[Assessment of global and regional left ventricular function with a 16-slice spiral-CT using two different software...

To determine global and regional left ventricular (LV) function from retrospectively gated multidetector row computed tomography (CT) by using two different semiautomated analysis tools and to correlate the results with those of magnetic resonance imaging (MRI).

Nineteen patients (5 females, 14 males, mean age 69 years) underwent 16-slice spiral-CT (MS-CT) with standard technique without administration of beta-blockers for a decrease in the cardiac rate. Ten series of images were reconstructed at every 10 % of the RR-interval. With commercially available software capable of semiautomated contour detection, end-diastolic and end-systolic LV volumes (EDV and ESV) were determined from short-axis multiplanar CT reformations (MPR). Axial images of the end-systolic and end-diastolic cardiac phase were transformed to 3D volumes (3D) to determine EDV and ESV by using a threshold-supported reconstruction algorithm dependent on the contrast enhancement of the left ventricle. Steady-state free-precession cine MR images were acquired in short-axis orientation on the same day in all but one patient. Regional wall motion was assessed qualitatively in 17 left ventricular segments and classified as normo-, hypo-, a- or dyskinetic. Bland-Altman analysis was performed to calculate limits of agreement and systematic errors between CT and MRI.

For MPR/3D, mean end-diastolic (144.4/142.8 mL +/- 67.5/67.1) and end-systolic (66.4/68.7 mL +/- 52.1/49.9) LV volumes as determined with MS-CT correlated well with MRI measurements (147.6 mL +/- 67.6 [ r = 0.98/0.96] and 73.3 mL +/- 55.5 [ r = 0.98/0.98], respectively [ p <.001]). LV stroke volume (77.6/74.1 +/- 19.2/23.4 mL for CT vs. 74.4 mL +/- 18.4 for MRI, r = 0.92/0.74) and LV ejection fraction (58.6/55.9 % +/- 13.5/13.7 for CT vs. 55.6 % +/- 13.5 for MRI, r = 0.95/0.91) also showed good correlation (p <.001). Regional wall motion analysis revealed agreement between CT and MRI in 316/323 (97.8 %) myocardial segments.

Semiautomated analysis of 16-detector row CT data sets enables global and regional volumetric and functional analysis. The CT results correlate well with MRI findings for short axis MPR and for 3D volume reconstructions, with a higher statistical spread for the 3D method. The underestimation of end-systolic and end-diastolic volumes with CT may be caused by partial volume averaging due to the lower temporal resolution as compared with MRI.