HAI Book 2025 - Flipbook - Page 587
Schwartz, Daniel
136
Cortical atrophy is correlated with the discrepancy between
neuroradiological assessment and binary quantitation of florbetapir PET in
nondemented older adults
Daniel Schwartz1,3, Ivan E. Villamizar Alvarado1,3,4, Laszlo Szidonya2, Gagandeep Choudhary2,
Jennifer Marcoe1,4, Amanda Mar1,4, Derrick Gillan2, Hollie Hendricks2, Jeffrey Kaye1,4, Lisa
Silbert1,4,5
1
OHSU Neurology, Portland, OR, US
OHSU Radiology, Portland, OR, US
3
OHSU Advanced Imaging Research Center, Portland, OR, US
4
Oregon ADRC, Portland, OR, US
5
Portland Veterans Affairs Healthcare System, Portland, OR, US
2
Background: With FDA approval of anti-amyloid therapies and newly proposed biomarker-based Alzheimer9s
disease (AD) diagnostic criteria[1], accurate assessment of individual cerebral amyloid status in early stage
disease is essential. Clinical determination of PET amyloid positivity is complicated by the need to evaluate tracer
intensity in neocortical gray matter (GM) relative to white matter (WM) that exhibits non-specific binding[2], an
assessment further obfuscated when cortical atrophy results in reduced GM signal. We sought to investigate
contributions of cortical thinning to discrepancies in clinical and quantitative amyloid status determinations in
dementia-free older adults.
Materials and Methods: A prospective cohort of 119 adults (age 78.5y) underwent florbetapir PET and MRI at
Oregon Health & Science University (Table 1). Each was acquired and processed according to ADNI guidelines[3]
using SUVR cutoff g1.11[4], and clinically assessed by an experienced neuroradiologist. Freesurfer 7.2.0[5]
determined cortical thickness and neocortical parcellations.
Results: Fifty-one individuals had an SUVRg1.11. Twenty-six clinical reads disagreed with quantitative
assessment; these participants were significantly closer to the cutoff in those with an SUVRg1.11 (t (50)=3.12,p
