Reproducibility of automated simplified voxel-based analysis of PET amyloid ligand [11C]PIB uptake using 30-min scanning data

Purpose Positron emission tomography (PET) with (11)C-labelled Pittsburgh compound B ([(11)C]PIB) enables the quantitation of beta-amyloid accumulation in the brain of patients with Alzheimer's disease (AD). Voxel-based image analysis techniques conducted in a standard brain space provide an objective, rapid and fully automated method to analyze [(11)C]PIB PET data. The purpose of this study was to evaluate both region- and voxel- level reproducibility of automated and simplified [(11)C]PIB quantitation when using only 30 min of imaging data. Methods Six AD patients and four healthy controls were scanned twice with an average interval of 6 weeks. To evaluate the feasibility of short scanning (convenient for AD patients), [(11)C]PIB uptake was quantitated using 30 min of imaging data (60 to 90 min after tracer injection) for region-to-cerebellum ratio calculations. To evaluate the reproducibility, a test-retest design was used to derive absolute variability (VAR) estimates and intra-class correlation coefficients at both region-of-interest (ROI) and voxel level. Results The reproducibility both at the region level (VAR 0.9-5.5%) and at the voxel level (VAR 4.2-6.4%) was good to excellent. Based on the variability estimates obtained, power calculations indicated that 90% power to obtain statistically significant difference can be achieved using a sample size of five subjects per group when a 15% change from baseline (increase or decrease) in [(11)C]PIB accumulation in the frontal cortex is anticipated in one group compared to no change in another group. Conclusion Our results showed that an automated analysis method based on an efficient scanning protocol provides reproducible results for [(11)C]PIB uptake and appears suitable for PET studies aiming at the quantitation of amyloid accumulation in the brain of AD patients for the evaluation of progression and treatment effects.