Assessment of left ventricular function and mass in dual-source computed tomography coronary angiography: Influence of beta-blockers on left ventricular function: Comparison to magnetic resonance imaging

Abstract

Purpose To quantify left ventricular (LV) function and mass (LVM) derived from dual-source computed tomography (DSCT) and the influence of beta-blocker administration compared to cardiac magnetic resonance imaging (CMR). Methods Thirty-two patients undergoing cardiac DSCT and CMR were included, where of fifteen received metoprolol intravenously before DSCT. LV parameters were calculated by the disc-summation method (DSM) and by a segmented region-growing algorithm (RGA). All data sets were analyzed by two blinded observers. Interobserver agreement was tested by the intraclass correlation coefficient. Results. 1. Using DSM LV parameters were not statistically different between DSCT and CMR in all patients (DSCT vs. CMR: EF 63 ± 8% vs. 64 ± 8%, p = 0.47; EDV 136 ± 36 ml vs. 138 ± 35 ml, p = 0.66; ESV 52 ± 21 ml vs. 52 ± 22 ml, p = 0.61; SV 83 ± 22 ml vs. 87 ± 19 ml, p = 0.22; CO 5.4 ± 0.9 l/min vs. 5.7 ± 1.2 l/min, p = 0.09, LVM 132 ± 33 g vs. 132 ± 33 g, p = 0.99). 2. In a subgroup of 15 patients beta-blockade prior to DSCT resulted in a lower ejection fraction (EF), stroke volume (SV), cardiac output (CO) and increase in end systolic volume (ESV) in DSCT (EF 59 ± 8% vs. 62 ± 9%; SV 73 ± 17 ml vs. 81 ± 15 ml; CO 5.7 ± 1.2 l/min vs. 5.0 ± 0.8 l/min; ESV 52 ± 27 ml vs. 57 ± 24 ml, all p < 0.05). 3. Analyzing the RGA parameters LV volumes were not significantly different compared to DSM, whereas LVM was higher using RGA (177 ± 31 g vs. 132 ± 33 g, p < 0.05). Interobserver agreement was excellent comparing DSM values with best agreement between RGA calculations. Conclusion Left ventricular volumes and mass can reliably be assessed by DSCT compared to CMR. However, beta-blocker administration leads to statistically significant reduced EF, SV and CO, whereas ESV significantly increases. DSCT RGA reliably analyzes LV function, whereas LVM is overestimated compared to DSM.