Impact of Postlabeling Delay on Arterial Spin-Labeling MRI-Based Measures of Cerebral Blood Flow in a Simultaneous PET/MRI Study of Aging.

BACKGROUND AND PURPOSE: Arterial spin-labeling (ASL) MRI with a single postlabeling delay (PLD) of 1800 msec is recommended for quantifying CBF when the PLD is sufficiently longer than the arterial transit time. However, this condition may not be satisfied in older participants or individuals with dementia, for whom multiple PLDs may be more beneficial. This work assessed the intermediate-term (4-week) test-retest reliability and accuracy of single- versus multiple-PLD ASL for quantifying CBF relative to [15O]water PET acquired using PET/MRI imaging in a study of aging and Alzheimer disease (AD). MATERIALS AND METHODS: Cognitively normal young (n=3) and older participants (n=5) and patients with AD (n=3) underwent PET/MRI sessions. ASL CBF was quantified separately using PLDs of 1800, 2000, 2200, and 2400 msec, and using multiple-PLD approach combining all 4 PLDs. A subset of participants (1 AD, 4 older cognitively normal) underwent repeated ASL MRI after approximately 4 weeks for test-retest evaluation, assessed using Bland-Altman plots and intraclass correlation coefficients (ICCs). Accuracy was assessed by comparing ASL CBF with [15O]water PET using Spearman correlation across cortical regions and vascular territories. RESULTS: The 4-week test-retest reliability was highest for single-PLD ASL using 1800 msec (ICC > 0.75) and lowest for single-PLD of 2400 msec (ICC: 0.1-0.8). Multiple-PLD ASL showed better reliability than the 2400 msec PLD but was not superior to shorter PLDs. Correlations with [15O]water PET were strongest for 1800 msec single-PLD and generally decreased with longer PLDs, except the occipital lobe, where all PLDs showed poor agreement. Multiple-PLD ASL showed similar correlation strengths as 2000 msec single-PLD. CONCLUSIONS: Single-PLD ASL with 1800 msec PLD demonstrated the best balance of reliability and agreement with [15O]water PET, suggesting it may be an effective choice for studying CBF in aging and AD populations. While multiple-PLD ASL offered comparable performance, it did not confer notable benefits compared with the best-performing single-PLD under similar acquisition time constraints. These findings underscore the importance of PLD selection in ASL protocols in older or cognitively impaired cohorts and highlight the value of assessing intermediate-term reliability in longitudinal studies.