With progressive implementation of amyloid PET imaging in clinical settings, issues regarding elderly pts compliance during scan have been proposed. [18F]flutemetamol (Vizamyl™) was recently approved to assess β-amyloid neuritic plaque density in cognitively impaired patients. The primary mean of assessing [18F]flutemetamol scans in clinical settings is by visual evaluation, using the late 20-min summation image, displayed using a rainbow colour scale. However, this acquisition time frequently appears too much long for people cognitively compromised. Aim of the study was to compare early 10-min (E-10min) to late 20-min (L-20min) summation images to assess potential use of the E-10min acquisition method in the clinical context.
Ten patients with diagnosis of MCI and dementia syndrome, using the new NIA-NIH criteria, were submitted to [18F]flutemetamol PET scan. Neuropsychological battery (abstract reasoning, memory, attention, language, praxis and visuo-perceptive functions), Geriatric Depression Scale and Summary Performance Physical Battery were assessed. Hand-Grip strength was measured by manual dynamometer. All patients underwent brain MRI or CT scan in the previous 3 months. PET scan began 90 minutes after injection of 185 MBq [18F]flutemetamol. using a whole-body hybrid system Discovery IQ (GE Healthcare) operating in 3D detection mode. PET images were visually interpreted by 2 board-certified independent nuclear medicine physicians. To validate visual results, we performed quantitative assessment of [18F]flutemetamol uptake using CortexID Suite (GE Healthcare) software implemented in Advantage GE workstation. CortexID Suite produces a Z-score surface map that highlights areas and degree of beta amyloid tracer uptake. The severity is measured in standard deviations from normal controls.
Visual assessment didn't change in any patient comparing E-10min vs L-20min independently of the reference region (pons/cerebellum). Quantitative analysis confirm visual assessment conclusion, in fact assessing E-10min vs L-20min with a Pair T test using Bonferroni correction and p=0.01, results are not statistically different. This means that, comparing E-10min vs L-20min, we have the same data for all areas of the Cortex ID Suite.
Our results show that acquisition time may be reduced to 10 min without decreasing diagnostic accuracy in the clinical context. Moreover, quantitative analysis confirm robustness of acquired data for research utilization.
© 2017 Published by Elsevier Inc.