Analysis of gray-matter quantitation showing brain ROIs (colored areas in 1st column) and ROIs overlaid onto raw (unmasked) transverse slices (2nd column), and same ROIs onto white-matter and CSF masked slices (3rd column).
Similar to brain scanning for Parkinson's disease, new radiotracers have been FDA-approved within the past years to image the deposition of beta-amyloid (associated with Alzheimer's disease) in the human brain. The gray matter analysis evaluates these tracers. The ADER analysis also makes use of a "standard geometry" permitting each patient's brain scan to be oriented and sized into a consistent orientation. Once registered to this standard geometry, predefined masks that exclude all activity outside of the brain (CSF) and exclude white-matter activity are applied to the patient's images. Three example transverse slices are shown in the right two panels of this figure where (after applying the predefined masks), the black background represents activity not analyzed and the uptake of a beta-amyloid targeting radiotracer is shown in levels of gray. The red-lined regions of interest (ROI's) are shown overlaid onto these images where the right-most column shows the brain after applying the predefined masks, and the middle column shows these same regions on the same brain images before the masks are applied. The left-hand column shows three example colored slices of the ROI regions that define the brain areas to be selected for analysis. Each color in the left-most column represents a brain region/structure (ROI) is defined in the AAL brain template. Here only three example slices are shown, however, this analysis is applied to all slices of the brain to extract Standard Uptake Value Ratios (SUVR) which are used for expressing uptake intensities for the new investigational drugs. SUVRs are also commonly applied clinically for quantitatively evaluating molecular targets such as beta-amyloid burdens in the human brain.
Results of 368 ADER automated quantitative SUVR values (y-axis) for beta-amyloid positive (“pos”) versus beta-amyloid negative (“neg”) subject groups. Determination of “pos” versus “neg” was based on 5 trained radiology readers. AUC = 0.867, sens & spec = 80% & 82%.
Published: G. Zubal, S. Bullich, A. M. Catafau, M. Ichise, P. M. Kuo, A. Stephens, and J. P. Seibyl, “An automated, operator-independent method for quantitative analysis of 18F-florbetaben PET scans,” EANM Conference Abstracts, vol. OP117, no. 14:41, Oct., 2014.
The results for ADER's automated quantitative SUVR results were obtained from 368 beta-amyloid scans that were interpreted by consensus as beta-amyloid positive ("pos") or beta-amyloid negative ("neg") by five trained radiology readers. The quantitative values delivered by ADER have an added advantage that the value of the SUVR (in proximity to the threshold cutoff = 1.6) gives an objective "probability" that the scan reflects a "pos" versus "neg" diagnosis.