Improved accuracy and consistency in measurement of flowing blood by using inversion recovery GE-EPI
RF pulse and EPI gradient pulse sequence with ECG gating for measurement of the of flowing fluid. The pulse sequence uses a nonselective adiabatic inversion pulse followed by a series of ECG-gated gradient EPI image acquisitions. One EPI acquisition is shown in the lower subregion. The two upper gray subregions show the relative order of the R-wave peak signal and the pulse sequence for the first image (left box) with the inversion and for the subsequent images (right box) without the inversion.
The setup diagram of the phantom study. The dark gray regions represent the tube with flowing water. The tubing is coiled within the magnet homogeneous region to ensure that the inversion pulse can invert sufficient fluid in the tubing. A big cylinder (gray region) with a hole at the center is used for RF coil signal loading.
Plot of the simulated magnetization after the inversion pulse as a function of distance along the tubing demonstrating increasing imperfect inversion for positions farther from the magnet isocenter (factor 3 described in the text). In the simulation, the longitudinal magnetization is assumed to change with distance to the imaging isocenter (located at ). The plotted symbols show the positions of the simulated data.
Plots of vs temperature measurements obtained from the phantom study: (I) measurements obtained using the CFN method; (II) linear fit to CFN measurements; (III) measurements obtained using the fitting method (IV); linear fit to fitting measurements.
measurements obtained from the human studies. The horizontal axis corresponds to the different regions in Fig. 6. (엯) measurements obtained with the fitting method; (×) measurements obtained with the CFN method; (⋆) measurements obtained from the null point defined by linear interpolation of the two signal measurements around the null point. The panels from 1 to 4 correspond to the four volunteers.
An ECG-gated EPI image from the first volunteer (Fig. 5, first panel). A, B, and C correspond to the ascending aorta, superior vena cava, and left branch of the pulmonary trunk, respectively.
The data and the fitting curve in the three different regions of volunteer 1. ’s of the fitting method for regions A, B, C, are 1308, 1533, . The comparable ’s of the CFN method are 1214, 1195, .
of the fitting and CFN method vs the residual percentage (the percentage of fluid that remains in the slice for the subsequent excitation), the true ’s are 900, 1200, , respectively.
Dependence of accuracy on magnetization change due to incomplete inversion (factor 3) or mixing with noninverted blood (factor 4). The upper plot shows five hypothetical cases where the fraction of noninverted blood increases with time due to factors 3 and 4 (described in the text). The lower plot shows the results of these different cases.
measurements obtained in the presence of different realizations of multiplicative (motion induced) noise. of the fitting and CFN method changes with different noise realizations. Real . , .
The difference between the true and the for the two different method (the CFN method and the fitting method) due to factor 3 (see the text).
results from the data of volunteers. Here three methods were used. I is the Fitting method, II is the CFN method, III is the linear interpolation method. A, B, C are the different regions of interest shown in Fig. 6.
Article metrics loading...
Full text loading...