Three-dimensional Principal Strain Patterns in Acute Myocardial Infarction

INTRODUCTION: Comprehensive characterization of myocardial viability continues to be of diagnostic interest to clinicians. Magnetic resonance (MR) tagging is considered the gold standard imaging tool for noninvasive assessment of functional viability in the form of mechanical strain tensors [1]. Principal strain orientations, which correspond to directions of maximal deformations (stretching, contraction, and elongation), can potentially provide a sensitive metric of functional viability. Recent reports have shown principal strain directions to change with ischemic injury [2]. However, these investigations were limited to the direction of maximal contraction, due to large effect of noise in calculation of other principal strain orientations. Prolonged examinations and the need for imaging orthogonal slices, susceptibility to misregistration and large amount of inaccuracy in calculation of angles, hindered the assessment of the clinical value of three-dimensional (3D) principal strain orientations and the adoption of this information in clinical practice. In this work, we developed an accelerated algorithm using zHARP tagging [3] and a generalized vector convolution to robustly measure 3D principal strain orientations and investigate their potential clinical value in the assessment of viability. Preliminary in-vivo results in a closed-chest porcine model of myocardial infarction (MI) confirm the potential of principal strain orientations to characterize abnormal function. THEORY: zHARP, a slice-following tagging imaging technique, provides 3D motion tracking of the imaged shortaxis slice with application of additional z-gradient in the slice-select direction. By forming rectilinear mesh of points on a stack of equally separated images, 3D Eulerian strain tensor is calculated for each spatial point. Eigensystem decomposition of the strain tensor yields three eigenvalues and three corresponding eigenvectors, representing the