Predictor value of pCASL maps on final core in acute ischemic stroke: an observational single‑center study

The three-compartment model of acute ischemic stroke (AIS) includes non-viable tissue (NVT), tissue at risk (TAR), and benign oligoemia. After assessing a multimodal computed tomography (CT)-based protocol, we assessed a hyperacute stroke imaging protocol with magnetic resonance imaging (MRI), aiming to better characterize the reliability of the two methods in predicting follow-up infarction.

We retrospectively reviewed 627 consecutive AIS patients evaluated for the selection for reperfusive treatments at comprehensive stroke center. We employed an MRI-based protocol for proper patient selection, including pseudocontinuous arterial spin labeling (pCASL) and diffusion-weighted imaging (DWI). To assess prognostic accuracy of pCASL maps in predicting final infact, we assumed the best correlation between NVT and final infarct in patients with favorable clinico-radiological outcomes. On the other hand, the TAR should better correlate with final infarct in untreated patients and in treatment failure.

We recruited 349 patients undergoing MRI-based protocol with DWI and pCASL, showing diagnostic accuracy of 84.8%. In anterior circulation perfusion deficit, DWI was highly reliable for NVT in patients with favorable clinical-instrumental outcomes. PCASL tended to overestimate TAR in patients with unfavorable outcome, but the rate of complete overlap with the final infarct was high. The previously assessed CT-based protocol showed a lower prognostic accuracy, as the CT-perfusion using time-to-maximum overestimated both the NVT and the TAR.

We found a highly reliable prognostic accuracy for DWI-pCASL. Even if pCASL proved reliable to identify the area of total hypoperfusion, there was an overestimation of TAR in many cases, with a tendency to incorporate the benign oligoemia.