Optimum Differentiation of Frontotemporal Lobar Degeneration from Alzheimer Disease Achieved with Cross‐Sectional Tau Positron Emission Tomography

• Published in: Annals of neurology 2022-12, Vol.92 (6), p.1016-1029
• Main Authors: Josephs, Keith A., Tosakulwong, Nirubol, Gatto, Rodolfo G., Weigand, Stephen D., Ali, Farwa, Botha, Hugo, Graff‐Radford, Jonathan, Machulda, Mary M., Savica, Rodolfo, Schwarz, Christopher G., Senjem, Matthew L., Boeve, Bradley F., Kantarci, Kejal, Jones, David T., Ramanan, Vijay K., Fields, Julie A., Reichard, Ross R., Dickson, Dennis W., Petersen, Ronald C., Jack, Clifford R., Lowe, Val J., Whitwell, Jennifer L.
• Format: Article
• Language: English
• Subjects: Accumulation; Alzheimer Disease - diagnostic imaging; Alzheimer Disease - pathology; Alzheimer's disease; Autopsies; Autopsy; Carbolines; Cross-Sectional Studies; Degeneration; Differentiation; Emission analysis; Frontotemporal Dementia; Frontotemporal Lobar Degeneration - diagnostic imaging; Frontotemporal Lobar Degeneration - pathology; Globus pallidus; Humans; Mesencephalon; Neurodegenerative diseases; Positron emission; Positron emission tomography; Positron-Emission Tomography - methods; Sensitivity; Tau protein; tau Proteins; Tomography
• ISSN: 0364-5134; 1531-8249; 1531-8249
• DOI: 10.1002/ana.26479

Objective This study was undertaken to assess cross‐sectional and longitudinal [18F]‐flortaucipir positron emission tomography (PET) uptake in pathologically confirmed frontotemporal lobar degeneration (FTLD) and to compare FTLD to cases with high and low levels of Alzheimer disease (AD) neuropathologic changes (ADNC). Methods One hundred forty‐three participants who had completed at least one flortaucipir PET and had autopsy‐confirmed FTLD (n = 52) or high (n = 58) or low ADNC (n = 33) based on Braak neurofibrillary tangle stages 0–IV versus V–VI were included. Flortaucipir standard uptake value ratios (SUVRs) were calculated for 9 regions of interest (ROIs): an FTLD meta‐ROI, midbrain, globus pallidum, an AD meta‐ROI, entorhinal, inferior temporal, orbitofrontal, precentral, and medial parietal. Linear mixed effects models were used to compare mean baseline SUVRs and annual rate of change in SUVR by group. Sensitivity and specificity to distinguish FTLD from high and low ADNC were calculated. Results Baseline uptake in the FTLD meta‐ROI, midbrain, and globus pallidus was greater in FTLD than high and low ADNC. No region showed a greater rate of flortaucipir accumulation in FTLD. Baseline uptake in the AD‐related regions and orbitofrontal and precentral cortices was greater in high ADNC, and all showed greater rates of accumulation compared to FTLD. Baseline differences were superior to longitudinal rates in differentiating FTLD from high and low ADNC. A simple baseline metric of midbrain/inferior temporal ratio of flortaucipir uptake provided good to excellent differentiation between FTLD and high and low ADNC (sensitivities/specificities = 94%/95% and 71%/70%). Interpretation There are cross‐sectional and longitudinal differences in flortaucipir uptake between FTLD and high and low ADNC. However, optimum differentiation between FTLD and ADNC was achieved with baseline uptake rather than longitudinal rates. ANN NEUROL 2022;92:1016–1029