A preprint posted to medRxiv on 24 June 2026 by a University of Cambridge group reports that the elevated TSPO PET signal seen in frontotemporal dementia corresponds to a heavier burden of CD68 positive phagocytic microglia, concentrated in white matter rather than spread through cortical grey matter. The study, led by Davi S. Vontobel with Maura Malpetti as senior author, paired [11C]PK11195 TSPO PET acquired during life with post mortem neuropathology in 10 people with FTD (five with FTLD-tau, five with FTLD-TDP) and five control donors. CD68 immunohistochemistry was performed across 17 brain regions, alongside multiplex immunofluorescence pairing TSPO with microglial and macrophage markers (IBA1, CD68), an astrocyte marker (GFAP) and an endothelial marker (CD31), plus three dimensional single cell reconstruction. The preprint has not been peer reviewed.

CD68 burden was higher in the FTD donors, concentrated in white matter, and correlated with regional TSPO PET binding across both pathology groups (beta = 8.40, P < 0.001), the authors report. Only the fraction of TSPO that co-localised with CD68 tracked the imaging signal, and there was no increase in TSPO per cell. On that reading, the scan is registering how many lysosome rich phagocytic microglia are present rather than how much TSPO each microglial cell has switched on. The authors conclude the finding supports TSPO PET as a marker of microglial burden in both FTLD-tau and FTLD-TDP.

The question matters because TSPO is expressed by microglia, astrocytes and endothelial cells, and the field has argued for years over which of those the PET signal reflects. The same Cambridge network reported a comparable result in a different condition last year: in Brain, Wijesinghe and colleagues studied eight people with progressive supranuclear palsy who had [11C]-PK11195 PET in life and three controls, and found regional binding correlated with post mortem CD68 positive microglia and with microglial, not astrocytic, TSPO. Separately, a 2023 Brain paper from the group reported that microglial activation measured in the frontal cortex predicted later cognitive decline in FTD, which is part of why the interpretation of the signal carries weight for trial design.

The counterpoint is scale and method. Fifteen brains, ten of them with FTD, is a small series for a claim about what a widely used tracer measures, and the design is observational: imaging done in life is being compared with tissue examined after death, so the association cannot establish that phagocytic microglia cause the signal or that either causes decline. The preprint has not been through peer review, and its findings apply to [11C]PK11195 in FTD rather than to every TSPO tracer or every dementia.