HAI Book 2025 - Flipbook - Page 80
Tissot, Cécile
Comparative sensitivity analysis of tau-PET tracers in Alzheimer9s disease
Cécile Tissot1, Nesrine Rahmouni2, Hsin-Yeh Tsai1, Joseph Therriault2, Stijn Servaes2, Firoza
Z. Lussier3, Stefania Pezzoli4, Jacob Ziontz,, Brian Gordon, Belen Pascual, Val Lowe, David
Soleimani-Meigooni, Hwamee Oh, William Klunk, Pedro Rosa-Neto, William Jagust, Tharick
Pascoal, Suzanne Baker
1
Lawrence Berkeley National Laboratory, Berkeley, CA, US
McGill University, Montreal, QC, CA
3
University of Pittsburgh, Pittsburgh, PA, US
4
University of California, Berkeley, Berkeley, CA, US
5
Washington University in Saint Louis, Saint Louis, MO, US
6
Houston Methodist, Houston, TX, US
7
Mayo Clinic, Rochester, MN, US
8
University of California, San Francisco, San Francisco, CA, US
9
Brown University, Providence, RI, US
2
Background: Tau-PET tracers are essential for visualizing pathology in Alzheimer9s (AD). High sensitivity is crucial
for early detection and monitoring tau deposition. This study compares the noise to dynamic range ratio (NRR) of
[18F]FTP, [18F]MK6240, [18F]PI2620, and [18F]RO948.
Methods: 328 individuals from the HEAD study (178 cognitively unimpaired elderly (CUE) and 150 cognitively
impaired) underwent [18F]FTP and [18F]MK6240 tau-PET scans; 71 additionally received [18F]PI2620 and [18F]RO948.
Noise was calculated as the standard deviation (SD) of CUE A´- participants aged f65 (SDCUEA´-f65). The dynamic
range was calculated as the SD across all subjects (SDrange). The lower the NRR (=SDCUEA´-f65/SDrange) the more
sensitive the tracer. Analyses were performed at both region-of-interest and voxel-wise levels.
Results: In the whole cohort, [18F]MK6240 demonstrated lower NRR than [18F]FTP in all on-target brain regions
(Figure1A). CUEA´-f65 may exhibit real tau-PET retention in Braak I, limiting conclusions there. Within the fourtracers subset, [18F]PI2620 showed the highest NRR in Braak II, followed by [18F]RO948, [18F]FTP and [18F]MK6240.
In the metatemporal-ROI, Braak III and IV, [18F]RO948 had the highest value, followed by [18F]PI2620, [18F]FTP and
[18F]MK6240. For Braak V and VI, [18F]RO948 again exhibited the highest values, followed by [18F]FTP, [18F]PI2620
and [18F]MK6240 (Figure1B). Voxel-wise analyses presented a similar pattern: [18F]MK6240 showed low NRR
values, [18F]PI2620 and [18F]FTP were comparable, and [18F]RO948 exhibited the highest (Figure2). The lower
[18F]MK6240 NRR result from large SDrange, while [18F]FTP's are due to smaller SDCUEA´-f65. In contrast, high
[18F]PI2620 and [18F]RO948 values are driven by larger SDCUEA´-f65.
Discussion: Our findings highlight significant variability in tau-PET tracers. [18F]MK6240 consistently
demonstrated lower NRR, implying better sensitivity across all regions; [18F]FTP outperforms [18F]RO948 and
[18F]PI2620 in metatemporal-ROI and Braak III-IV. [18F]RO948 (excluding hippocampus) consistently exhibited
worse sensitivity. These results provide insights into the differential tracer sensitivity, helping guide their optimal
use in detecting and tracking tau.
Keywords: Tau, noise, sensitivity, Alzheimer’s disease, head-to-head
HAI2025 - 80
