An automatic method for the quantification of uptake with Florbetapir imaging

Objectives To evaluate an automatic method for the quantification of Florbetapir PET uptake, implemented in Scenium Ratio Analysis software, with a large clinical population composed of young cognitive normal (YCN), old cognitive normal (OCN), mildly cognitive impaired (MCI), and probably Alzheimer’s disease (PAD) subjects.

Methods The automatic method in Scenium Ratio Analysis is based on a semi-automatic method to compute the cortex-to-cerebellum SUVr described in Fleisher et al.[Archives of Neurology 2011]. The method includes registration of PET images to a standard Florbetapir PET template in MNI atlas space. Cortex-to-cerebellum SUVrs were computed for 6 target cortical regions of interest (frontal, temporal, parietal, anterior cingulate, posterior cingulate and precuneus) using the whole cerebellum as a reference. The proposed method differs from the reference method in 3 main points: automatic vs. semi-automatic, affine vs. non-linear registration, revised ROIs. The Florbetapir PET scans of 210 subjects composed of 81 YCN (age 28.1±7.9 years), 43 OCN (age 72.5±11.9 years), 30 MCI (age 75.9±10.9 years), and 56 PAD (age 77.2±9.9 years) were analyzed with Scenium Ratio Analysis whose results were compared to the reference method.

Results SUVr from Scenium Ratio analysis and semi-automatic method were highly correlated (Table1). SUVr for 3/210 were significantly different across two methods, resulting from sub-optimal registration in the automatic method. Table 1 summarizes the distribution of automatic SUVr across clinical populations and its correlation with semi-automatic method.

Conclusions Results show that Scenium Ratio Analysis enables automatic computation of SUVr for quantification of Florbetapir PET images with numerical results comparable to the ones obtained in Fleisher et al.[Archives of Neurology 2011]

• Download figure
• Open in new tab
• Download powerpoint

Table 1: SUVr using Scenium Ratio Analysis from Florbetapir PET images and its correlation with semi-automatic method for YCN, OCN, MCI and PAD.