MRI May Widen Therapeutic Window for Ischemic Stroke

STANFORD, Calif., Nov. 3 -- For patients with an acute ischemic stroke who arrive at emergency department too late for thrombolysis, an MRI scan may show whether there is still room for benefit even though the three-hour therapy window has shut.

Among 74 consecutive patients with acute ischemic strokes who arrived at the ED more than three hours after the onset of symptoms, MRI scans revealed that some had salvageable tissue and reversible symptoms, reported Gregory W. Albers, M.D., of Stanford, and colleagues in the DEFUSE trial.

On the other hand, some patients with similar clinical presentations had severe damage that would not be helped by reperfusion, and tPA would offer no clinical benefit, the investigators reported in the online edition of Annals of Neurology.

"The DEFUSE results demonstrate that acute MRI scans are feasible in acute ischemic stroke patients at selected centers, and that MRI findings appear to differentiate patient subgroups that will benefit from therapies that lead to early reperfusion from those who are unlikely to benefit or may be harmed," they wrote.

The DEFUSE (Diffusion and Perfusion Imaging Evaluation For Understanding Stroke Evolution) study examined the question of whether specific pre-determined patterns or profiles seen on MRI -- specifically a perfusion-diffusion mismatch -- could predict clinical responses to reperfusion therapy in patients treated between three and six hours after the onset of symptoms of acute ischemic stroke.

"A perfusion/diffusion mismatch has been proposed as a surrogate for the ischemic penumbra, and patients with a mismatch are hypothesized to be more likely to benefit from early reperfusion than patients with other MRI patterns," the authors wrote.

In the prospective, multicenter study, the Stanford researchers and colleagues in other stroke centers in the United States, Canada, and Belgium looked at 74 consecutive stroke patients.

Inclusion criteria included a clinical diagnosis of ischemic stroke with a NIH Stroke Scale Score of more than 5, and ability to be treated with tPA within three to six hours of symptom onset.

Patients were excluded if they were comatose, had rapid improvement of symptoms, stroke within the last six week, premorbid modified Rankin Scale score of 3 or higher, seizure at symptom onset, previous intracranial hemorrhage, evidence of acute hemorrhage or clearly identifiable hypodensity involving more than one third of the middle cerebral artery territory on the baseline non-contrast brain CT.

Patients who agreed to participate underwent an MRI scan of the brain after a baseline CT scan. Regardless of the MRI results, the patients were then treated with 0.9 mg/kg tPA, delivered as a 10% IV bolus over 1 minute, followed by the remainder of the dose over 60 minutes. The therapy was delivered as soon as possible after the MRI scan, but not later than six hours after symptom onset.

The patients underwent repeat MRI scans three to six hours after initiation of thrombolytic therapy and at day 30. Patients who had neurological deterioration during their hospital CT underwent an additional CT and/or MRI scan to document brain hemorrhage.

For the entire patient population, the association between early reperfusion and desirable clinical outcomes was modest and not statistically significant.

They used the baseline MRI profiles to categorize patients into subgroups, and compared clinical responses on the basis of whether early reperfusion had been achieved.

In an interim analysis, they determined that a specific profile they called the "malignant" profile was associated with severe intracranial hemorrhage and poor outcome after reperfusion in three patients.

The malignant profile was characterized by a large diffusion-weighted imaging lesion volume, and/or a large perfusion-weight imaging volume with long delays on the Tmax map.

Patients who exhibited a perfusion/diffusion mismatch, but did not have the Malignant profile, were deemed to have the "target" profile; that is, they were likely to be good candidates for thrombolytic therapy.

Among all patients with a perfusion/diffusion mismatch, early reperfusion was associated with significantly increased odds of achieving a favorable clinical response, with an odds ratio of 5.4 (P=0.039).

The response was even better among patients with the target mismatch profile, with an odds ratio of 8.7 (P=0.011). The target mismatch profile was present in 33 of the mismatch patients (48% of patients with interpretable baseline perfusion-weighted imaging and 45% of all enrolled patients). Early reperfusion could be assessed in 31 of these patients (two had technically inadequate follow-up perfusion-weighted imaging).

"Sixty-seven percent of these patients had a major improvement in neurological function," Dr. Albers said. "This often meant the difference between inability to speak with paralysis of one side of the body and a complete, or nearly complete, recovery."

In contrast, patients with the no-mismatch profile did not appear to benefit from early reperfusion, and patients with the malignant profile were more likely to have fatal intracranial hemorrhage.

The authors cautioned that their study was designed to determine whether MRI could reliably predict which patients could benefit from thrombolytic therapy, and without a control group was not intended as a study of the efficacy of tPA in acute stroke.

"By having this additional information available, we should be able to make a much more sophisticated decision about which therapies are optimal for an individual patient, especially as you get into the longer time windows," Dr. Albers said.

He is working with a team of radiologists, physicists and programmers to develop analysis software that will provide real-time prognostic information for patients with acute stroke in a variety of urgent care settings.