Longitudinal tau quantification may provide a useful marker of drug efficacy in clinical trials. Different tau PET tracers may have different sensitivity to longitudinal changes, but without a head-to-head dataset or a carefully designed case-matching procedure, comparing results in different cohorts can be biased. In this study, we compared the tau PET tracers, ¹⁸F-MK6240 and ¹⁸F-flortaucipir (FTP), both cross-sectionally and longitudinally by case-matching subjects in the AIBL and ADNI longitudinal cohort studies.

A subset of 113 participants from AIBL and 113 from ADNI imaged using ¹⁸F-MK6240 and ¹⁸F-FTP respectively, with baseline and follow-up, were matched based on baseline clinical diagnosis, MMSE, age and amyloid (Aβ) PET centiloid value. Subjects were grouped as 64 Aβ- cognitively unimpaired (CU), 22 Aβ+ CU, 14 Aβ+ mild cognitive impairment (MCI) and 13 Aβ+ Alzheimer’s disease (AD). Tracer retention was measured in the mesial, temporoparietal, rest of the cortex, and a meta-temporal region composed of entorhinal, inferior/ middle temporal, fusiform, parahippocampus and amygdala. T-tests were employed to assess group separation at baseline using SUVR Z-scores and longitudinally using SUVR%/Yr.

Both tracers detected statistically significant differences at baseline in most regions between all clinical groups. Only ¹⁸F-MK6240 showed statistically significant higher rate of SUVR increase in Aβ+ CU compared to Aβ- CU in the mesial, meta-temporal and temporoparietal regions.

¹⁸F-MK6240 appears to be a more sensitive tracer for change in tau level at the preclinical stage of AD.