Background: Imaging with amyloid-β PET can potentially aid the early and accurate diagnosis of Alzheimer's disease. Florbetaben (¹⁸F) is a promising ¹⁸F-labelled amyloid-β-targeted PET tracer in clinical development. We aimed to assess the sensitivity and specificity of florbetaben (¹⁸F) PET in discriminating between patients with probable Alzheimer's disease and elderly healthy controls.
Methods: We did a multicentre, open-label, non-randomised phase 2 study in 18 centres in Australia, Germany, Switzerland, and the USA. Imaging with florbetaben (¹⁸F) PET was done on patients with probable Alzheimer's disease (age 55 years or older, mini-mental state examination [MMSE] score=18-26, clinical dementia rating [CDR]=0·5-2·0) and age-matched healthy controls (MMSE ≥ 28, CDR=0). Our primary objective was to establish the diagnostic efficacy of the scans in differentiating between patients with probable disease and age-matched healthy controls on the basis of neocortical tracer uptake pattern 90-110 min post-injection. PET images were assessed visually by three readers masked to the clinical diagnosis and all other clinical findings, and quantitatively by use of pre-established brain volumes of interest to obtain standard uptake value ratios (SUVRs), taking the cerebellar cortex as the reference region. This study is registered with ClinicalTrials.gov, number NCT00750282.
Findings: 81 participants with probable Alzheimer's disease and 69 healthy controls were assessed. Independent visual assessment of the PET scans showed a sensitivity of 80% (95% CI 71-89) and a specificity of 91% (84-98) for discriminating participants with Alzheimer's disease from healthy controls. The SUVRs in all neocortical grey-matter regions in participants with Alzheimer's disease were significantly higher (p < 0·0001) compared with the healthy controls, with the posterior cingulate being the best discriminator. Linear discriminant analysis of regional SUVRs yielded a sensitivity of 85% and a specificity of 91%. Regional SUVRs also correlated well with scores of cognitive impairment such as the MMSE and the word-list memory and word-list recall scores (r -0·27 to -0·33, p ≤ 0·021). APOE ɛ4 was more common in participants with positive PET images compared with those with negative scans (65%vs 22% [p=0·027] in patients with Alzheimer's disease; 50%vs 16% [p = 0·074] in healthy controls). No safety concerns were noted.
Interpretation: We provide verification of the efficacy, safety, and biological relevance of florbetaben (¹⁸F) amyloid-β PET and suggest its potential as a visual adjunct in the diagnostic algorithm of dementia.
Funding: Bayer Schering Pharma AG.
Copyright © 2011 Elsevier Ltd. All rights reserved.