Multisequence stress CMR shows promise

Stress-rest SPECT imaging may be a generally accurate test for coronary artery disease, but it is painful, time-consuming, and logistically complex in practice. New multisequence cardiac MR tests are proving in clinical trials to be as sensitive for the detection of coronary artery disease as stress-rest SPECT and far more specific.

Although stress thallium imaging can be performed in a 30-minute morning session, the patient must return in the afternoon for the resting portion of the procedure. If the results of that test are positive, the patient must be rescanned immediately and again the following day to measure myocardial viability.

A CMR stress-rest perfusion imaging exam can take less than an hour to capture all the diagnostic information needed to measure the presence and extent of myocardial ischemia and the viability of surviving myocardial tissue following infarction.

Studies presented at the SCMR meeting strongly suggest that although a 15-minute visual reading of the images may not be as diagnostically powerful as a 30-minute quantitative evaluation of results, it is still good enough to outperform stress-rest SPECT. And in at least one study, CMR surpassed stress echocardiography, according to Dr. Dipan J. Shah, director of the Nashville Cardiovascular MRI Institute.

Shah and his former colleagues at Duke University's Cardiovascular MRI Center devised a protocol that combines the strengths of several MR tests of myocardial performance. The multisequence approach was designed to address such issues as the low signal to noise of first-past contrast-enhanced MR myocardial perfusion, which is the primary way CMR has been used to test for coronary artery disease.

"No standards for a first-pass protocol exist. No consensus regarding pulse sequences exists, and no consensus has been formed on whether a rest perfusion study should be performed or whether a stress or rest perfusion sequence should be performed first during the procedure," he said.

The Duke protocol begins with cine MR imaging to capture wall motion data. The patient is infused with adenosine, a pharmacologic vasodilator, for about three minutes. Gadolinium contrast is injected immediately before first-pass stress perfusion imaging.

The patient rests for 15 minutes on the scanner bed while the adenosine clears. Another bolus of gadolinium is injected before first-pass rest perfusion imaging. After a 5-minute pause, delayed-enhancement MR images are acquired as a test of myocardial viability. The average procedure is completed in 45 minutes.

Imaging involves using a saturation recovery sequence with acquisition performed every heartbeat. The resolution of 3.5 x 2.5 x 8-mm slice thickness is nearly 10 times better than nuclear stress imaging, Shah said.

Two primary conclusions about coronary artery disease can be gleaned from the protocol, according to Shah: Hyperenhancements that appear on delayed enhancement imaging establish the diagnosis of CAD, and hypoenhancement and the absence of perfusion abnormalities during stress imaging rule out its presence. For the sake of simplicity, wall thickening is examined only in equivocal cases.

About 700 patients at Duke and the Heart Group, a cardiology group practice in Nashville, were studied with the multimodality protocol by the end of January. A prospective study based on the experience of 100 of those patients demonstrated how the strengths of one phase of the test compensate for the weaknesses of other phases, said Dr. Igor Klem, principal investigator and a senior cardiovascular MRI fellow at Duke.

The combined first-pass stress-rest and delayed enhancement protocol generated overall sensitivity and specificity rates for CAD diagnosis of 86% and 88%, respectively (see table). The overall performance was 46 percentage points more sensitive than delayed enhancement MR alone and 20 percentage points more specific than first-pass perfusion MR alone, Klem said.

EXPERIENCED READERS REQUIRED

Dr. Nidal Al-Saadi, a cardiologist at Charite Clinic in Berlin, reported that a multimodality protocol similar to the Duke University approach can be highly accurate even when experienced readers adopt a fast visual approach to perform their interpretation.

An overall accuracy rate of 88% with negative and positive predictive values of 96% and 83%, respectively, were tabulated in the study of 151 patients with known or suspected coronary artery disease. Two experienced readers performed a visual analysis based on a 16-segment model of the myocardium using a five-point scale to determine the presence or absence of an infarct. Regional perfusion was classified according to its transmural extent and defect reversibility. MR results were compared with x-ray angiography. A 50% or greater stenosis was used as the threshold for the presence of coronary artery disease.

Al-Saadi's results, however, contradicted the experience of Dr. Eike Nagel, director of cardiovascular MRI at the German Heart Institute in Berlin. In a study of 86 patients reported at the 2003 SCMR conference, Nagel found that the sensitivity and specificity of first-pass MR tests fell 14 percentage points and 32 percentage points, respectively, when physicians switched to visual analysis (Circulation 2003;108:432-438).

Dr. Scott Bingham, a radiologist at the Central Utah Multi-Specialty Clinic in Provo, found in his study of 138 patients with suspected CAD that the combination of a multimodality CMR test and a careful visual inspection can be accurate. Readings were completed before x-ray angiography was performed in 32 patients and several months later on a blinded basis. The two sets of interpretations generated sensitivities of 92% and 88%, respectively. The specificities for the pre- and postangiography visual readings were 81% and 78%, respectively.

Dr. Sven Plein, a lecturer in radiology at Leeds General Infirmary in the U.K., found a multimodality perfusion CMR protocol more sensitive than TIMI, a widely accepted clinical and biochemical marker protocol for determining whether patients with suspected coronary artery disease should be referred for cardiac catheterization.

In Plein's study, the two techniques were performed in 68 patients with non-ST-segment-elevation acute coronary syndrome.

Their status was based on the presence of chest pain and the results of abnormal electrophysiology tests. The CMR protocol included first-pass rest and adenosine-stress perfusion, wall motion, and late-enhancement MR imaging. A careful visual analysis of the images took an average of 30 minutes, the same amount of time needed for Nagel's readers to complete a quantitative analysis in his 2003 study. The clinicians sought to identify cases involving greater than 70% coronary artery stenosis requiring revascularization.

Although comprehensive CMR and TIMI correctly identified the severity of coronary artery disease in 83% of the cases, CMR was far more sensitive to the presence of disease (p< 0.05). The sensitivity rates for CMR and TIMI were 96.4% and 50%, respectively.

The combined results presented at the 2004 SCMR meeting led Shah to conclude that CMR stress testing has great promise.

"When a systemic approach is used, it is able to achieve sensitivities and specificities comparable to SPECT and stress echo," he said. "It can be performed in a rapid fashion and is amenable to a rapid clinical assessment."