In this study of consecutive patients with biopsy proven extra-cardiac sarcoidosis who were prospectively screened for cardiac involvement, the principal finding was that DE-CMR identified myocardial abnormalities in significantly more patients than a standard clinical evaluation based on the consensus guidelines from the Japanese Ministry of Health (26% vs. 12%). An additional finding was that DE-CMR positive patients had a higher rate of adverse events including cardiac death during follow-up compared with DE-CMR negative patients, however the cohort was small and there were few events.
The prevalence of cardiac sarcoidosis of 26% found by DE-CMR is consistent with autopsy series in the United States, which have found myocardial lesions in 20–27% of patients with sarcoidosis.2,3
For many reasons, the lower detection rate from the standard assessment is not surprising. As a first step in the diagnosis of cardiac sarcoidosis, JMH guidelines require an abnormal electrocardiogram, yet even in patients with findings of extensive myocardial infiltration on autopsy, the premortem electrocardiogram may be normal in 25%.3
More importantly, necropsy studies have demonstrated that many sarcoid patients suffering sudden death have only limited cardiac involvement, with lesions that are small and patchy.3,4,24
Premortem, it is unlikely that these lesions would have been associated with gross changes in ventricular cavity size or function, or detected by traditional cardiac imaging techniques. In the present study, the median extent of damage detected by DE-CMR was only 6.1% of myocardial mass. Additionally, on a regional basis, the majority of affected segments (70%) had non-transmural involvement (<50% hyperenhanced). The latter is important since non-transmural scarring—even if overall scar size is considerable—is frequently associated with normal ventricular wall motion9
and normal myocardial perfusion by nuclear scintigraphy.10
Thus, the ability of DE-CMR to detect small and/or nontransmural myocardial damage is likely the reason why a more than two-fold higher rate of cardiac disease was identified compared with the standard assessment.
Improved diagnostic sensitivity may come at the cost of lower specificity. On this point we note that none of the abnormalities detected by cardiac imaging, either as part of the JMH guidelines or from DE-CMR, are specific for cardiac sarcoidosis. For instance, any cardiomyopathic process may meet the JMH criteria of abnormal wall motion. Similarly, hyperenhancement on DE-CMR representing scar tissue and/or necrosis,12,19,20
may result from a variety of etiologies including ischemic heart disease. To account for this, as part of the study design, we excluded patients with prior myocardial infarction or known coronary disease. Furthermore, all patients with hyperenhancement underwent coronary angiography and none had evidence of obstructive coronary disease. As a consequence, in our study population of patients with biopsy proven sarcoidosis, it is highly likely that hyperenhancement represents cardiac sarcoidosis rather than another nonischemic process (e.g. cardiac amyloidosis, etc.), since it is improbable that patients would be afflicted concurrently with two rare disorders.
Our pathology data are also consistent with the interpretation that hyperenhancement in the study population represents cardiac sarcoidosis. Of the 4 patients with pathology confirmed cardiac sarcoid (2 at autopsy, 2 by endomyocardial biopsy), all 4 were identified by DE-CMR whereas the standard assessment identified only 2. Moreover, in the 2 with autopsy, gross examination of the heart demonstrated a nearly 1-to-1 concordance between the location, shape, and extent of myocardial lesions and in vivo
hyperenhancement (). Interestingly, of the 13 patients undergoing endomyocardial biopsy, only 2 were positive for cardiac sarcoidosis. This diagnostic yield of 15% is similar to that found by Uemura et al.7
who reported a positive biopsy rate of only 19% in a study cohort in whom the clinical diagnosis of cardiac sarcoidosis had already been made. Importantly, Uemura et al.7
concluded that patients with clinical evidence of cardiac involvement be treated for cardiac sarcoidosis despite negative myocardial biopsies, since inhomogeneous myocardial involvement and sampling error will frequently lead to false negative biopsies. Our results suggest that 6 of 11 patients had false negative biopsies (), and the DE-CMR images indicate a mechanism: none of the 6 had widespread hyperenhancement of the right ventricular side of the interventricular septum, which was the source of tissue for endomyocardial biopsy.
These findings suggest that DE-CMR provides additional value to standard assessment for the diagnosis of cardiac sarcoidosis, and bolster similar conclusions from prior investigations.13,25
However, from a clinical care standpoint, arguably the prognostic implications are most important. In this regard, the present study suggests that damage detected by DE-CMR may be associated with future adverse clinical events in patients with sarcoidosis. Even if small, regions of myocardial damage identified by DE-CMR may provide substrate for ventricular arrhythmias and conduction disturbances,26
and DE-CMR has shown prognostic utility independent of common clinical and functional predictors in other cardiac disorders.27–30
Nonetheless, in the current study events were few and caution should be used in interpreting these results until confirmed in a larger cohort.
The current report appears to be the first to prospectively demonstrate a link between a noninvasive index of cardiac involvement and clinical outcome in patients with sarcoidosis. Yazaki et al,31
retrospectively identified 75 Japanese patients with the clinical diagnosis of cardiac sarcoidosis. Multivariable analysis demonstrated that left-ventricular enlargement by echocardiography, sustained ventricular tachycardia, and NYHA functional class were independent predictors of all-cause mortality. Notably, in this cohort many patients had heart failure symptoms and NYHA functional class was the most powerful prognostic predictor. Smedema et al.32
reviewed data in 101 consecutive patients with pulmonary sarcoidosis assessed at two university medical centers in the Netherlands. Although a battery of tests, including CMR, was performed in many patients, and during a mean follow-up of 1.7 years, there were 4 deaths and 9 received a pacemaker and/or implantable cardioverter/defibrillator, the study was primarily descriptive and there was no data relating results of noninvasive testing with clinical outcome. Additionally, it appears that only a minority of patients underwent inversion-recovery DE-CMR, based on another report the same year involving many of the same patients.13
In most patients, CMR images were acquired using an older spin-echo sequence, and this is important since there are considerable differences in image quality between these techniques.33
Recently, Mehta et al.34
reported on a prospective evaluation of 62 ambulatory patients with biopsy-proven extra-cardiac sarcoidosis who were screened for cardiac involvement. Using a structured algorithm, 26 patients underwent CMR and 25 had PET. Based on abnormalities found on either CMR or PET, 24 patients (39%) were given the diagnosis of cardiac sarcoidosis. Over a mean follow-up of 1.8 years, no patient died or had ventricular arrhythmias, and the authors concluded that sarcoidal lesions seen on CMR or PET do not predict arrhythmias in patients with preserved cardiac function. At first glance, this report appears to contrast with the current investigation. However, we note that hyperenhancement on DE-CMR was found only in 8 patients, and the chance of a type II error seems high. Conversely, in patients with negative DE-CMR scans, the results of this report appear to corroborate our finding of a low event rate (< 2% per year) and a benign course.
Our study has a number of limitations. If patients systematically had ambulatory electrocardiography in addition to the resting 12-lead and had undergone every non-CMR cardiac imaging procedure, the number of patients with cardiac involvement by JMH criteria would likely have been higher. However, we attempted to compare DE-CMR with a clinically plausible approach (an average of 1.6 non-CMR imaging studies were performed per patient), and we note that an earlier investigation utilizing both resting and ambulatory ECG and multiple non-CMR imaging modalities demonstrated similar findings to the current study—namely that DE-CMR was more sensitive for cardiac involvement than standard clinical approaches.13
T2-weighted sequences aimed at detecting acute edema/inflammation were not performed, and there may be some uncertainty regarding the precise pathophysiological interpretation of myocardial hyperenhancement. However, in our study cohort, nearly all patients had chronic sarcoidosis (median of 7 years since diagnosis) and none had acute cardiac symptoms. Moreover, repeat DE-CMR in a subset of patients demonstrated that the location, shape, and extent of hyperenhancement were unchanged (). Along with the pathology findings, these suggest that hyperenhancement represents scar and/or chronic granulomatous tissue rather than acute necrosis with inflammation in our population. There appears to be racial differences in both the prevalence of cardiac sarcoidosis and its morbidity and mortality.5
Most of our patients were African-American, and caution should be used in extrapolating our findings to other patient populations. While we believe our study represents the largest experience with contemporary DE-CMR in patients with sarcoidosis, to date, the cohort was relatively small, from a single-center, and clinical events were few. Our findings will need to be replicated in a larger, preferably multi-center investigation.
The presence of severe left-ventricular dysfunction (LVEF ≤ 30–40%) is often a key indication for clinical management decisions including implantable cardioverter-defibrillator therapy.35
Yet in our study only 2 of the 8 patients with adverse events during follow-up had an LVEF less than 40%. These data suggest that in patients with sarcoidosis—as in other populations36
—there may be a paradox in that although low LVEF identifies a group with relatively high risk, the majority of adverse events occur in patients with preserved LVEF. Since DE-CMR detected 6 of 8 patients with adverse events in the current study, future investigations could focus on whether DE-CMR can improve the management of patients with sarcoidosis, including the early diagnosis of cardiac involvement and the potential for prophylactic defibrillator use.