An MRI-based automated myocardium boundary detection technique using Displacement Encoding with Stimulated Echoes (DENSE) images

摘要

Displacement Encoding with Stimulated Echoes (DENSE) is a Magnetic Resonance (MR) elastography technique developed to encode phase information from myocardial tissue displacement, which occur during the mixing period of the sequence and is stored in the longitudinal magnetization direction for prolonged periods. While there exists a number of semi-automated algorithms for processing phase data in complex DENSE images, with displacements retrieved for Left Ventricular (LV) strain analysis, there remains a need for fast, sophisticated techniques that address myocardial segmentation in DENSE images. This study proposes a novel, automated technique for detecting myocardial boundaries, using the DENSE phase data and image quantization via thresholding, which can significantly expedite the strain computation process.