• Solveig Tiepolt, Henryk Barthel, Daniel Butzke, Swen Hesse, Marianne Patt, Hermann-Josef Gertz, Cornelia Reininger, and Osama Sabri.
• Department of Nuclear Medicine, University of Leipzig, Liebigstr 18, 04103 Leipzig, Germany. solveig.tiepolt@medizin.uni-leipzig.de
• Eur. J. Nucl. Med. Mol. Imaging. 2013 Jan 1; 40 (2): 238-44.

PurposeFlorbetaben is a β-amyloid-targeted PET tracer with significant potential for augmenting the toolbox in the clinical diagnosis of Alzheimer's disease (AD). In dementia imaging, shortening of scan duration may simplify future clinical use. The aim of this retrospective study was to investigate the effect of scan duration on diagnostic accuracy.MethodsPET scans obtained from 25 AD patients and 25 healthy volunteers (HVs) were analysed. In each subject, scans of three different durations (5, 10 and 20 min; all starting 90 min after injection) were obtained, randomized, and visually assessed by three experts blinded to the subject's identity and group affiliation. Presence/absence of β-amyloid and diagnostic confidence (0-100 %) were scored, and 10 % of the scans were re-read. Further, randomly selected datasets of ten AD patients and ten HVs were quantified using an established VOI-based approach and using a voxel-based approach.ResultsThe sensitivity and specificity of the blinded read were 80 % and 96 %, respectively, for all scan durations. Diagnostic confidence was high (97 ± 6 %, 97 ± 6 % and 95 ± 8 % for the 20-min, 10-min and 5-min scans, respectively; n.s.), as was interreader agreement (kappa(20 min) = 0.94, kappa(10 min) = 0.94, kappa(5 min) = 0.89; n.s.). Intrareader agreement was highest for the 20-min scan (kappa = 1.00) and lower for the 10-min scan (kappa = 0.71) and 5-min scan (kappa = 0.80; p = 0.002 and 0.003 vs. the 20-min scan). For all scan durations, composite SUVRs (Cohen's d effect size 4.5, 3.9 and 4.8 for the 5-min, 10-min and 20-min scans; p < 0.0001 each) and individual brain volumes affected by β-amyloid (Cohen's d effect size 1.6, 1.8 and 2.0 for the 5-min, 10-min and 20-min scans; p < 0.005 each) were significantly higher in AD patients than in HVs.ConclusionReduction in scan duration did not relevantly affect the accuracy of florbetaben PET scans in discriminating between AD patients and HVs. Thus, a reduction in scan duration seems conceivable for the future clinical use of florbetaben.

33 keywords...

hide…