Title :
Myocardial kinematics from tagged MRI based on a 4-D B-spline model
Author :
Tustison, Nicholas J. ; Dávila-Román, Victor G. ; Amini, Amir A.
Author_Institution :
Cardiovascular Image Anal. Lab., Washington Univ., St. Louis, MO, USA
Abstract :
Current research investigating the modeling of left ventricular dynamics for accurate clinical assessment of cardiac function is extensive. Magnetic resonance (MR) tagging is a functional imaging method which allows for encoding of a grid of signal voids on cardiac MR images, providing a mechanism for noninvasive measurement of intramural tissue deformations, in vivo. We present a novel technique of employing a four-dimensional (4-D) B-spline model which permits concurrent determination of myocardial beads and myocardial strains. The method entails fitting the knot planes of the 4-D B-spline model for fixed times to a sequence of triplets of orthogonal sets of tag surfaces for all imaged volumetric frames within the constraints of the model´s spatio-temporal internal energy. From a three-dimensional (3-D) displacement field, the corresponding long and short-axis Lagrangian normal, shear, and principal strain maps are produced. As an important byproduct, the points defined by the 3-D intersections of the triplets of orthogonal tag planes, which we refer to as myocardial beads, can easily be determined by our model. Displaying the beads as a movie loop allows for the visualization of the nonrigid movement of the left ventricle in 3-D.
Keywords :
biomechanics; biomedical MRI; cardiology; image coding; kinematics; muscle; physiological models; splines (mathematics); 3-D intersections; 4-D B-spline model; functional imaging method; left ventricular dynamics modeling; magnetic resonance imaging; medical diagnostic imaging; movie loop; myocardial beads; myocardial kinematics; nonrigid movement visualization; orthogonal tag planes triplets; signal voids grid encoding; spatio-temporal internal energy; tagged MRI; Capacitive sensors; Image coding; Kinematics; Magnetic resonance; Magnetic resonance imaging; Myocardium; Noninvasive treatment; Spline; Surface fitting; Tagging; Algorithms; Heart Ventricles; Humans; Image Enhancement; Imaging, Three-Dimensional; Magnetic Resonance Imaging; Motion; Movement; Pattern Recognition, Automated; Stress, Mechanical; Ventricular Function, Left;
Journal_Title :
Biomedical Engineering, IEEE Transactions on
DOI :
10.1109/TBME.2003.814530