HAI Book 2025 - Flipbook - Page 246
Betthauser, Tobey
57
Longitudinal amyloid PET trajectories in the HABS-HD cohort
Tobey Betthauser1, Jacob Morse1, Margo Heston1, Rebecca Langhough1,2, Finnuella Carey1,
Julie Wisch3, Beau Ances3, Annie Cohen4, Arthur Toga5, Meredith Braskie5, Megan
Zuelsdorff1,6, Sid O9Bryant7, Karin Meeker7
1
University of Wisconsin-Madison, School of Medicine and Public Health, Department of Medicine, Madison, WI, US
University of Wisconsin-Madison, Wisconsin Alzheimer’s Institute, Madison, WI, US
3
Washington University School of Medicine, Department of Neurology, St Louis, MO, US
4
University of Pittsburgh, Department of Psychiatry, Pittsburgh, PA, US
5
Mark and Mary Stevens Neuroimaging and Informatics Institute, Keck School of Medicine, University of Southern
California, Los Angeles, CA, US
6
University of Wisconsin-Madison, School of Medicine and Public Health, Department of Nursing, Madison, WI, US
7
University of North Texas Health Science Center, Institute for Translational Research, Fort Worth, TX, US
2
Background: Amyloid PET and longitudinal accumulation trajectories are well-characterized in cohorts primarily
comprised of non-Hispanic White (NHW) participants. This work investigates whether sampled iterative local
approximation (SILA), previously validated for amyloid trajectory modeling in NHW cohorts, generalizes to other
ethnoracial groups from the HABS-HD-ATN study.
Methods: 556 participants with longitudinal Florbetataben PET SUVR were included (Table 1). Cortical amyloid was
quantified averaging SUVR across FreeSurfer-defined regions (whole cerebellum reference region, 90-110 min).
A+ was defined as global SUVR>1.076 based on GMM. SILA was applied across the entire sample to model
longitudinal amyloid accumulation and estimate individualized A+ age and time (A+ age minus age at scan) using
each person9s last scan as a reference. SILA SUVR residuals were calculated at the first scan to investigate
amyloid accumulation trajectories differences. ANOVA/Chi2 tested for differences in age, %A+, global SUVR, and
for SUVR prediction residual differences between groups stratified by A+/- and secondarily within A- and A+
strata. LME tested for possible differences in amyloid accumulation rates between groups, stratified by A+/(ageXgroupXA+/-).
Results: Ethnoracial groups differed on age, FBB SUVR, and %A+ (age: NHW >Hispanic >Black). NHW had higher
SUVR and higher %A+ than Black and Hispanic groups (Table 1, Figure 1). ANOVA indicated main effects of A+/and group for SILA prediction residuals, which were lower for NHW and for A- groups (Figure 2). No significant
group differences were observed for SUVR residuals within A+/- strata (p(A-)=0.10, p(A+)=0.10. LME indicated no
significant group differences in amyloid accumulation rates (interaction p=0.75).
Discussion: Preliminary analyses suggest SILA is amenable to amyloid trajectory modeling across ethnoracial
groups and amyloid PET trajectories are similar across groups but require further investigation with larger
longitudinal samples and more statistical power. Further work will investigate possible differences in A+ onset
age and possible differential effects of APOE.
HAI2025 - 246
