Effect of Delayed Transit Time on Arterial Spin Labeling Correlation With Dynamic Susceptibility Contrast Perfusion Magnetic Resonance in Moyamoya Disease

Objectives: Because arterial spin labeling (ASL) is completely noninvasive and provides absolute cerebral blood flow (CBF) information within a brief period, the technique has been increasingly used for patients with acute or chronic cerebrovascular disease. However, the effect of delayed transit time on ASL can generate errors in the quantitative estimation of CBF using ASL. Furthermore, in the clinical setting, in which transit time is uncertain, the variability of the transit time in patients reduces the validity of CBF on ASL images. Therefore, we evaluated the effect of delayed transit time on ASL images compared with dynamic susceptibility contrast (DSC) perfusion magnetic resonance (MR) in patients with moyamoya disease. Materials and Methods: Arterial spin labeling and DSC perfusion MR images were acquired in 54 patients with moyamoya disease. Vascular territory and anatomical structure-based regions of interest (ROIs) were applied to the CBF and time-to-peak (TTP) maps from DSC and a CBF map using ASL. The change of the correlation coefficient (r) between normalized CBFs (nCBFs) from DSC and ASL was evaluated with categorization by the TTP. In addition, the dependence of the difference between the nCBF values from DSC and ASL on the TTP obtained using DSC was also analyzed. Results: The nCBF values from DSC and ASL were strongly correlated (r = 0.877 and r = 0.867 for the internal carotid artery (ICA) and middle cerebral artery territory-based ROIs, respectively; P < 0.0002 for both; r = 0.783 for the anatomical structure-based ROIs; P < 0.0084). However, correlations between nCBFs from DSC and ASL tended to be weaker when the TTP increased, with recovery when the TTP was extremely delayed (>25 seconds). The TTP delay had a positive effect on the difference between the nCBF values from the DSC and ASL for the ICA territory-based and anatomical structure-based ROIs (standardized coefficients, 0.224 for the ICA territory-based ROIs; P = 0.0410; 0.189 for the anatomical structure-based ROIs; P < 0.0084). Conclusions: Our results demonstrate that the correlation between the CBF values from the ASL and DSC tends to be weaker when the transit time is more delayed, with the restoration of the strength of the correlation when the TTP is extremely delayed (>25 seconds). Understanding the effect of delayed transit time on the CBF from ASL perfusion MR in a clinical setting would facilitate the proper interpretation of ASL images.