Stanford University Home

Stanford News Archive

Stanford Report, April 26, 2000

New MRI technique pinpoints oxygen-starved areas in stroke patients' brains  

BY KRISTA CONGER

Magnetic resonance imaging (MRI) allows physicians to see inside the body, looking through bone to diagnose abnormalities in the soft tissues including the brain. Now a study by Stanford physicians indicates that a technique that uses the same equipment and technology is even more sensitive than standard MRI in pinpointing oxygen-starved areas in the brains of stroke patients. Almost half of the new extra-sensitive scans picked up lesions that were undetectable by conventional methods, and in some cases the findings radically changed the course of therapy for the patient.

"We looked at three categories of findings that could be clinically useful: location, number and age of the lesions," said Gregory Albers, MD, professor of neurology and lead author of the study. "Almost 50 percent of the time we found something potentially important that we couldn't see on the conventional MRI," he said. The results are published in the April 25 issue of the journal Neurology.

It's important for physicians to accurately determine the location and number of stroke-damaged brain areas before they begin treatment. A misdiagnosis might steer a surgeon to the incorrect blood vessel, or obscure the true cause of the symptoms. Many small, scattered lesions can indicate that blood clots are originating from a source outside the brain, often due to cardiac problems. Treatment to prevent these traveling clots is different from that used to prevent clots originating in the blood vessels of the neck or the brain.

To make an accurate diagnosis and begin the appropriate treatment for a suspected stroke victim, Stanford Stroke Center physicians routinely use MRI to generate a series of "snapshots" of different slices of the patient's brain. The quick, noninvasive scan can help pinpoint the location and nature of the oxygen-starved tissue.

But the conventional MRI, which works by applying a magnetic field to the body and detecting the release of energy absorbed by hydrogen nuclei in bones and tissues, isn't perfect. Often it is difficult to detect a very recent stroke, or to identify particularly small or deep lesions. A standard MRI is also not very good at differentiating between recent damage and necrotic tissue left over from a previous stroke (that may not be responsible for the patient's current symptoms).

The new method, diffusion weighted imaging (DWI), uses the same equipment and scanning procedure as the MRI. However, DWI specifically hones in on the hydrogen-rich water molecules in the brain and measures how free they are to move about, or diffuse. Water in the extracellular space can squeeze around individual neurons, but water trapped inside a cell has much less room to maneuver.

"It's like going to a party where there are only a few people and it's easy to mingle with each other," said Albers, director of the Stanford Stroke Center. "In contrast, when the party gets really crowded it's much more difficult to circulate," he said.

In a DWI, the restricted diffusion of water molecules inside a cell serves as a flag for cells in trouble. When a dying cell is no longer able to maintain proper internal chemical concentrations, water crowding into the cell to compensate for the imbalance shows up on the DWI as very bright areas against a dark background. This "light bulb" effect using DWI makes it easier for physicians to identify the size and location of stroke-damaged tissue compared with conventional MRI.

Previous studies have indicated that DWI is a useful tool for identifying dead or dying cells very soon after the onset of symptoms, but it was unclear how often the technique would be useful for patients who are imaged beyond six hours after stroke symptom onset. Most people don't appear in the emergency room until several hours after their symptoms begin.

In the new study, the patients were first examined by a physician who made a preliminary diagnosis of the suspected location and type of lesion based on their symptoms. Each patient underwent first a conventional MRI, and then a DWI. The time from onset of symptoms until scans varied from six hours to 48 hours, with an average time of 27 hours.

Of the 40 patients who participated in the study, the results of the DWI found that the true symptomatic lesion was located in a different region of the brain than that identified by conventional MRI in seven people, or about 18 percent of the total. Five people, about 13 percent of the total, were found to have multiple new lesions when the conventional MRI had only identified one.

Monitoring the rate of water diffusion is also a good way to determine the age of the lesion. Dead cells disintegrate in about 20 to 30 days, leaving their trapped water free to slosh about in the empty space remaining. The freely diffusing water is easily detectable, and indicates that the lesion is at least three weeks old. In Albers' study, 8 patients, or 20 percent of the total, were found to have only old lesions and no new ones, essentially ruling out the possibility that they had suffered a recent stroke.

All together, the DWI findings changed the diagnosis of 19 of the 40 patients participating in the study.

In addition to Albers' study, another study from the Stanford Stroke Center in which Michael Marks, MD, associate professor of radiology is the senior author, appears in the same issue of Neurology. His study confirms the usefulness of DWI in helping to determine the appropriate treatment for stroke patients within seven hours of symptom onset. Marks compared DWI to computed tomography (CT), another common imaging technique that uses X-rays to visualize the brain.

In this study the researchers compared DWI and CT scans taken within seven hours of symptom onset in 19 patients. DWI was able to correctly pinpoint the location of the lesion in all of the cases, but CT scans were correct in only 42 percent to 63 percent of the patients.

Although neither study analyzed whether the clinical outcomes were different when DWI changed a patient's diagnosis, both urge that further study is warranted to address the issue.

In addition to Albers and Marks, Stanford physicians participating in the two studies include Martin Lansberg, MD, formerly a research assistant at the Stroke Center and now an intern in Baltimore; David Tong, MD, assistant professor of neurology and neurological sciences; Michael O'Brien, formerly a stroke fellow and now in private practice in San Jose; Michael Moseley, PhD, associate professor of radiology; and Christopher Beaulieu, MD, PhD, assistant professor of radiology. SR