Three-dimensional segmentation of the left ventricle in late gadolinium enhanced MR images of chronic infarction combining long- and short-axis information

Segmentation of left ventricle in late gadolinium enhanced MRI through 2D‐4D registration for infarct localization in 3D patient‐specific left ventricular model

“…To date, most LV segmentation studies using LGE scans have made use of an ECG‐gated trigger time to identify the closest cine phase, from which the corresponding LV contours were projected/registered onto 2D LGE scans as a priori segmentation. Dikici et al, by contrast, interpolated the cine image from 2 nearby phases before nonrigidly registering it to the LGE image, after which the a priori segmentation contours from the interpolated cine scan was deformed to match the LV borders in the 2D environment of the LGE scan.…”

“…Other studies, on the other hand, used predelineated myocardial cine contours as a priori knowledge for LV segmentation, taking advantage of the high blood pool-myocardium contrast in cine images. 9,12,[15][16][17] These approaches are plausible because cine data is frequently analyzed and contoured for functional assessment. Furthermore, a handful of automatic 4D or 3D+time cine segmentation techniques, [18][19][20][21][22][23][24][25] previously validated against public challenge datasets, are available for use as part of the LGE LV segmentation framework meant for facilitating infarct delineation and quantification.…”

“…However, possible interslice shifts (i.e., out-of-plane shifts) 26 in these studies were not addressed, which may cause the segmented model to deviate from the true heart shape, leading to inaccurate estimation of LV function. 27 Ciofolo et al, 15 Wei et al, 12 and Tao et al, 17 by contrast, have developed more advanced 3D segmentation frameworks, deforming either the 2D cine contour and/or 3D cine mesh model to delineate the LV in LGE scans. The focus of these approaches was solely to segment the LV in LGE scans to facilitate quantification of infarct scarring.…”

Accuracy of respiratory motion measurement of 4D-MRI: A comparison between cine and sequential acquisition

“…25,27,[32][33][34][35][36] An in-house developed  program was built to facilitate the simulation and evaluation. Figure 3 illustrated the overall study design and workflow of the digital phantom simulation study.…”

“…To evaluate and compare the performance of the cine-mode and the sequential-mode 4D-MRI techniques, we conducted virtual experiments on a 4D digital extended cardiac-torso (XCAT) digital human phantom [29][30][31] by simulating stepby-step the image acquisition and reconstruction of the two 4D-MRI techniques. 25,27,[32][33][34][35][36] An in-house developed  program was built to facilitate the simulation and evaluation. Figure 3 illustrated the overall study design and workflow of the digital phantom simulation study.…”