The present study addressed the impact of DCMR on clinical management in a large unselected patient population with chest pain syndromes. The main findings are 1) DCMR is applicable in a clinical routine setting with a high success rate and few stressor related complications during a reasonably short examination time of less than 30 minutes, 2) DCMR proved useful as an arbiter for clinical decision making with regard to invasive versus medical treatment in patients with suspected and known CAD, 3) the positive predictive value of DCMR to detect coronary luminal narrowing >50% is high, 4) a positive DCMR is a powerful predictor of future cardiac events, and 5) a negative DCMR test result infers a low risk for subsequent cardiac events (about 1% in the two years after stress testing).
One of the most important clinical questions that non-invasive stress testing has to address is whether patients should be advised to undergo invasive angiography or to continue with medical treatment. As a result of major advances in imaging technology, several diagnostic strategies have become available over the past decades. Although exercise electrocardiography is advocated as a first-line procedure , sensitivity may be as low as 45 percent and many patients cannot exercise sufficiently due to poor functional status . In order to determine a patient's most appropriate management multiple tests may be conducted and frequently yield conflicting results. Thus, in a large number of individuals a single imaging test in conjunction with pharmacological stress as the initial strategy may be the most effective approach in patient care. Whereas echocardiography and radionuclide imaging have been evaluated extensively, data on CMR based management strategies are scarce. DCMR has matured into a technically robust method with similarly high values for sensitivity and specificity of ≈ 85% for the detection of myocardial ischemic reactions in the presence of obstructive coronary lesions  and has been shown to provide relevant prognostic information [[10, 11, 18]]. Previous studies focused on the prognostic value of DCMR in low/intermediate versus high risk patient groups as defined by conventional cardiovascular risk factors and reported a relative merit of stress magnetic resonance testing . The design of the present study, however, was unique in that it established the utility of DCMR testing as the sole clinical decision maker in a routine clinical setting: our study attributed DCMR testing an active role in clinical decision making with treatment directed either to a medical or invasive strategy. Consequently, while corroborating the usefulness of DCMR testing, our data closely reflects clinical reality in a tertiary care center setting and as such will be applicable to a similar clinical scenario.
The overall safety profile and frequency of adverse events of DCMR observed in our study are in agreement with previous reports using CMR and other well established methodologies using high dose dobutamine-atropine stress protocols [9, 20]. Results of DCMR were communicated in a timely and definitive manner to the referring physician so that they formed the basis for subsequent clinical decision making.
In our study most patients with a positive DCMR underwent invasive angiography with the intention to perform revascularization. Compared to prior results regarding the diagnostic accuracy of DCMR, this study confirms the high predictive value for the detection of angiographically relevant obstructive coronary stenoses in a population with known or suspected CAD [5, 6]. DCMR-positive patients who did not undergo invasive angiography within 90 days frequently sustained hard cardiac events. Interestingly, the number of patients who sustained early events was relatively high compared to results of the recently published COURAGE and BARI-2D trials which compared invasive vs. conservative management of stable CAD [21, 22]. However, there are certain methodological differences between these trials and our study. First, all patients in the above mentioned trials had to have angiographically proven significant coronary stenosis as an inclusion criterion. In the present study, however, patients were classified with regard to inducible ischemia on the myocardial level. Thus, generalization from these trials to the present patient population is limited. Second, subgroup analyses from COURAGE and BARI-2D showed that outcome is worse with complex CAD and high extent of inducible ischaemia, and that early revascularisation in addition to optimal medical therapy was better than optimal pharmacological therapy alone [3, 23]. Since early revascularization is likely to improve outcome in these high-risk patients, the pivotal role of cardiac imaging as an arbiter in clinical decision making is further corroborated. In our study patients with early events had a significantly greater extent of ischemia suggesting that an early referral to invasive angiography may be advisable in this patient group. Prior studies using SPECT  and echocardiography  demonstrated that the extent and severity of stress inducible ischemia is associated with a worse outcome, however, similar data using DCMR testing is still limited.
The vast majority of DCMR-negative patients in our study did not undergo invasive angiography during follow-up time. Similar to SPECT imaging and stress echocardiography, a normal DCMR has generally been associated with a hard annual cardiac event rate of ≈ 1% [[10, 11, 26, 27]]. Data from randomized trials proved that the low rate of cardiac events in patients with negative stress examinations cannot be improved by revascularization as indicated in current guidelines [1, 28]. These patients can be safely treated initially with medical therapy and should only be investigated further if their symptoms cannot be controlled. The impetus for DCMR-driven management has been data from a previously published study dealing with the prognostic value of DCMR and demonstrating a warranty period of two years in case of a negative DCMR test result .
In our study 10 percent of the patients with a negative DCMR were subsequently referred for invasive angiography largely owing to recurrent anginal symptoms. In 52 percent revascularization was performed indicating that these patients may have been misclassified as DCMR-neg. The overall "false-negative" rate in our study, however, was low. Previous studies using stress echocardiography have shown that chest pain in the absence of identifiable wall motion abnormalities represents an independent predictor of future cardiac events and should be considered in the interpretation of a normal examination . In addition, it most likely also reflects clinical practice since physicians facing a patient with uncontrolled symptoms are more likely to refer for invasive angiography based on their clinical judgment regardless of the results of prior non-invasive testing. Furthermore, the sensitivity of stress inducible wall motion abnormalities as a marker of ischemia is known to be slightly lower compared to myocardial perfusion imaging techniques. Thus, the addition of perfusion imaging during high dose dobutamine may be helpful in detecting patients with ischemia . Nevertheless, all DCMR-neg patients who underwent invasive angiography without revascularization did not sustain any hard cardiac events supporting the high negative predictive value of the test.
Delayed enhancement (DE) images were not acquired for the purpose of this study. The presence and extent of DE has already been demonstrated to carry independent prognostic value [30, 31]. Recently, the combination of CMR vasodilator stress myocardial perfusion and DE was shown to provide complementary prognostic implication for cardiac events . Thus, with regard to prognostication the addition of DE to DCMR may be beneficial in certain patient populations. Visual analysis of invasive angiography by experienced interventional cardiologists was used to determine the degree of coronary luminal narrowing. Quantitative coronary angiography (QCA) is often used as a reference standard in clinical trials, but its usage during routine coronary angiography is rather limited though it may be performed to assist planning of a revascularization procedure. The aim of the present study, however, was to define the role of DCMR testing within a widely seen clinical scenario.