sd), respectively. Cross-sectional rotation and the decrease in cross-sectional area of the cavity were both greater in the volunteers than in the patients: 6.4 +/- 1.5 vs. 3.0 +/- 0.8 degrees (p < 0.001), and 945 +/- 71 vs. 700 +/- 176 mm2 (p = 0.02), respectively. In the patients, asymmetry of wall motion, as expressed by a sine wave dependency of contraction around the circumference, was significantly enlarged (p = 0.02). The proposed method of kinematic analysis can be used to assess cardiac deformation in humans. We expect that by analyzing images of more cross-sections simultaneously, the 3D location and the degree of infarction can be assessed efficiently.

Relation between torsion and cross-sectional area change in the human left ventricle.

During the ejection phase, motion of the left ventricular (LV) wall is such that all myocardial fibers shorten to the same extent. In a mathematical model of LV mechanisms it was found that this condition could be satisfied only if torsion around the long axis followed a unique function of the ratio of cavity volume to wall volume. When fiber shortening becomes non-uniform due to cardiac pathology, this pathology may be reflected in aberration of the torsional motion pattern. In the present study we investigated whether the predicted regular motion pattern could be found in nine healthy volunteers, using Magnetic Resonance Tagging. In two parallel short-axis cross-sections, displacement, rotation, and area ejection were derived from the motion of tags, attached non-invasively to the myocardium. Information from both sections was combined to determine area ejection, quantified as the change in the logarithm of the ratio of cavity area to wall area, and torsion, represented by the shear angle on the epicardium. Linear regression was applied to torsion as a function of area ejection. The slope thus found (-0.173 +/- 0.024 rad, mean +/- S.D.) was similar to the slope as predicted by the model of LV mechanics (-0.194 +/- 0.026 rad). In conclusion, the relation between area ejection and torsion could be assessed noninvasively in humans. In healthy volunteers, the relation was close to what was predicted by a mathematical model of LV mechanics, and also close to what was found earlier in experiments on animals.