HAI Book 2025 - Flipbook - Page 406
Lim, Seokbeen
100
Investigation of optimal intensity threshold for tau-PET overlap
index
Seokbeen Lim1, Jeyeon Lee2, Hoon-Ki Min1, Christina M. Moloney3, Carly T. Mester4, Sujala
Ghatamaneni1, Matthew L. Senjem5, Aivi T. Nguyen6, Jonathan Graff-Radford7, Christopher G.
Schwarz1, Jeffrey L. Gunter1, Kejal Kantarci1, Bradley Boeve7, Prashanthi Vemuri1, David T.
Jones7, Clifford R. Jack1, Ronald C. Petersen7, Melissa E. Murray3, Val J. Lowe1
1
Department of Radiology, Mayo Clinic, Rochester, MN, US
Department of Biomedical Engineering, Hanyang University, Seoul, KR
3
Department of Neuroscience, Mayo Clinic, Jacksonville, FL, US
4
Department of Health Sciences Research, Mayo Clinic, Rochester, MN, US
5
Information Technology, Mayo Clinic, Rochester, MN, US
6
Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, US
7
Department of Neurology, Mayo Clinic, Rochester, MN, US
2
Background: The Overlap Index (OI) we previously published (Lee et al., 2022) is a reliable method for detecting
tau accumulation using tau-PET imaging (flortaucipir; AV1451). It identifies voxel-wise standardized uptake value
ratio (SUVr) elevations across serial scans. Previous studies selected OI intensity thresholds based on OI and SUVr
relationships. However, parameter optimization is needed to confirm correlations between neuropathologic
Braak stages and Meta-ROI OI. This study aims to optimize the OI intensity threshold in tau-PET by comparing it to
neuropathologic Braak staging.
Methods: This study included 57 participants from the Mayo Clinic Study of Aging or Mayo Alzheimer's Disease
Research Center, all of whom underwent autopsy and at least two serial AV1451 tau-PET scans and 3T T1-weighted
MRI within and average of 16 months. Tau-PET SUVr images were normalized to cerebellar crus grey matter, and
MR images were co-registered to a single-subject MRI mean. Tau-PET scans were resampled into this space, and
OI was calculated by selecting a ROI and applying intensity thresholds (1.231.4). Clusters with fewer than 20
contiguous voxels were excluded. Neuropathologic Braak staging was determined using AT8
immunohistochemistry. Sensitivity, specificity, and accuracy were calculated to compare detection performance
between OI and SUVr.
Results: OI demonstrated higher sensitivity and accuracy than meta-ROI SUVr across all thresholds in detecting
tau pathology, particularly in early Braak stages (I-IV). OI identified NFTs in the entorhinal cortex, especially in
moderate to frequent NFT cases, outperforming SUVr. OI detected FTLD-tau and GGT pathology in individuals at
Braak stage 0, which SUVr didn9t. While SUVr had higher specificity, OI at an intensity threshold of 1.3 offered
balanced performance for early tau pathology detection.
Conclusion: The OI in serial tau PET imaging outperforms SUVr in detecting localized tau accumulation by
capturing elevated signals in voxel subsets within the ROI, with a 1.3 optimal threshold.
HAI2025 - 406
