Reliability of quantitative transverse relaxation time mapping with T 2 -prepared whole brain pCASL

Abstract

Arterial spin labeling (ASL) is increasingly applied for cerebral blood flow mapping, but T 2 relaxation of the ASL signal magnetization is often ignored, although it may be clinically relevant. To investigate the extent, to which quantitative T 2 values in gray matter (GM) obtained by pseudocontinuous ASL (pCASL) perfusion MRI can be reproduced, are reliable and a potential neuroscientific biomarker, a prospective study was performed with ten healthy volunteers (5F,28 ± 3y) at a 3 T scanner. A T 2-prepared pCASL sequence enabled the measurement of quantitative T 2 and perfusion maps. T 2 times were modeled per voxel and analyzed within four GM-regions-of-interest (ROI). The intraclass correlation coefficients (ICCs) of the quantified ASL-T 2 varied across brain regions. When averaged across subjects and postlabeling delays (PLDs), the ICCs ranged from reasonable values in parietal regions (ICC = 0.56) to smaller values in frontal regions (ICC = 0.36). Corresponding subject-averaged within-subject coefficients of variation (WSCVs) showed good test–retest measurement precision (WSCV PLD≤ 0.14 for all PLDs), but more pronounced inter-subject variance. Reliability and precision of quantified ASL-T 2 were region-, PLD- and subject-specific, showing fair to robust results in occipital, parietal and temporal ROIs. The results give rise to consider the method for future cerebral studies, where variable perfusion or altered T 2 times are suspected.