MR perfusion imaging. What will be its impact for detection of coronary disease in the future?

Abstract

Revealing stress-induced myocardial ischaemia is a well-established method for evaluating the presence and pathophysiological severity of coronary artery disease. In clinical practice, exercise or pharmacological stress nuclear imaging are the most widely used techniques for this purpose. According to two meta-analyses, exercise1 and dipyridamole and adenosine stress2 single-photon emission computed tomography (SPECT) have respectively 87, 89, and 90% sensitivity and 64, 65, and 75% specificity for detection of angiographically significant coronary artery disease. Detection of myocardial ischaemia by nuclear imaging also has prognostic value for predicting cardiac events3,4 such as death or myocardial infarction. Nevertheless, nuclear imaging has several significant limitations: indeed the test is quite lengthy, since it usually requires a stress and a rest study to be performed in separate sessions. In addition, conventional cardiac nuclear imaging has poor spatial resolution and lacks the ability to perform quantitative measurement of perfusion. It also suffers from attenuation artefacts. More importantly, it exposes patients to significant doses of ionizing radiation, with the potential risk of radiation-induced cancer. Finally, nuclear imaging is far from being a perfect test for detection of coronary artery disease. As indicated above, the test lacks specificity for detection of coronary disease. Therefore, alternatives for conventional nuclear perfusion imaging would be clearly desirable. Although positron emission tomography perfusion imaging can overcome many of the limitations of conventional nuclear imaging,5 its complexity and cost, and the requirement to have either a cyclotron or a rubidium generator on-site, have prevented widespread use of the technique. Cardiac magnetic resonance (cMR) may be another more … *Corresponding author. Tel: +32 2 7642803, Fax: +32 2 7642811, Email: Bernhard.gerber{at}clin.ucl.ac.be