Arterial Spin Labeling at 3.0 Tesla in Subacute Ischemia

Abstract

Arterial spin labeling (ASL) is a promising but clinically not established non-invasive method to assess cerebral perfusion. The purpose of this study was to compare perfusion imaging with pulsed ASL (pASL) to conventional dynamic susceptibility contrast (DSC) perfusion-weighted imaging (PWL) using commercially available equipment and postprocessing (3.0Tesla, 32-channel head coil) in patients with subacute ischemia. The pASL and DSC-PWI techniques were compared in 15 patients with subacute ischemia (age 49-88 years, 6 females and 9 males, time from onset to scan 4-161h). Image inhomogeneity was assessed with the non-uniformity index. Image quality, delineation of hypoperfusion and degree of hypoperfusion were rated by two readers using a 5-scale grading system. The volume of hypoperfusion was quantified planimetrically. Image quality and image inhomogeneity were superior in DSC time-to-peak (TTP) compared to pASL cerebral brain flow (CBF; both p < 0.05). The delineation of hypoperfusion was better in DSC-TTP (p < 0.05) and the hypoperfusion was graded as more severe in DSC-TTP (p < 0.05). The volume of hypoperfusion did not differ between pASL-CBF and DSC-TTP, however, in pASL-CBF five cases with small infarctions (lacunar and pontine) were false negative compared to DSC-relative CBF. The mismatch frequency was lower in pASL (13%) than in DSC-rCBF (20%) and DSC-TTP (47%). Using a commercially available sequence and a 32-channel head coil at 3.0 Tesla pASL-CBF is feasible but limited compared to DSC-PWI in the assessment of ischemic stroke. In its present form pASL has a reserve role in clinical practice for situations when gadolinium diethylenetriaminepentaacetic acid (Gd-DTPA) is contraindicated.