Towards a universal cortical tau sampling mask: Neuroimaging / Optimal neuroimaging measures for early detection. (7th December 2020)

• Record Type: Journal Article
• Authors: Dore, Vincent; Rowe, Christopher C; Bourgeat, Pierrick; Burnham, Samantha C; Huang, Kun; Krishnadas, Natasha; Masters, Colin L; Fripp, Jurgen; Villemagne, Victor LL
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 5
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.040667
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: The introduction of the A/T/(N) framework raised several issues regarding the definition of T+. In which region should tau scans be sampled? Based on which tracer? In this work, we developed a "universal" cortical tau mask derived from three major tau ligands for the AD continuum, over which several different global/regional sampling VOI or composites can be applied. Method: 221 healthy elderly controls (HC) and 67 Alzheimer's disease (AD) patients underwent tau scans with either 18 F‐AV1451 (HC=54/AD=24), 18 F‐MK6240 (HC=157/AD=22) or 18 F‐PI2620 (HC=10/AD=21). All HC were Ab‐ (<20CL) and all AD were Ab+(20CL). The tau scans were spatially normalized using CapAIBL and the cerebellum cortex was used as reference region. For each tracer, a difference image between the means of the Ab‐HC and Ab+AD patients was then generated. Difference images were subsequently thresholded at 1/3 of the difference between Ab‐HC and Ab+AD in the inferior temporal lobe.A MRI‐derived grey matter mask at PET resolution was applied to the composite mask to only sample cortical region. The mask was finally mirrored and fused to remove the hemispherical asymmetry due to the distribution of tau deposits in the assessed participants. Agreement between masks was assessed by dice‐scores. Result: Visually, the three tracer‐specific masks appeared very similar (Fig 1, rows 1‐3). None of the known off‐target binding regions was discernible in the resulting masks. There was good agreement between all masks, with dice‐scores of 0.60 and 0.66 for cortical regions. Conclusion: We constructed an "universal" tau mask for the AD continuum based on three commonly used tau tracers aiming at standardizing tau sampling (and quantification) across tracers and across centres. The "universal" tau mask can also be sub‐segmented into smaller regions to focus on specific areas or to construct a subset of composite masks that might better capture early tau deposition and spreading. Further refinement of the tau mask will require the addition of more tau tracers. (RO948, GTP1, PBB3, etc.).