Visual Assessment Versus Quantitative Assessment of 11C-PIB PET and 18F-FDG PET for Detection of Alzheimer's Disease

UNLABELLED: Amyloid-beta (Abeta) imaging with N-methyl-(11)C-2-(4'-methylamino-phenyl)-6-hydroxy-benzothiazole ((11)C-6-OH-BTA-1; also known as (11)C-PIB) shows a robust increase in cortical binding in Alzheimer's disease (AD). The aim of this study was to explore the clinical potential of Abeta imaging for the diagnosis of AD by comparison of the accuracy of visual reading of (11)C-PIB images with quantitative analysis and (18)F-FDG. METHODS: Fifteen AD patients (age, 71.1 +/- 11.3 y mean +/- SD; mini-mental state examination MMSE, 18.9 +/- 9.3 mean +/- SD) and 25 healthy control (HC) subjects (age, 71.9 +/- 6.82 y; MMSE >or= 28) underwent 90-min dynamic (11)C-PIB PET and 20-min static (18)F-FDG PET. (11)C-PIB images, generated from data acquired between 40 and 70 min after injection, and (18)F-FDG images were rated separately by 2 readers as normal, possible AD, or probable AD. Quantitative analyses used the distribution volume ratio (DVR) of frontal cortex, parietotemporal cortex, posterior cingulate, and caudate nucleus for (11)C-PIB and standardized uptake value ratio (SUVR) of parietotemporal cortex and posterior cingulate for (18)F-FDG, using cerebellar cortex as the reference region. Receiver-operating-characteristic (ROC) analysis was performed to compare the accuracy of quantitative measures. To determine the effect of age on diagnostic accuracy, the median age of the AD subjects (74 y) was chosen to separate the cohort into younger (64.4 +/- 5.8 y) and older (78.6 +/- 4.1 y) groups. RESULTS: Visual agreement between readers was excellent for (11)C-PIB (kappa = 0.90) and good for (18)F-FDG (kappa = 0.56). (11)C-PIB was more accurate than (18)F-FDG both on visual reading (accuracy, 90% vs. 70%, P = 0.05) and ROC analysis (95% vs. 83%, P = 0.02). Accuracy declined more with (18)F-FDG than with (11)C-PIB in the older group. CONCLUSION: Visual analysis of (11)C-PIB images appears more accurate than visual reading of (18)F-FDG for identification of AD and has accuracy similar to quantitative analysis of a 90-min dynamic scan. The accuracy of (11)C-PIB PET is limited by cortical binding in some healthy elderly subjects, consistent with postmortem studies of cerebral Abeta. Longitudinal follow-up is required to determine if this represents detection of preclinical AD.