Marc Fisher, MD (pictured below)

Department of Neurology, University of Massachusetts School of Medicine

Gregory W. Albers, MD

Stanford University School of Medicine

Acute ischemic stroke therapy with intravenous (i.v.) tPA or intra-arterial devices is most effective when initiated early after stroke onset because the goal of these therapies is to salvage a portion of the ischemic territory at risk of infarction, i.e. that ischemic penumbra to reduce infarct size and improve outcome. Advanced imaging, especially with diffusion and perfusion MRI afford the opportunity to identify the extent of the ischemic penumbra in individual patients and therefore to extend the therapeutic time window of tPA and devices. With DWI/PWI the ischemic penumbra can be approximated as the ischemic region that is abnormal on PWI but normal on DWI, a mismatch region. Over the past few years, increasing evidence to support the utility of diffusion/perfusion (DWI/PWI) MRI for patient selection has emerged. In the first DEFUSE study not only was it shown that patients with a target DWI/PWI mismatch benefit from i.v. tPA in the 3-6 hour time window, but that a malignant pattern with a large > 100cc DWI lesion and/or large severe baseline PWI region were at substantial risk for intracerebral hemorrhage with i.v. tPA use. The EPITHET study employed pretreatment DWI/PWI prior to randomization to either i.v. tPA or placebo in the 3-6 hour time window. Patients treated with i.v. tPA and a DWI/PWI mismatch >20% tended to have smaller ischemic lesions over time and an improved clinical outcome. The study provided important lessons for the design of future MRI-based treatment trials including; the mismatch should be larger than 20% for patient selection, the perfusion lesion should be identified by a Tmax value of 5-6 seconds and a sample size of approximately 200 patients per group would be adequate to detect a clinically significant treatment effect. The recently published DEFUSE 2 study provided additional information about the utility of MRI DWI/PWI mismatch in identifying patients likely or unlikely to benefit from endovascular device treatment up to 12 hours from stroke onset. In DEFUSE 2 patients with a target DWI/PWI mismatch who reperfused after device therapy had a 56% rate of favorable outcome at 90 days, while patients who did not reperfuse only achieved a 31% favorable outcome rate, odds ratio 4.0 (95% confidence interval 1.3-12.). In patients without a target mismatch, the favorable outcome rates were 25% and 22% respectively. DEFUSE 2 did not have an untreated control group, but the demonstration of a significantly greater favorable outcome rate in reperfused target mismatch patients as compared to nonreperfused target mismatch patients and the fact thatreperfusion was not associated with clinical benefits in patients without a mismatch, supports the hypothesis that the DWI/PWI mismatch approach to identifying the ischemic penumbra is useful for predicting which patients will respond to therapy, especially at a relatively late treatment time point.

To conclusively prove that a DWI/PWI mismatch identifies stroke patients who will benefit from i.v. tPA or device treatment randomized, double-blinded, placebo controlled trials are needed. The EXTEND trial is underway and is randomizing stroke patients with a 20% or greater mismatch on MRI or CT perfusion in the 4.5-9 hour time window to i.v. tPA or placebo. Recently both the TREVO and SOLITAIRE devices demonstrated very high rates of successful endovascular recanalization and improved outcome as compared to the older MERCI device up to 8 hours after stroke onset. Both of these newer devices will now apparently be compared to best medical therapy at delayed time windows after stroke onset with patient inclusion to include mismatch on penumbral imaging. If one or more of these current trials demonstrate a significantly better clinical outcome with treatment, then penumbral imaging, especially DWI/PWI, will move to the forefront in patient treatment decision paradigms, especially beyond 3-4.5 hours after stroke onset.