There is a “place” for MR imaging in acute stroke

Gregory Albers 2
Gregory Albers

Magnetic resonance (MR) imaging may have advantages over other imaging modalities in later time windows when it comes to selecting patients for endovascular treatment of acute ischaemic stroke. That is according to Gregory Albers (Stanford University Medical Center, Stanford, USA). He made this assessment during a presentation at the International Stroke Conference (ISC; 17–19 February, Los Angeles, USA).

Making the case for MR imaging over computed tomography (CT), Albers explained that in the SWIFT PRIME trial, patients could be selected using either MR imaging or CT. While the majority of patients were selected using CT, of the patients selected using MR imaging in SWIFT PRIME, 63% had a good outcome in the endovascular arm selected by MR, and 33% in the control group, with a 30% difference.

Further, Albers reported, in the Rankin shift primary analysis of SWIFT PRIME, in the MR selected subgroup (34 patients) the control group had mRS (modified Rankin Scale) 4 at 90 days, and the thrombectomy group had mRS of 2 (p=0.022). Infarct volume at 27 hours was 61ml in the control group and 24ml in the thrombectomy group (p=0.052). There was no difference in CT perfusion versus MR group in age, NIHSS (National Institutes of Health Stroke Scale), or door to randomisation time (median 67 min vs. 68.5 min). The door to randomisation was only 90 seconds longer in these patients.
He explained that this can be attributed to the fact that many of these patients were “drip and ship” or transfer patients—therefore, the hospitals were aware of them and were able to have the MR ready upon their arrival.

“Why choose MR imaging over CT? Diffusion weighted imaging is the gold standard for estimating ischaemic core, it may have some greater advantages in later time windows: it always gives you full brain coverage and it detects smaller lesions. So if your patient comes in and their hemiplegia is because of a medial medullary infarct, which technique is going to show you that? Not the CT,” Albers maintained.

The caveat however, Albers said, is that MR imaging should not delay acute stroke therapy. He acknowledged that this is likely the reason most stroke teams do not use it because it would delay acute stroke therapies, but reiterated that if the hospital is prepared for the patient and has a system that can do the MR, the MR can be done very quickly.

In the recently initiated DEFUSE 3 (a prospective, randomised, multicentre trial of patients with an acute ischaemic stroke who can be treated between six and 16 hours of stroke onset) he explained, there is a six minute MR imaging protocol with 45 seconds of automated processing, “so if you have scanner availability, you should be able to get those images quickly”.

“CT perfusion has some limitations for core. If you had bad ischaemia a few hours ago and now it is resolved, you will not see the core on CT perfusion, but you will see it on diffusion weighted imaging. The reason is you are using cerebral blood volume or cerebral blood flow to determine your core on CT perfusion; your core is going to give you an answer for what the cerebral blood volume or the cerebral blood flow is at the moment of your scan, it gives you no history. So if you had critical ischaemia a couple hours ago, you will not see it—after reperfusion the cerebral blood flow is normal, after reperfusion the cerebral blood volume is normal, maybe even elevated, so we are still at the advantage with diffusion weighted imaging in that situation,” Albers argued.

At his own centre in Stanford, the acute stroke imaging protocol is rapid MR imaging for acute stroke transfers. If they are told that the patient is coming, the MR technician is pre-notified, the MR is then open by the time the patient arrives, and they are taken straight down to the MR, and with the six minute MR protocol, seven minutes after the patient gets onto the scanner, the team gets the results.

“But I must admit that we do not do that when they roll into the emergency room because if we have very little pre-notification, even at Stanford, we cannot get an instant MRI scan. We can get a CT much faster, so our CT/CTA/CT perfusion rapid processing is two minutes,” Albers conceded.

Another possible advantage for MR, he added, is the late time window.

“Time is brain, that is our mantra, and what it says is that when you reperfuse a patient at eight hours, the good outcome rate is not terrific. In the recent studies 30% of the patients had a good outcome if the reperfusion occurred at eight hours. Think about this, what if we had a way to identify that 30% and that was the only 30% that you took to your cath lab, what would your good outcome rate be at eight hours? It would be pretty good,” Albers pointed out.

As for the way forward, Albers maintained that, at least in the later time window, there is a need for sophisticated imaging. The following trials are currently using a combination of CT perfusion and/or MR perfusion and will help to determine if endovascular therapy improves functional outcome in stroke patients with favourable clinical and imaging characteristics: DEFUSE 3 (6–16 hours), DAWN (6–24 hours), POSITIVE (6–12 hours), EXTEND (4.5–9 hours) and ECASS 4 (4.5–9 hours) is using MR only.