Feature Tracking in Cardiac Magnetic Resonance Imaging

Purpose: An accurate and practical method to measure parameters like strain in myocardial tissue is of great clinical value, since it has been shown that strain is a more sensitive and earlier marker for contractile dysfunction than ejection fraction (EF). Current CMR technologies for strain assessment are time consuming and difficult to implement in daily clinical practice. Feature tracking is a technology that can lead to more automatization and robustness of quantitative analysis of medical images with less time consumption than comparable methods. Methods: Feature tracking (FT) technique was tested in computed-generated phantom models with varying image resolution, frame rate, and tissue dynamics. The phantom motion was known exactly and allowed a careful check of the intrinsic accuracy. The same technique was also tested in 7 patients undergoing clinical MRI: 2 normal, 1 non-ischemic, 2 ischemic cardiomyopathy, 2 hypertrophic cardiomyopathy. Analysis of clinical CMR involved calculation of ejection fraction (EF) as well as radial, circumferential and longitudinal strain Results: The method is well suited for tracking left ventricular myocardium in radial, circumferential as well as longitudinal direction. In addition, strain can be assessed at epicardial, endocardial and mid myocardial levels 3/19 Conclusions: This new method offers a robust and time saving procedure to quantify myocardial tissue displacement, velocity and deformation parameters on regular SSFP sequences of CMR imaging. It therefore can be implemented easily into clinical practice.