MRI software helps to detect damage from stroke

By adapting a technique in oncology, physicians used standard MRI scans to more accurately measure damage to the blood brain barrier in stroke patients. If further validated, the results may offer a method to identify which patients might benefit or be harmed if treated with tissue plasminogen activator (tPA). The study was published online Dec. 20 in PLOS ONE.

Richard Leigh, MD, of the neurology department of Johns Hopkins University School of Medicine in Baltimore, and colleagues noted that dynamic susceptibility contrast (DSC) MRI can be used to measure cerebral blood volume in brain tumors and blood flow to the brain in stroke patients. In oncology, it is a tool for determining tumor grade and in stroke it may help to evaluate treatment options for patients.

“However, the approach used in the brain tumor literature, which assumes symmetric perfusion of the brain, can be subject to error when applied to patients with perfusion deficits, such as acute stroke patients,” Leigh and colleagues wrote. “The delay in contrast delivery to areas of hypoperfusion makes calculation of K2 [a measure of permeability used in oncologic applications] inaccurate. Thus we developed a technique that applies an arrival time correction (ATC) prior to calculation of K2.”

They designed their proof-of-concept study to compare permeability measurements made with and without ATC in stroke patients with known blood brain barrier disruption to determine if ATC improved imaging’s ability to identify contrast leakage. They retrospectively reviewed MRIs of stroke patients admitted at one center between Aug. 1, 2010, and March 31, 2011, for evidence of blood brain barrier disruption on T1 post-contrast imaging. To be included in the analysis, patients also needed to have a successful DSC MRI and acute stroke on perfusion weighted imaging.

Nine patients and 13 MRI scans qualified. Researchers performed a receiver operating characteristic (ROC) analysis to compare detection of blood brain barrier damage before and after ATC. They found that ATC improved the ability to detect blood brain barrier damage. The area under the curve improved from 0.53 to 0.70; sensitivity from 0.51 to 0.67 and specificity from 0.57 to 0.66. Visual inspection of the ROC curve showed the uncorrected analysis performed worse than random guess at some thresholds.

“Our study confirms that DSC permeability imaging is improved using this correction,” they wrote. They added that while other approaches had used DSC MRI in acute stroke, none had applied ATC before calculating permeability, and that permeability imaging “had not found its way into clinical practice.”

"A better characterization of blood-brain barrier damage opens the door to new approaches to treating stroke patients," Leigh said in a release. He described the technique as a tool to tailor patient care. "Rather than lumping everyone together, we can figure out—on a case-by-case basis—who should and who shouldn't get which treatment. In the long run, we can increase the number of patients we can help and decrease the number who have complications."

The authors wrote that the software needed to be further tested in more studies before the approach could become part of clinical practice. They currently are investigating its use in a larger group of stroke patients.