MRIs could save stroke victims from brain damage
The MRI scan [left] is of a 77-year-old woman nearly six hours after the onset of stroke symptoms. The brightly colored areas show the sections of her brain that are not receiving adequate blood flow. Four-and-a-half hours later, after administering the medication tPA , the second scan displays less color, suggesting that the drug had remedied the problem. The patient had nearly a complete recovery over the next month.
A stroke victim arrives in the emergency room, and within minutes, the doctor must make a decision: Should drugs be administered to open up the blocked blood vessel and prevent further brain damage? Or is this patient at high risk for suffering a brain hemorrhage if the blocked vessel is opened?
Greg Albers, MD, the director of the Stanford Stroke Center, and his team report in the November issue of Annals of Neurology that new magnetic resonance imaging techniques can discriminate between stroke patients who are likely to benefit from a stroke medication—even when administered beyond the currently approved three-hour time window—and those for whom treatment is unlikely to be beneficial and may cause harm.
For years, Albers, professor of neurology and neurological sciences, has been using new MRI techniques to visualize the damage from stroke while it is actually happening. The goal was to differentiate brain tissue that is potentially salvageable from tissue that is already irreversibly injured from a stroke. As his group accumulated MRI scans of stroke patients, they noticed patterns in the images that seemed to identify which patients were most likely to benefit from opening up blocked blood vessels.
"One of the criticisms was that these detailed brain images looked beautiful and interesting, but there was no proof that they should be used to influence treatment or that they would result in improved outcomes," said Albers. "How do you know that these MRI patterns can predict whether the therapy is likely to be beneficial?"
To answer these questions, Albers and his colleagues designed a study to see if obtaining an MRI profile from stroke patients before beginning treatment could identify which patients would benefit from clot-dissolving drugs administered between three and six hours after stroke onset and which patients were unlikely to benefit, or potentially might be harmed. Albers was the principle investigator of the three-year-long study, which was funded by National Institutes of Health and included sites in the United States, Canada and Belgium.
Strokes result from decreased blood flow to an area of the brain. Once brain cells are deprived of oxygen and nutrients carried in the blood, they begin to malfunction. Symptoms of a stroke include weakness, paralysis or numbness on one side of the body, problems speaking or loss of vision.
About 85 percent of strokes are caused by clots blocking blood vessels in the neck or brain. In 1996, the clot-busting drug tPA was approved by the U.S. Food and Drug Administration. The drug can restore blood flow to regions of the brain injured by stroke. The study that led to its approval indicated that tPA should be used only in patients who were treated within three hours of the onset of stroke symptoms and who also had a CT scan indicating that there was no bleeding in the brain.
Despite the need for early intervention, less than a quarter of stroke patients make it to a hospital within three hours. Albers has been trying to pin down what factors might allow the tPA treatment window to remain open longer.
In the late 1990s, Albers was one of the principal investigators of a study that attempted to extend the time window for this therapy to six hours based on the CT scan approach. Unfortunately, this study failed. The researchers suspected that this failure occurred because CT scans were unable to differentiate patients who could benefit at three to six hours from those who did not benefit.
"CT scans do not demonstrate how much brain tissue is still salvageable and how much is irreversibly injured," said Albers. "Therefore, with only the CT image it is difficult to know for any given patient whether opening the blocked vessel is going to be a good thing, a bad thing or have no effect. If there is already a large area of severely injured tissue, opening up a blocked vessel can result in serious, even fatal, brain hemorrhage."
Standard CT scans can differentiate strokes caused by ruptured blood vessels from ones caused by blocked vessels, but the location and extent of the brain injury is typically not evident for at least eight hours after symptoms begin. An MRI can immediately demonstrate areas of brain injury, outline areas of critically reduced blood flow and clarify which blood vessel is blocked. These subtleties can determine whether opening the vessel is likely to be beneficial, Albers said.
Patients with radically different situations in their brains can have identical symptoms when brought into the emergency room. Two study participants, for example, were both unable to speak and were paralyzed on their right side. Although their CT scans looked the same, their MRI patterns were completely different; one revealed minimal irreversible injury but considerable tissue at risk while the other revealed that extensive severe injury had already occurred.
The study's original hypothesis was that that MRI patterns would allow the patients to be divided into subgroups based on how much brain tissue was already damaged and how much had insufficient blood flow. Patients whose scans indicated substantial areas of insufficient blood flow but little permanent damage were predicted to benefit most from tPA administration. The drug does not repair existing damage.
The team enrolled 74 consecutive stroke cases that met various criteria, including having the treatment administered between three and six hours of symptoms onset. The team obtained MRI scans for each patient immediately before, and approximately four hours after, administration of intravenous tPA. No other study has performed these advanced MRI techniques immediately before and so soon after treatment.
In this study, the MRI results were not used to make treatment decisions. The scans were analyzed later at Stanford and the pre-treatment MRI patterns were compared to how the patients fared three months later.
During the trial, the researchers discovered a significant twist to their hypothesis. Three patients developed fatal cerebral hemorrhage after tPA treatment. All three had a unique MRI pattern prior to treatment and the successful opening of their blocked blood vessel. This finding led the investigators to define a new profile that predicts a high risk of dangerous bleeding in the brain following tPA therapy.
Still, for those patients who had a pattern indicating that a favorable response to tPA was likely, the benefits of opening the blocked vessel were dramatic.
"Sixty-seven percent of these patients had a major improvement in neurological function," 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."
This is the first study to show that certain MRI patterns predict a very good response upon opening of the blood vessel and that for other patterns, opening the vessel may have no beneficial effect or can even cause harm. A 20-minute MRI scan has the potential to indicate who is likely to benefit and who is not. "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," said Albers.
Albers is currently working with a team of radiologists, physicists and programmers to optimize the software to analyze the MRI blood flow scans. The goal is to broaden the technique so that it can eventually be used in any hospital with an MRI machine, delivering prognostic information in real time, so that "we won't have to treat stroke by a stopwatch any longer," said Albers.
In addition to Albers, there were nine researchers from Stanford: Roland Bammer, PhD, assistant professor of radiology; Anna Finley, MD, clinical assistant professor of neurology; Wataru Kakuda, MD, visiting research associate in neurology; Stephanie Kemp, social science research assistant in neurology; Maarten Lansberg, MD, instructor in neurology; Michael Marks, MD, professor of radiology and, by courtesy, of neurosurgery and chief of interventional neuroradiology; Michael Moseley, PhD, professor of radiology; Neil Schwartz, MD, professor of radiology, and Christine Wijman, MD, assistant professor of neurology and neurological sciences and, by courtesy, of neurosurgery.