Comparison of left ventricular function evaluation using gated myocardial perfusion imaging with different software of quantitative analysis.
- Author:
Yu ZENG
;
Lisha JIANG
;
Lina ZHOU
;
Xiaochuan YANG
;
Luyi ZHOU
- Publication Type:Journal Article
- MeSH:
Aged;
Algorithms;
China;
Female;
Humans;
Male;
Middle Aged;
Myocardial Perfusion Imaging;
Regression Analysis;
Software;
Tomography, Emission-Computed, Single-Photon;
Ventricular Function, Left
- From:
Journal of Biomedical Engineering
2013;30(6):1239-1243
- CountryChina
- Language:Chinese
-
Abstract:
In the present study, the accuracy of the 4-dimensional model single photon emission computed tomography (4D-MSPECT) and quantitative gated single photon emission computed tomography (QGS) was investigated for assessing left ventricular end-diastolic (EDV), end-systolic volume (ESV) and ejection fraction (EF) from gated myocardial perfusion imaging (G-MPI) using left ventriculography (LVG) as reference. From December 2008 to June 2011, 85 patients, who underwent rest G-MPI and LVG (within 30 days) in West China Hospital, Sichuan University, were retrospectively recruited. EDV, ESV, and EF were calculated from G-MPI using 4D-MSPECT and QGS. Eighty-five patients (47 men, 38 women; age 57 +/- 13 years) were finally analyzed. Correlation between results of G-MPI and LVG was high for EDV, r = 0.89 (4D-MSPECT), r = 0.81 (QGS); ESV, r = 0.97 (4D-MSPECT), R = 0.95(QGS); EF, r = 0.95 (4D-MSPECT), r = 0.93 (QGS). 4D-MSPECT and QGS underestimated EDV significantly compared with LVG [(125 +/- 20) mL (4D-MSPECT), (118 +/- 39) mL (QGS), (131 +/- 33)mL (LVG)]. The ESV, 4D-MSPECT and QGS values did not differ significantly from LVG [(47 +/- 32) mL (4D-MSPECT), (53 +/- 29) mL (QGS), (49 +/- 37) mI (LVG)]. For LVEF, only QGS yielded values were significantly lower than LVG [61% +/- 21% (4D-MSPECT), 55% +/- 17% (QGS), and 63% +/- 19% (LVG)]. EDV, ESV, and EF as determined by 4D-MSPECT and QGS from G-MPI agree well with relevant values with LVG. However, Algorithm-inherent also showed slightly over- or under-estimation of volumes. Therefore, separated normal databases should be set up for each algorithm.
