HARP is an image analysis technology that works hand-in-hand with tagged magnetic resonance (MR) data. Tagged MR images are enhanced with regular patterns that encode the motion of the underlying tissue; making it easier to visualize and quantitatively analyze the motion. Tagging is especially good for imaging repetitive muscle motion, such as the muscles that give rise to the contraction of the heart. Prior to the invention of HARP in the late 1990’s, tagged MR images were either visually observed or perhaps analyzed by hand in order to better understand the mechanics of cardiac motion—in other words, the function of the heart. However these were very time consuming and hindered a wider use of tagging.

Although great strides were made in medical science related to cardiac mechanics, routine clinical use of tagged MR data was impossible prior to HARP—due to the long processing times. The key discovery in HARP is that motion can be automatically and robustly decoded using the harmonic phases of the tagged images. By carrying out key processing steps in the Fourier transform domain—also called k-space within the magnetic resonance physics community, weak signals, tag fading, and even motion artifacts are largely irrelevant to the analysis. The harmonic phases can be decoded to recover the motion of any point within the heart muscle, which allows the program to automatically track the contraction of the heart and recover the mechanical strain taking place at each point in the heart, which allows the program to characterize the regional function of the heart.

Computation of the motion of several short-axis sections including the left ventricle of the heart is sufficient for the program to compute several standard measures of heart performance; for example ejection fraction. Computation of the strain within several short-axis sections is sufficient to identify regions of the heart muscle that are not performing well. Since MR tagging data are image sequences—that is, they are movies—this characterization can reveal both spatial as well as temporal anomalies. This makes HARP useful for finding ischemic or infarcted tissues as well as for dyssynchronous behavior that is indicative of electrical problems within the heart. Besides strain, the rotation and twisting of the left ventricle is also measured using HARP.

HARP is known to only characterize the two-dimensional motion and strain of the heart muscle—the so-called in-plane motion. Typically, HARP is most useful when coupled with short-axis images of the left ventricle; in this case the in-plane motions that are measured can be characterized as either radial (toward or away from the long axis of the left ventricle) or circumferential (around the ventricle) measures. If long-axis images are acquired, then HARP can yield both radial and longitudinal—from apex to base—measures. This restriction to two dimensions is not a limitation of HARP but of the conventional implementation of MR tagging on commercial MR scanners. With relatively simple modifications to conventional tagging pulse sequences, out-of-plane tags can be generated and through-plane motion and strain can be studied.