DOES ACQUISITION TIME AFFECT THE RESULTS OF 18F-FLORBETABEN IMAGING?

Abstract

Background: 18F-Florbetaben is a PET radioligand approved for determining binary amyloid status. An acquisition period of 90110 minutes post-injection provides high contrast as pseudo-equilibrium is achieved. In the US, clinical scans may be acquired for 15-20 minutes beginning anytime 45-130 minutes post injection. We sought to determine if acquisition time affects scan interpretation. Methods: Twenty-four subjects were studied (age1⁄4 66.5 6 8.85 years), including 21 patients meeting clinical criteria for amnestic mild cognitive impairment or mild Alzheimer’s disease (MMSE 24.46 4.7) and 3 cognitive controls. 18F-Florbetaben images were acquired 50-110 min post-injection in 5-min frames. Images were binned into 50-70 and 90-110 min scans and visually read as positive or negative by three trained readers blind to cognitive status and timewindow.We determined agreement among the 3 readers (inter-reader agreement) and agreement between 50-70 and 90-110 min scan reads (intra-reader agreement). Standardized uptake value ratios (SUVRs) were calculated, using cerebellar gray matter as reference, for both 50-70 and 90-110 min image data. Results: High agreement was found between majority read results and SUVR for both imaging windows. SUVRs from 50-70 min scans correlated with those from 90-110 scans (r 1⁄4 0.987, p<0.0001). SUVRs were significantly greater in visually positive than in visually negative scans for both time windows (p<0.005). SUVRswere still increasing at 50-70min but became near-constant after 90 min (Fig. 1). Inter-reader agreement was greater for 50-70 min than for 90-110 min scans (Fleiss Kappa1⁄4 0.808 vs. 0.547). Regarding intra-reader agreement, 6 of 8 discrepant cases between 50-70 and 90-110 min scans were generated by onereader, who had extensive prior experience reading 50-70 min time window. For 6 subjects, CSF or tau PET data were available and supported the positive 50-70 min read, consistent with the assessment of the other readers and SUVR findings. Conclusions: Majority read and SUVR data suggest that both time windows are suitable for scan assessment. However, if prior reading experience is mainly available for one imaging window, time window should be known to the reader, as kinetic features influence the image, and SUVR data could support interpretation.