W M Schaefer1, H-J Kaiser, H Kuehl, K-C Koch, B Nowak, U Buell. 1. Department of Nuclear Medicine, University Hospital, University of Technology, Pauwelsstrasse 30, 52074 Aachen, Germany. wschaefer@ukaachen.de
Abstract
AIM: Using 8-frames/cardiac cycle with gated SPECT underestimates end-diastolic volumes (EDV) and ejection fractions (LVEF), and overestimates end-systolic volumes (ESV). However, using 16-frames/cardiac cycle significantly decreases the signal-to-noise-ratio. We analyzed 16-frames and rebinned 8-frame gated SPECT data using common 4D-MSPECT and QGS algorithms. PATIENTS, METHODS: 120 patients were examined using gated SPECT on a Siemens Multispect 3 (triple-head gamma camera) 60 minutes after intravenous administration at rest of about 450 MBq (two-day protocol) or about 750 MBq (one-day protocol) (99m)Tc-tetrofosmin. Reoriented short axis slices (16-frames) were summed framewise (1+2,3+4, etc.) yielding 8-frame data sets. EDV, ESV and LVEF were calculated for both data sets using 4D-MSPECT and QGS. RESULTS: QGS succeeded with 119, 4D-MSPECT with 117 patients. For the remaining 116 patients, higher EDV (+0.8ml/+3.8 ml) and LVEF (+1.5%/+2.6%; absolute) and lower ESV (-1.7ml/-0.9 ml) (4D-MSPECT/QGS) were found for 16-frame runs. Bland-Altman limits were smaller for QGS than 4D-MSPECT [EDV 32/12 ml, ESV 21/10 ml, LVEF 17/7% (4D-MSPECT/QGS)]. CONCLUSION: Both algorithms showed the expected effects. Contour finding using QGS failed with only one data set, whereas contour finding using 4D-MSPECT failed with three data sets. Since the effects observed between the 8- and the 16-frame studies are relatively small and quite predictable, 8-frame studies can be employed in clinical routine with hardly any loss at all, plus contour finding appears less susceptible to error.
AIM: Using 8-frames/cardiac cycle with gated SPECT underestimates end-diastolic volumes (EDV) and ejection fractions (LVEF), and overestimates end-systolic volumes (ESV). However, using 16-frames/cardiac cycle significantly decreases the signal-to-noise-ratio. We analyzed 16-frames and rebinned 8-frame gated SPECT data using common 4D-MSPECT and QGS algorithms. PATIENTS, METHODS: 120 patients were examined using gated SPECT on a Siemens Multispect 3 (triple-head gamma camera) 60 minutes after intravenous administration at rest of about 450 MBq (two-day protocol) or about 750 MBq (one-day protocol) (99m)Tc-tetrofosmin. Reoriented short axis slices (16-frames) were summed framewise (1+2,3+4, etc.) yielding 8-frame data sets. EDV, ESV and LVEF were calculated for both data sets using 4D-MSPECT and QGS. RESULTS: QGS succeeded with 119, 4D-MSPECT with 117 patients. For the remaining 116 patients, higher EDV (+0.8ml/+3.8 ml) and LVEF (+1.5%/+2.6%; absolute) and lower ESV (-1.7ml/-0.9 ml) (4D-MSPECT/QGS) were found for 16-frame runs. Bland-Altman limits were smaller for QGS than 4D-MSPECT [EDV 32/12 ml, ESV 21/10 ml, LVEF 17/7% (4D-MSPECT/QGS)]. CONCLUSION: Both algorithms showed the expected effects. Contour finding using QGS failed with only one data set, whereas contour finding using 4D-MSPECT failed with three data sets. Since the effects observed between the 8- and the 16-frame studies are relatively small and quite predictable, 8-frame studies can be employed in clinical routine with hardly any loss at all, plus contour finding appears less susceptible to error.
Authors: Berlinda J van der Veen; Arthur J Scholte; Petra Dibbets-Schneider; Marcel P M Stokkel Journal: Eur J Nucl Med Mol Imaging Date: 2010-05-26 Impact factor: 9.236