Abstract
1287
Objectives We have developed and tested PC executable code as a software application package which automatically places regions of interest (ROIs) onto nuclear medicine brain scans and extracts PET SUVr quantitative results for evaluating beta-amyloid burdens in human images.
Methods Over 50 PET scans were downloaded from the ADNI-LONI website, representing a combined group of healthy controls and subjects with Alzheimer’s Disease (AD). Our Alzheimer's Disease Evaluation of Radiotracers software package (ADER): 1) reads the reconstructed images, 2) registers, in 3D, the subject’s brain volume to a beta-amyloid PET template, 3) overlays the regions defined in the Automated Anatomical Labeled (AAL) template, 4) identifies areas of white and gray matter, and 5) uses predetermined ROIs from the AAL template to extract radiotracer brain uptake intensities to calculate regional SUVr values.
Results When compared to the analysis using interactively placed ROIs, ADER delivers human beta-amyloid concentrations with the same inter-rater variability as when the same PET scans are manually analyzed by human operators.
Conclusions This work is an extension of earlier work developed for analyzing iodinated beta-amyloid SPECT tracers [J Nucl Med. 2008; 49 (Suppl. 1):378P]. ADER, our fully automated and non-subjective brain analysis software package: a) decreases processing time to evaluate beta-amyloid concentrations in human brains, b) points the way to an objective method for comparing clinical results across imaging centers, c) offers an evaluation method for testing other radiotracers targeting beta-amyloid. This automated objective software method has been tested and statistically compared to over 150 manual analyses of both PET and SPECT brain scans.
Research Support DOE SBIR DE-FG02-08ER8517