HAI Book 2025 - Flipbook - Page 403
Olafson, Emily
99
Characterization and correction of extracerebral [18F]MK-6240 tau
PET off-target signal in Alzheimer's disease
Emily Olafson1, Matteo Tonietto2, Nick Doren2, Sandra Sanabria Bohórquez1
1
Research and Early Development (gRED), Genentech, Inc., South San Francisco, CA, US
Research and Early Development (pRED), Hoffmann-La Roche, Basel, CH
2
Introduction: [18F]MK-6240 is a next-generation tau-PET tracer with high affinity for tau aggregates in
Alzheimer9s disease (AD). It exhibits off-target signals in extracerebral regions such as the meninges and sinus
that may impact signal quantification. In this study, we aimed to characterize the extracerebral [18F]MK-6240
signal and to develop a method to mitigate its impact on tau-PET quantification.
Methods: We analyzed [18F]MK-6240 PET images from the Lantheus database (Table 1). Non-negative matrix
factorization (NMF) was used to identify interpretable clusters of the extracerebral signal in cognitively
unimpaired amyloid negative subjects (CU/A´-). Eroded regions of interest (ROIs) for the meta temporal (MT) and
cerebellum were created to measure cerebral signal without influence from extracerebral compartments (see
Figure 1A). Spill-over was quantified as the percent difference in SUV (calculated using subject weight and
injected dose) between the full and eroded (%Δ[eroded-full]) (Figure 1A). Using the eroded cerebellum as a
reference, coefficients capturing the meningeal spill-over into the MT were calculated by modeling %Δ[erodedfull] as a function of the meninges SUVR (Figure 2A, B). SUVRs in the MT in amyloid positive (A´+) subjects were
adjusted using the meninges coefficients derived from CU/A´- subjects (Figure 2B). The similarity between the
eroded and adjusted MT SUVR was calculated with R2.
Results: NMF identified four components, with the meninges and sinus clustering separately (Figure 1B). In both
ROIs, the %Δ[eroded-full] was significantly related to weights for only the meningeal NMF component. Applying
HAI2025 - 403