A dual-time-window protocol to reduce acquisition time of dynamic tau PET imaging using [18F]MK-6240

Background [ 18 F]MK-6240 is a PET tracer with sub-nanomolar affinity for neurofibrillary tangles. Therefore, tau quantification is possible with [ 18 F]MK-6240 PET/CT scans, and it can be used for assessment of Alzheimer’s disease. However, long acquisition scans are required to provide fully quantitative estimates of pharmacokinetic parameters. Therefore, on the present study, dual-time-window (DTW) acquisitions was simulated to reduce PET/CT acquisition time, while taking into consideration perfusion changes and possible scanning protocol non-compliance. To that end, time activity curves (TACs) representing a 120-min acquisition (TAC 120 ) were simulated using a two-tissue compartment model with metabolite corrected arterial input function from 90-min dynamic [ 18 F]MK-6240 PET scans of three healthy control subjects and five subjects with mild cognitive impairment or Alzheimer’s disease. Therefore, TACs corresponding to different levels of specific binding were generated and then various perfusion changes were simulated. Next, DTW acquisitions were simulated consisting of an acquisition starting at tracer injection, a break and a second acquisition starting at 90 min post-injection. Finally, non-compliance with the PET/CT scanning protocol were simulated to assess its impact on quantification. All TACs were quantified using reference Logan’s distribution volume ratio (DVR) and standardized uptake value ratio (SUVR 90 ) using the cerebellar cortex as reference region. Results It was found that DVR from a DTW protocol with a 60-min break between two 30-min dynamic scans closely approximates the DVR from the uninterrupted TAC 120 , with a regional bias smaller than 2.5%. Moreover, SUVR 90 estimates were more susceptible (regional bias ≤ 19%) to changes in perfusion compared to DVR from a DTW TAC (regional bias ≤ 10%). Similarly, SUVR 90 was affected by late-time scanning protocol delays reaching an increase of 8% for a 20-min delay, while DVR was not affected (regional bias