[18F]RO948 Tau PET in early and late onset Alzheimer’s disease: associations with plasma p-tau 217, atrophy and cognition

Purpose The aim of this study was to compare regional brain tau extent across the disease spectrum in early- and late-onset Alzheimer’s disease using the second-generation tau PET tracer [ 18 F]RO948, and to investigate the relationship between PET-derived patterns, plasma p-tau217, brain atrophy, and cognition. Methods We examined 57 participants: 39 patients from the Cognitive Assessment Unit at Karolinska University Hospital (Stockholm/Sweden)—diagnosed as MCI Aβ– ( n = 10), MCI Aβ + ( n = 16; 8 EO and 6 LO), and AD ( n = 13; 5 EOAD and 8 LOAD)—and 18 cognitively normal controls. All participants underwent [ 18 F]RO948 tau-PET, structural MRI, cognitive testing, and plasma p-tau217 analysis. Results Both EO and LO MCI (Aß +) showed higher [ 18 F]RO948 binding in amygdala, entorhinal cortex, hippocampus ( p < 0.001) and extending to inferior temporal regions ( p < 0.01) compared to controls, with some MCI-EO cases already showing advanced neocortical tau burden. At the AD stage, EOAD patients showed greater neocortical tau extent in temporo-parietal and frontal cortices than LOAD. P‑tau217 positively correlated with [ 18 F]RO948 in amygdala, entorhinal cortex, hippocampus (rho = 0.81) followed by temporal (rho = 0.71), and parietal cortices (rho = 0.56, all p < 0.001). The Tau‑PET/p-tau217 association was different between EOAD and LOAD with greater differences in parietal cortices. Tau‑PET also showed stronger region‑specific associations with cognitive impairment than p‑tau217, especially in EOAD. Conclusions [ 18 F]RO948-PET captured interindividual heterogeneity in tau accumulation that was not reflected by plasma p-tau217. This dissociation was particularly evident in neocortical regions, where LOAD showed elevated plasma p-tau217 levels despite low tau PET binding. These findings highlight the value of tau-PET as a sensitive biomarker for detecting early disease heterogeneity, assessing disease severity, and improving patient stratification for disease-modifying treatments.