Estimation of hypoperfused tissue volume in large vessel occlusions: pseudo-continuous arterial spin labeling versus dynamic susceptibility contrast perfusion-weighted imaging.

Abstract

Background: Currently, the selection of patients with acute anterior large vessel occlusions (LVOs) for endovascular thrombectomy (EVT) is primarily based on dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI) or computed tomography (CT) perfusion imaging. This study investigated the consistency between hypoperfused tissue (HPT) (time to maximum >6 s, Tmax >6 s) volumes estimated by corrected and uncorrected multidelay pseudo-continuous arterial spin labeling (pCASL) and DSC-PWI in patients with anterior LVOs and also evaluated the diagnostic performances in selecting patients with acute LVOs for EVT.

Methods: This retrospective study enrolled patients with acute (n=108) and symptomatic chronic (n=90) LVOs. Shapiro-Wilk tests and receiver operating characteristic (ROC) analyses were used. Intraclass correlation coefficient (ICC) compared the consistency of HPT volume calculated by DSC-PWI and multidelay pCASL.

Results: Multidelay pCASL with different thresholds in acute LVOs were 128.8 [interquartile range (IQR), 76.2-181.1] mL in uncorrected relative cerebral blood flow (rCBF) <40%, 84.1 (IQR, 36.8-133.9) mL in uncorrected CBF <20 mL·100 g^-1·min^-1 , and 74.4 (IQR, 26.2-118.0) mL in corrected CBF <20 mL·100 g^-1·min^-1, which were comparable to the volume of 69.5 (IQR, 20.0-121.4) mL automatically determined by Tmax >6 s in DSC-PWI, and showed substantial consistency after correction (ICC =0.742). Multidelay pCASL with different thresholds in symptomatic chronic LVOs was 78.3 (IQR, 53.5-129.4) mL, 59.8 (IQR, 16.6-98.5) mL and 36.4 (IQR, 10.1-85.3) mL, which were comparable to the volume of 0 (IQR, 0-36.4) mL in DSC-PWI, and showed substantial consistency after correction (ICC =0.617). Using DEFUSE 3 as the reference standard, the CBF corrected by arterial transit time (ATT) showed good performance in selecting patients for EVT (area under the curve 0.804, 95% confidence interval: 0.717-0.891).

Conclusions: The volume of HPT defined by corrected CBF <20 mL·100 g^-1·min^-1 is consistent with that of DSC-PWI in acute and chronic symptomatic LVOs patients. Multidelay pCASL adjusted by ATT is more applicable to clinical routine.