DE-CMR demonstrates its predictive powers

May 11, 2005

Prognostication has become a key feature of cardiac MRI. That role was established in 2003 when Dr. Raymond Kim, Dr. Robert Judd, and their colleagues at Duke and Northwestern universities demonstrated that delayed enhancement MR could reliably differentiate between patients who will benefit from revascularization following myocardial infarction and those who will not.

Prognostication has become a key feature of cardiac MRI. That role was established in 2003 when Dr. Raymond Kim, Dr. Robert Judd, and their colleagues at Duke and Northwestern universities demonstrated that delayed enhancement MR could reliably differentiate between patients who will benefit from revascularization following myocardial infarction and those who will not.

Studies suggesting that CMR's prognostic powers can apply to other questions that arise during the days following a myocardial infarction were featured at the SCMR meeting in January. Dr. Rishi Kaushal, a CMR researcher at the University of California, Los Angeles, presented results demonstrating that infarction size measured with cine and DE-CMR can predict the risk of mortality the injury poses. Kaushal's study addressed a paradox in the criteria for selecting candidates for an implantable cardioverter defibrillator (ICD). Epidemiologic data indicate that patients thought to be at low risk of sudden cardiac death have a higher mortality rate than patients in the high-risk group.

Patients with an ejection fraction between 31% and 40% as well as unsustained ventricular and inducible tachycardia during invasive electrophysiological testing qualify for an ICD. However, clinical trials have shown that even patients with noninducible tachycardia during invasive electrophysiological testing remain at high risk of sudden death. In fact, among patients with low ejection fraction, electrophysiological testing is no longer required to stratify their risk.

Kaushal's investigation was based on a study of 100 patients with CAD. They were initially imaged with cine and DE-CMR as part of a routine cardiac examination. The mean left ventricular ejection fraction among the group was 34%. Kaushal also measured the location and size of the previous MI. Regional wall thickening and contrast enhancement were measured using a 12-segment anatomic model. MI was determined with planimetry.

Outcome data collected an average of 25 months later found that 17 subjects had died. Univariate analysis indicated that infarction of greater than 15% of left ventricular mass and injection fraction were the only significant predictors of death (see graph). Using Cox multivariate analysis, Kaushal determined that an infarction of greater than 15% was the best predictor of mortality. The results led him to conclude that the extent of MI can predict whether a patient is susceptible to a death that could potentially be avoided with ICD implantation.


Mark Doyle, Ph.D., a CMR physicist at Allegheny General Hospital in Pittsburgh, and Dr. Gerald Pohost, director of cardiovascular imaging at the University of Southern California, presented study results suggesting that late-enhancement imaging may help cardiologists more accurately assess the severity of coronary artery disease among women.

Women with intermediate or high Framingham risk scores (Adult Treatment Panel III) are classified with severe CAD, but follow-up quantitative coronary angiography often shows that the disease is less serious.

Doyle and Pohost presented results from a multicenter study that was part of the Women's Ischemia Syndrome Evaluation sponsored by the National Heart, Lung, and Blood Institute. The findings indicate that combining ATP III criteria and a global myocardial perfusion index (GMPi) calculated from resting perfusion MRI will help clinicians better assess the clinical risks their female CAD patients face.

Researchers at nine sites examined 133 women with suspected MI and average ATP III scores of 9.8. About 20% of the patients had a greater than 50% occlusion of a coronary artery. They underwent quantitative coronary angiography and first-pass myocardial perfusion MR to measure the GMPi. This calculation reflects the status of myocardial function using ventricular time-to-signal peak and left ventricular blood pool data.

Ten major adverse coronary events, such as death and MI, were documented as the subjects were monitored for the next 38 months. Patients with intermediate or high ATP III scores accounted for 38% of patients with an adverse event, experiencing 80% of the events. Using the GMPi, the researchers found that 90% of the patients who fell into the two lowest quartiles of myocardial function as measured with MRI experienced an adverse event.

Using Cox modeling, the researchers determined that ATP III and GMPi together could separate high-risk from low-risk patients. Doyle and Pohost concluded that the combined ATP III and GMPi risk was more accurate than the Framingham index alone for identifying women who would potentially benefit from coronary angiography and excluding women who would probably not benefit from the quantitative test.


The ability of first-pass stress-rest perfusion MRI to confirm MI is well established, but until now researchers could not say with certainty that the exam would help clinicians evaluate candidates for revascularization. Dr. Oliver Bruder's long-term study of patients with suspected CAD suggests that high-dose dobutamine perfusion MR can identify patients with a high risk of a major cardiac event.

Bruder's study at Elisabeth Hospital in Essen, Germany, was based on 382 patients who underwent high-dose dobutamine perfusion MRI to examine clinically suspected CAD. From that total, 156 patients with inducible ischemia were referred to coronary catheter angiography. Sixty-four who underwent revascularization were excluded from further study. The medical histories of the remaining 320 patients were tracked for an average of 16 months.

The results demonstrated the value of wall motion measured during first-pass perfusion imaging to ascertain the probability of a future adverse cardiac event. Nearly one of four (24.5%) patients who exhibited abnormal wall motion during stress-rest perfusion imaging experienced a major event in the follow-up period. The event rate for patients with normal wall motion was 8.5%.


Dr. Matthias Regenfus, a researcher at the University of Erlangen in Germany, conducted a tracking study of 102 patients who underwent percutaneous reperfusion after MI. His findings indicate that DE-CMR performed during the week after intervention can identify patients most likely to have future cardiac problems.

The MR protocol, performed on a 1.5T system optimized for cardiovascular imaging, consisted of cine imaging with TrueFISP and DE-CMR acquired with a segment inversion recovery sequence, performed 10 minutes after injection of gadolinium

contrast. A 17-segment model was used to score regional wall thickening and contrast enhancement on a five-point scale. Left ventricle function was determined using planimetry.

Regenfus tracked the patients for an average of 2.5 years to record incidences of cardiac-related mortality and morbidity. Eleven deaths and infarctions were documented, and 26 patients were revascularized or hospitalized with unstable angina or congestive heart failure.

Significantly compromised ejection fraction was the most powerful predictor of these events (p = 0.006). Among patients who died or experienced reinfarction, the dysfunctional area circumscribed with CMR (0.65 vs. 0.48, p = 0.08) and the dysfunctional but viable area identified with delayed enhancement (0.16 vs. 0.27, p = 0.008) were significantly higher than in patients without such events.

Predicting events

A positive delayed enhancement exam provided clinicians with the most potent data for predicting future cardiac events among nine variables tested by Dr. Raymond Kwong, codirector of CMR at Brigham and Women's Hospital. Kwong and colleagues interviewed the families of 279 patients about 14 months after the patients underwent a late-enhancement MR viability study to investigate suspected CAD. During the follow-up period, 71 patients experienced an adverse cardiac event: 30 died, four suffered an acute MI, 20 reported unstable angina, and 70 developed decompensated congestive heart failure requiring hospitalization.

Compared with nine variables-including a prior MI, a 40% reduction of left ventricular ejection fraction, ECG indications of CAD, and wall motion abnormalities-a positive myocardial delayed enhancement test produced the strongest association with the probability of a future adverse event, according to Dr. Carmen W.S. Chan, who presented the results at the SCMR meeting.