QUANTITATIVE DISEASE PHENOTYPING BY SINGLE-TIME-POINT VOXEL-BASED AMYLOID PET DIFFERENTIATES AMYLOID-BETA ACCUMULATOR SUBTYPES

Abstract

Listen

Translate article

There is considerable interest in characterizing subgroups of patients in the earliest/preclinical disease stage, for therapeutic intervention in AD. Amyloid-beta (Abeta) early-accumulators, defined as subjects with abnormal CSF Abeta-42 but without global amyloid PET positivity present (CSF+/PET-), have been recently found (Palmqvist et al., Brain 2016) to accumulate Abeta in the brain faster than non-accumulators (CSF-/PET-), and at a similar rate as late-accumulators (CSF+/PET+). We have previously developed a fully automated amyloid PET quantitation method for calculating voxel-based-SUVR using PET in conjunction with MRI, for improved localization of deposition, and enhanced performance for identifying subjects with early disease (Mikhno et al., AAIC 2015; HAI 2016). The aim of the present study was to assess the ability of the voxel-based quantitation to differentiate accumulator subgroups in non-demented individuals, with a single integrated measure, but without the need for an invasive lumbar puncture. A total of 205 non-demented subjects (not used for training) were included from the ADNI-GO/ADNI2 cohort. [18F]florbetapir PET amyloid positivity status was established with syngo. PET Amyloid Plaque (sPAP, cutoff >1.12), a region-based quantitation method approved for clinical use. CSF amyloid positivity was established using an Abeta-42 cutoff of 192 ng/l +/- 5% to exclude borderline cases (Palmqvist 2016), and the median batch value used to further reduce variability. Voxel-based-SUVRs from corresponding PET/T1 MRI scan pairs were assessed for differences and discrimination between early-accumulator (CSF+/PET-), late-accumulator (CSF+/PET+), and non-accumulator (CSF-/PET-) subgroups. 28 early-, 83 late-, and 94 non-accumulator subjects were analyzed. Receiver operating characteristic area under the curve (AUC) was 0.967 (non- vs. late-accumulators, two-tailed t-test p=2.9e-35), 0.907 (early- vs. late-accumulators, p=3.2e-11), and 0.660 (non- vs. early-accumulators, p=0.0036). Voxel-based-SUVR showed significant group differences and high accuracy for separating early-accumulators (CSF+/PET-) and late-accumulators (CSF+/PET+), as well as late- and non-accumulators (CSF-/PET-). To our knowledge, this is the first study to show a significant difference between early- and non-accumulators, with a single-time-point integrated SUVR measure, and without the need for training on CSF data. Use of an integrated quantitative phenotype may have potential applications in disease understanding and trial enrichment.