乌鸦传媒

Assessment with two types of MRI techniques showed potential for identifying stroke patients who are within the time window for thrombolytic therapy, data from a multicenter study showed.

A positive result with diffusion-weighted MRI (DWI) coupled with a negative fluid-attenuated inversion recovery (FLAIR) MRI study had a 62% sensitivity and 78% specificity for predicting whether patients were within the 4.5-hour therapeutic window.

The findings add to a growing volume of evidence that DWI-FLAIR mismatch could play a major role in assessing patients with ischemic stroke for thrombolytic therapy, as reported online in The Lancet Neurology.

"Patients with an acute ischemic lesion detected with DWI but not with FLAIR imaging are likely to be within a time window for which thrombolysis is safe and effective," Gotz Thomalla, MD, of University Medical Center Hamburg-Eppendorf in Hamburg, Germany, and co-authors wrote in conclusion.

They cautioned, however, that their study "does not provide evidence for efficacy and safety of MRI-based thrombolysis inpatients with unknown time of symptom onset," and suggested further study to determine that, "with DWI and FLAIR MRI used to enroll patients."

Current criteria for administration of intravenous thrombolysis in stroke exclude patients who have an unknown time of onset. Given that an estimated 25% of ischemic strokes occur during sleep, a substantial population of patients who might otherwise qualify for thrombolytic therapy are excluded because of unknown time of stroke onset, the authors wrote in their introduction.

Moreover, clinical and radiographic findings suggest that patients waking with stroke symptoms have a presentation similar to those with a known symptom onset of three to six hours, they continued.

Use of multiple MRI techniques in combination may have potential to identify stroke patients who are likely to benefit from thrombolysis. For example, DWI can detect changes in water diffusion within three minutes of ischemia onset, and FLAIR can detect a net increase in water within one to four hours after symptom onset.

Although sensitive to subacute ischemic brain lesions, FLAIR usually cannot detect ischemic lesions within the first few hours, the authors wrote.

The mismatch in visibility has led to speculation that a combination of DWI and FLAIR imaging might improve identification of stroke patients who are likely to benefit from thrombolysis, the authors wrote.

A single-center study showed that DWI-FLAIR mismatch had a specificity and positive predictive value exceeding 90% for identification of patients within three hours of symptom onset. However, the technique had poor sensitivity and negative predictive value (Ann Neurol 2009; 65: 724-732).

Several other single-center investigators have provided additional support for use of DWI-FLAIR mismatch to evaluate stroke patients with unknown time of onset, but all were retrospective.

But before a randomized controlled trial could be done, "the diagnostic accuracy of this approach needs to be confirmed in a large multicenter study," Thomalla and co-authors noted.

In an effort to collect that needed data, they conducted a retrospective multicenter cohort study involving stroke patients who had a known time of symptom onset. All patients underwent MRI with DWI and FLAIR within 12 hours of symptom onset. Standardized MRI protocols were used at all participating centers.

Two experienced neurologists performed image analysis, and if they disagreed about patient status, they reviewed the findings with a third neurologist and reached a consensus as to whether a patient had symptom onset within a 4.5-hour time window.

The final analysis included 543 patients who had a mean age of 66 and median NIH stroke score of 8.

DWI revealed ischemic stroke lesions in 516 (95%) patients, whereas 271 (50%) patients had lesions that were visible on FLAIR.

DWI-FLAIR mismatch identified patients who were within the 4.5-hour therapeutic window with:

• 62% sensitivity
• 78% specificity
• 83% positive predictive value
• 54% negative predictive value

Multivariate analysis of factors associated with lesion visibility on FLAIR imaging revealed three independent predictors: longer time to MRI (P<0.0001), younger age (P=0.0009), and larger lesion volume on DWI (P=0.0226).

"This study sets the stage for optimization and validation of the FLAIR-DWI mismatch biomarker," Michael D. Hill, MD, and Richard Frayne, MD, of the University of Calgary in Canada, wrote in a related commentary.

"Validation would allow this biomarker to be used as a participant selection method for randomized trials of thrombolytic stroke therapy in patients with stroke on awakening or with unwitnessed stroke onset.

"Such trials are on the drawing board and it will be these kinds of imaging biomarkers -- which are simple, practical, and easily implementable at many sites -- that will allow relevant candidates to be selected for enrollment in these trials," Hill wrote.

Comment