Early phase amyloid PET versus FDG-PET in atypical dementia: the AMY-ITA multicenter study.

Background Amyloid PET and 18F-FDG PET scans are commonly used in patients with uncertain diagnosis of Alzheimer’s disease (AD). A few studies showed that early frames of amyloid PET correlate well with FDG PET images, providing perfusion-like information, thus being a potential surrogate for FDG-PET scan. Objective To investigate whether early florbetaben-PET (FBB-PET) imagies are comparable to those of FDG-PET in terms of regional uptake deficits in patients with a clinical suspicion of atypical AD. In addition, the change in clinical judgment by neurologists at different centers after performing FBB-PET was evaluated. Finally, through the DORIAN software, the concordance between the visual analysis of FBB- PET with a semiquantitative analysis was analyzed. Materials and Methods AMY-ITA is an ongoing multicenter prospective study conducted in 8 Italian centers. So far, 83 patients have been enrolled, collecting FDG-PET and FBB-PET images. Clinical data were collected by the neurologist, who estimated the suspicion of Alzheimer's disease for all patients prior to the AMY-PET result and then confirmed or did not confirm the initial diagnosis. CSF analysis was available for 17 patients. In visual analysis of ET images, the brain was divided into 8 different regions in both FDG and FBB-PET early-frames scans. Each region was analyzed visually and blinded in both PET scans, defining the tracer uptake abnormality, using a scale of 0 to 3. A statistical analysis was then performed using Spearman's and Wilcoxon's tests. DORIAN software was used, which through SUVr, ELBA, TDr and RANK methods produced a semiquantitative analysis. Results: The most frequent clinical variants of AD were early primary progressive aphasia, posterior cortical atrophy and young onset-AD. Visual analysis of the brain revealed similar patterns between FBB-PET and FDGPET images of early frames; however, abnormal uptake scores were higher in FDG-PET in all regions. Spearman's test showed a statistically significant correlation in all brain regions (ρ 0.733 to 0.893, P<0.001). 64% of patients tested positive on AMT-PET scans and the discordance rate between the initial suspicion of amyloid positivity and the FBB-PET result was 33%. In 61% of cases the neurologist rejected the initial diagnosis and changed the clinical management. FBB-PET findings was to confirm AD in case of uninformative CSF data. A 95.5% concordance was found between visual and semiquantitative analysis of Amy-PET data. Conclusions Visual analysis of early-phase FBB-PET acquisitions correlated well with FDG-PET images in atypical forms of AD, offering a surrogate marker of brain metabolism. As a consequence, FBB-PET with analysis of early frames may convey added information on metabolism, reducing patient radiation exposure and health costs by avoiding FDG-PET. FBB-PET is a confirmed valid biomarker also in detecting amyloid deposition in atypical AD variants. DORIAN semiquantitative analysis confirmed visual analysis result.