Perfusion MR succeeds by evaluating all results

Academic medical practitioners no longer consider first-pass stress and perfusion MRI, wall motion, and delayed enhancement imaging as separate tests. During a plenary session at the SCMR meeting, researchers agreed that these sequences generate the highest accuracy for detecting myocardial infarction and assessing myocardial viability when they are performed and analyzed together. The optimal order for performing these tests and interpretation of the results, however, still generate considerable controversy.

Justifying a comprehensive approach is difficult because most studies have not tested the technique's strengths and weaknesses. Most trials described in the medical literature compare CMR perfusion, function, and delayed enhancement against the results from competing modalities. The trials have shown, for example, that stress-rest perfusion imaging has the potential to outperform thallium or sestamibi SPECT for detecting myocardial ischemia associated with coronary artery disease, said Dr. Igor Klem, a senior cardiac fellow at Duke University. The spatial resolution of perfusion MR is 10 times higher than that of perfusion SPECT, but perfusion MR is technically demanding.

"So is perfusion MR ready for prime time?" Klem asked at the SCMR meeting.

His answer is no, if the wide variability of reported sensitivities and specificities for the test are taken at face value. But accuracy rates close to 90% are possible with careful consideration of how the components of a comprehensive MR myocardial perfusion/viability exam should fit together.

Patients at Duke are referred for coronary angiography after a diagnosis of coronary artery disease by delayed enhancement MRI and stress-rest perfusion imaging. Cine imaging measurement of wall motion is not used because of its low sensitivity and specificity for chronic ischemia, Klem said.

Delayed enhancement, however, is highly accurate for the diagnosis of acute and chronic disease. Its value when interpreted in tandem with stress and rest perfusion was demonstrated in a study Klem presented at the meeting in San Francisco.

The study's 100 consecutive patients had suspected CAD but no prior history of the disease. DE-MRI revealed evidence of infarction in 19 patients whose conditions were confirmed with cardiac catheterization in all but one case. The researchers then turned to the stress perfusion component of the protocol for patients with negative DE-MR exams. These patients were deemed free of significant CAD if their stress CMR exams were negative as well.

A positive stress test, however, led the reader to examine the rest perfusion exam as well. A positive finding there indicated that the abnormalities stemmed from image artifacts, not actual pathology. These patients were diagnosed as disease-free. Negative rest perfusion results in the presence of positive delayed enhancement and stress perfusion tests indicated the presence of CAD, Krem said.

The protocol is diagnostically more powerful than delayed enhancement alone, he said. Its sensitivity, specificity, and accuracy rates for the detection of CAD were 89%, 87%, and 88%, respectively, compared with 84%, 58%, and 68% for delayed enhancement alone.

BRIGHAM AND WOMEN'S APPROACH

Brigham and Women's Hospital has also adopted a comprehensive approach to CMR, although its protocol emphasizes the importance of wall motion data collected with functional MR, especially in interpretation of the studies.

A rationale for the Brigham and Women's model was defined as its CMR researchers examined why stress echocardiography shows low sensitivity and high specificity for the presence of myocardial infarction, while rest-stress SPECT imaging produces high sensitivity and low sensitivity rates for that application. They discovered that the ischemic cascade, the pattern of events beginning with coronary artery occlusion and ending with infarction, caused the mirror-image effect, said Dr. Raymond Y. Kwong, codirector of CMR. As a result, ischemic wall motion abnormalities depicted with stress echo are consistently smaller on a case-by-case basis than perfusion deficits measured with SPECT.

This temporal and spatial discordance led Kwong and his colleagues to conclude that the optimal CMR protocol should include wall motion and perfusion to temper its effect. Stress and rest functional CMR at Brigham and Women's are performed first and can involve up to a 40-mkg administration of gadolinium contrast. Infusion of the contrast is continued for first-pass perfusion imaging. These initial sequences are completed in about 25 minutes. After 10 minutes for recovery, delayed enhancement perfusion MR is performed. The entire exam is completed in slightly more than an hour.

During interpretation, the wall motion is examined first to map areas of akinesis and dyskinesis. Delayed enhancement can be relied on almost exclusively for CAD diagnosis in regions of akinesis/dyskinesis, Kwong said.

In cases with normal kinetic signs, he looks for inducible ischemia by considering the low-dose dobutamine stress, cine, and delayed enhancement results. Cine function is graded on a four-point scale, with one signifying the most severe conditions. Ischemia is probably present when a deficit identified with perfusion persists beyond peak stress.

"In a nutshell, we look for a deterioration in the grade of segmental function," he said.

Hypokinetic cases are harder to interpret. Their evaluation requires the clinician to more closely define the clinical question for the case at hand. The patient's clinical history and the risk of adverse events during surgery should be considered. Kwong described a patient with a severe reduction of systolic contraction to illustrate the challenge posed by hypokinetic cases. No delayed enhancement was observed, but stress perfusion revealed a deficit indicative of hibernating myocardium. Cardiac catheterization of this patient showed a tight LAD stenosis. Intervention led to complete functional recovery.

ASSESSING VIABILITY

Combining delayed enhancement MR and contractile reserve data has also improved the predictive power of CMR myocardial viability studies performed at Wake Forest University. The pathophysiological basis for delayed enhancement is straightforward, and the procedure itself is easy to interpret, said Dr. W. Gregory Hundley, an associate professor of cardiology. Persistent hyperenhancement observed in DE-MR images corresponds with myocellular death and fibrotic replacement. Because of MRI's inherently high spatial resolution, the clinician can appreciate the extent of subendocardial injury and predict the likelihood of functional recovery following revascularization.

The sensitivity of DE-MRI decreases, however, when the infarction is not transmural, Hundley said. When the injury involves less than 75% of the wall diameter, DE-MRI cannot differentiate between normal myocardium and infarcted, remodeled, hibernating, or stunned tissue.

But a study by Dr. Ernst Wellnhofer in Dr. Eike Nagel's laboratory at the German Heart Institute in Berlin indicates that low-dose dobutamine wall motion imaging can compensate for that weakness (Circulation 2004;109 [10]:2172-2174). In a study of 29 patients with CAD, DE-MRI correctly identified 73% of hibernating segments. A low-dose dobutamine wall motion test predicted 85% of the hibernating segments. The value of DE-MRI for predicting the success of revascularization improved when wall motion results were included, especially for patients with intermediate amounts of scarring, Hundley said.