Effects of Acquisition Parameter Modifications and Field Strength on the Reproducibility of Brain Perfusion Measurements Using Arterial Spin-Labeling

Background and Purpose: To encourage clinical adoption of ASL-based perfusion MRI, we investigated the reproducibility of CBF measurements and the effects of protocol variations at clinical field strengths in adult participants with a broad age range. This study increases the knowledge on the tolerance to variations in scan parameters compared to the recommended ASL implementation.
Materials and Methods: Thirty-four volunteers (mean age 57.8±17.0y, range 22-80y) underwent two separate scan sessions on clinical field strengths (1.5T and 3T, single vendor), using a fifteen channel head coil. Both sessions contained repeated 3D and 2D pseudo-continuous ASL (pCASL) vendor-recommended protocols, followed by three 3D pCASL scans; two with post-labeling delays (PLD) of 1600ms and 2000ms and one with increased spatial resolution. All pCASL scans were acquired with a single PLD. The effect of scan parameter variations on CBF and spatial coefficient of variation (CoV) was examined, as well as the reproducibility of the recommended protocols, both intrasession (two identical protocols scanned 5 min apart) and intersession (first 2D and 3D protocol of the first and second session).
Results: Intrasession CBF reproducibility was similar across image readouts and field strengths ( coefficient of variation (CV) ranging from 4.0% to 6.9%) and did not show a statistically significant correlation with age. Intersession wsCV ranged from 6.8% to 14.8%, scanning twice at 3T versus mixed 3T-1.5T respectively. At 3T, there is sufficient SNR to increase the spatial resolution of the 3D protocol, causing less mixing of gray matter (GM) and white matter signal, therewith decreasing the bias in GM CBF between 2D and 3D protocols (ΔCBF = 2.49 (p<0.001) ml/100g/min. Changes in PLD resulted in a modest bias (ΔCBF ranging from -3.78 (p<0.001) to 2.83 (p<0.001) ml/100g/min).
Conclusion: Our data shows that ASL imaging is reproducible at both field strengths and does not show a statistically significant correlation with age. Furthermore, 3T offers more tolerance for scan parameter variations and allows for protocol optimizations such that 3D and 2D protocols can be compared.