Early Neurological Deterioration and Hypoperfusion Volume Ratio on Arterial Spin Labeling in Patients with Acute Ischemic Stroke

Abstract

BackgroundArterial spin labeling (ASL) is a magnetic resonance imaging (MRI) technique used to quantify cerebral blood perfusion by labeling blood water as it flows throughout the brain. Hypoperfusion volume ratio (HVR) can be calculated using proportional hypoperfusion volume on ASL-based cerebral blood flow (CBF). This study aimed to explore the relation between HVR and early neurological deterioration (END) in AIS patients. Subjects and MethodsPatients with AIS were recruited consecutively, and ASL and regular MRI scans were performed. HVR was calculated from 1.5 and 2.5s post labeling delay (PLD) ASL-CBF maps. END was defined as ≥2 points increment of NIHSS within 72 hours of stroke onset. Univariate and multivariate analysis were used to evaluate the relation between HVR and END. Receiver operating characteristic (ROC) curves were used to determine the ability of HVR in predicting END. ResultsOf the 52 enrolled patients, 18 (34.5%) were determined with END. In patients with END, the median hypoperfusion volume was 20 mL [Inter Quartile Range)IQR, 6-72.5 mL] at 1.5s PLD, and 11.2 mL (IQR, 5.3-26 mL) at 2.5s PLD; Sixteen (88.9%) patients had HVR ≥50%, and 13 (72.2%) patients hypoperfusion volume at 2.5s PLD ASL were greater than diffusion-weighted imaging (DWI) infarct volume. In patients without END, median hypoperfusion volume was 7 mL (IQR, 4-30 mL) at 1.5s PLD, and 4 mL (IQR, 1.5-8.5 mL) at 2.5s; Eleven (32.4%) patients had HVR ≥50%, and 10 (29.4%) patients hypoperfusion volume at 2.5s PLD ASL were greater than DWI infarct volume. The proportion of HVR ≥50% and hypoperfusion volume >DWI infarct volume were more frequent in patients with END than patients without (all P<0.001). After adjusted for age, admission NIHSS, proportion of hypoperfusion volume > DWI infarct and arterial transit artifact (ATA) by logistic regression analysis, HVR ≥50% (OR=13.1, P=0.003) was an independent risk factor for END. ROC analysis demonstrated that the HVR could predict END with an area under the curve of 0.794 (P=0.001). ConclusionsHVR obtained from the 1.5 and 2.5s PLD ASL may be a useful predictor of END in AIS. The value of HVR may be a marker for hemodynamic impairments.