In AD, PET imaging to detect Ab has a great potential for early and accurate diagnosis. Florbetaben is currently
under clinical development as a promising tracer candidate for this purpose. The aim of this multicenter, phase
2 trial was to determine the diagnostic efficacy of florbetaben in differentiating ADs from healthy controls (HCs).
In 18 centers, 150 subjects were recruited/imaged with florbetaben PET: 81 patients with probable AD (DSM-IVTR
and NINCDS-ADRDA criteria, age ≥55 yrs, MMSE=18-26, CDR=0.5-2) and 69 age-matched HCs (MMSE≥28,
CDR=0). The PET data were visually analyzed by 3 blinded readers. Further, semi-quantitative analysis was done by
adapting a modified AAL volume of interest (VOI) template and obtaining SUV-ratios (SUVRs, reference: cerebellar
cortex). To optimize this VOI method, grey matter was automatically segmented on the MRIs.
According to visual analysis, the 90-110 min p.i. PET data were 80% sensitive and 90% specific in discriminating
ADs from HCs. VOI analysis of the 90-110 min p.i. PET data revealed significantly (p<0.0001) higher SUVRs for
ADs versus HCs, in particular in frontal, laterotemporal, parietal, and cingulate cortices. Grey matter segmentation
led to 10-22% improved SUVR discrimination between AD and HCs (p<0.0001) for cortical regions compared
to the non-segmented approach. In AD patients, APOE4 alleles were found more frequently for PET-positives as
compared to PET-negatives (65 vs 22%, p=0.027). There were significant correlations of SUVRs to APOE4 status
(number of APOE4 alleles) in gyrus rectus, temporal and cingulated cortices for the AD group (p<0.05, all r≥0.3)
but not for HCs (all p>0.1).
Florbetaben brain PET is accurate in differentiating clinically diagnosed ADs from HCs. The correlations observed
between PET data and APOE4 genotypes confirm preclinical data showing that florbetaben binds to Ab. Further
development of florbetaben PET as a visual adjunct for improved AD diagnosis is encouraged.
This trial was supported by Bayer Healthcare.
