Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) is a rare autosomal dominant disorder that frequently manifests as ventricular tachycardia or sudden cardiac death in young patients. Histopathologically, it is characterized by progressive fibro-fatty replacement of the myocardium, leading to right ventricular dysfunction 1 2 3 4 5 and ventricular ectopy. Early diagnosis of ARVC is considered imperative for prognosis and management, which usually consists of implantation of an automatic implantable defibrillator. Definitive diagnosis of ARVC remains challenging due to a low disease prevalence and the lack of a single conclusive diagnostic test 6 . Per the 1994 task force criteria (TFC) of the European Society of Cardiology and of the Scientific Council on Cardiomyopathies of the International Society and Federation of Cardiology, a clinical diagnosis of ARVC is established on the basis of structural and functional alteration, electrocardiographic abnormalities, and a positive family history 7 .

Cardiovascular magnetic resonance (CMR) has shown promise in the evaluation of patients with suspected ARVC, given its ability to provide highly accurate anatomical and functional information 8 9 10 11 12 13 14 . In 2010, TFC formally incorporated CMR findings (such as the presence of regional wall motion abnormalities and quantitative functional assessment) into the diagnostic criteria for ARVC 15 . The presence of fatty myocardial infiltration by CMR, which was initially thought to be important in the diagnosis of ARVC due to its unique pathological process, is included in neither the 1994 nor 2010 TFC due to its high prevalence in the normal aging patient population 16 17 18 . Vermes et al. have found that the 2010 TFC resulted in a higher specificity but a lower sensitivity in identifying patients with ARVC, compared to the 1994 TFC 19 . However, the utility of the 2010 imaging criteria has not been validated in large cohorts.

In this study, we sought to evaluate the effect of revised 2010 Task Force Criteria (TFC) on the prevalence of major and minor CMR criteria for ARVC versus the 1994 TFC with and without the presence of fatty infiltrate and abnormal Late Gadolinium Enhancement (LGE) as an additional criterion. In addition, we assessed the utility of CMR in identifying alternative diagnoses for patients referred for ARVC evaluation.

Our institutional IRB granted a waiver for this retrospective research. HIPAA compliance was maintained throughout the study. No outside funding was received, and the authors maintained full control over the data.

Although ARVC is a rather rare disorder, is a frequent reason for referral to CMR. In this light, many radiologists and cardiologists who interpret CMR have become quite familiar with this relatively uncommon disease entity. According to the 2010 Task Force Criteria, meeting even a single major criterion (which can be CMR based) allows for a possible diagnosis of ARVC (the final diagnosis of which requires further clinical input). Therefore, there is a significant burden upon cardiac imagers to provide an accurate assessment and documentation of CMR findings, as this imaging test may have important repercussions for clinical management, including the indication for an implantable cardiac defibrillator placement.

In our study of almost one thousand adult patients referred for CMR during a 15-year period at a tertiary academic referral center with the primary indication of evaluation for ARVC, we report several significant findings. First, the vast majority of patients (943 out of 968, or 97.4%) undergoing cardiac MRI do not meet any criteria for ARVC according to the 2010 Task Force Criteria. This greater selectivity is related to the introduction of more objective and quantitative parameters for diagnosis with the 2010 criteria as compared to the 1994 criteria. The low positivity rate of CMR for ARVC could be considered intuitive given the rarity of the disease, and may possibly reflect a low threshold for ordering a CMR to rule out this entity. This finding becomes important for lower volume centers to recognize, especially if disease prevalence at their centers is similarly low or even lower than at our tertiary referral center.

In addition, we have shown that there is significant re-classification of patients comparing the use of the 1994 versus the 2010 Task Force Criteria. The greatest shift in classification occurred in those patients who met the minor criteria according the 1994 Task Force Criteria. Interestingly, the vast majority of these were re-classified as having no criteria according to the 2010 criteria, yet 8 patients were upgraded to meeting major criteria. These findings suggest that the 2010 criteria are “stricter” by nearly eliminating the category of patients satisfying only minor criteria. This near dichotomization of cardiac MRI categorization (into major and no criteria) suggests a potentially increased importance for CMR in making the diagnosis of ARVC, as meeting the CMR major criterion allows for a possible diagnosis of ARVC (imaging findings alone are insufficient for diagnosis). Conversely, meeting no CMR criteria will eliminate any imaging criteria contribution to the overall clinical diagnosis of ARVC, so that significant non-imaging criteria must therefore be present to establish the diagnosis.

Our findings reinforce and confirm the predictable results of Vermes et al‘ s study by showing an overall reduction in the prevalence of any positive CMR criteria with the revised 2010 Task Force guidelines 15 19 . However, our study included a much larger sample size (968 versus 294 patients), and we included all consecutive patients referred with an indication of evaluating for ARVC whereas Vermes et al. only evaluated patients referred by electrophysiologists. Therefore, our findings may be more applicable to a community practice population (the nature of our referral base made complete electrophysiology records unavailable in a number of externally referred patients). A key difference is that although the number of patients with any positive CMR criteria was substantially reduced with the 2010 criteria, the number of patients with major criteria was relatively similar using both TFC classification schemes. Vermes et al. stated that their relatively large number of patients meeting major criteria according to the 1994 criteria was due to a local tendency to overcall “microaneurysms”, however this terminology was apparently not described in the 1994 criteria (only aneurysms are discussed). Our readers tended to use descriptive terminology (e.g. localized bulge throughout systole and diastole), in addition to the keyword such as “aneurysm”. In addition, aneurysms represented a minority of the regional wall motion abnormalities that were detected on the examined studies.

Aside from these findings, we have taken a step further by evaluating two other items of practical importance: (1) assessing for ARVC “mimics” and other diagnostic findings, (2) evaluating the value of intra-cardiac fat for CMR diagnosis of ARVC. Nearly 10% of the patients in our cohort had either potential ARVC mimics or other clinically important cardiac diagnoses. These findings are in accord with the report by Quarta et al. 21 . Although the importance of dedicated CMR sequences to distinguish between ARVC and cardiac sarcoidosis and myocarditis have previously been described 22 23 24 25 , we found many patients with other cardiomyopathies, as well as RV volume and pressure overload conditions (atrial septal defect, tricuspid regurgitation, etc.) that may explain a patients’ clinical presentation that led to a CMR investigation for ARVC criteria.

Furthermore, much emphasis has historically been placed on fat assessment of the right ventricular myocardium by CMR, with the inclusion of dedicated fat suppression sequences for this purpose at many institutions including our own. Traditionally, the conventional wisdom had been that MRI characterization of fat could serve as a “pathologic equivalent” and thus meet the MRI equivalent of the major 1994 TFC (which required pathologic tissue confirmation of fat). Therefore, we analyzed whether adding the presence of fat in the RV myocardium as a CMR criterion would provide any untility in our patient population. As expected, we found that many patients have some degree of RV myocardial fat, and that there were no patients not fulfilling 1994 criteria whom would have met 2010 criteria (had RV myocardial fat been hypothetically included as a criterion). These findings suggest a limited utility for dedicated, time-intensive sequences for fat evaluation in patients undergoing CMR evaluation for ARVC.

Lastly, abnormal late gadolinium enhancement imaging abnormal in contrast-enhanced MRI can be detected in the right and left ventricular myocardium in patients with suspected arrhythmogenic right ventricular cardiomyopathy suggesting areas of intramyocardial fibrous tissue 26 . In our study, we found that less than half of the patients with positive RV LGE also met either the 1994 or 2010 CMR criteria. All the patients with positive LV LGE but without CMR criteria were ultimately diagnosed with ARVC mimics. These findings suggest that late gadolinium enhancement play a very important role in evaluation of ARVC and is also helpful to identify ARVC Mimics.

LV involvement of arrhythmogenic right ventricular has been identified as more common finding in patients with ARVC. Sen-Chowdhry et al. reported among 200 affected patients, 76.3% patients were found with CMR evidence of LV involvement 27 . In our study, we found 30.9% of the patients with any 1994 or 2010 CMR criteria also had an LV size or function abnormality.

There are, however, several limitations of our retrospective, imaging-based, study. First, we did not have complete clinical information for all patients in order to fully assess clinical Task Force Criteria beyond CMR (e.g. signal-averaged ECG, Holter monitor, etc.), and therefore, we did not have a complete “gold standard” for determining a probable or definitive diagnosis of ARVC. This was secondary to a wide referral base in which many patients only came to our center only for CMR, and therefore did not receive further clinical follow-up at our institution. Second, we pooled cases over many years with varying image quality, during which time technical pulse sequences improved, and clinical readers gained experience. The segmentation of the right ventricle to calculate volumetric parameters was initially not standardized and in some older cases was not performed. Manual RV tracing for quantization is known to be time consuming and to be operator-dependent 28 . However, our intention was to include all cases across this time period (which fell between the 1994 and 2010 TFC, thus eliminating potential bias due to knowledge of the 2010 TFC) to make our results more practically applicable. In addition, in our analysis of ARVC mimics and other cardiac diagnoses, we did not systematically assess whether CMR was the first or only test to detect these diagnoses, as some of these patients may have had additional testing performed at other institutions that was not accessible to us.

Considerations that could be incorporated into future guideline documents for CMR imaging of ARVC include whether other quantitative methods for right ventricular assessment (e.g. tricuspid annular systolic plane excursion [TAPSE], time-volume curves 28 , phase-contrast MRI 29 , or perhaps myocardial strain analysis) may be useful alternatives to traditional quantification of RV volumes and functional parameters 30 . Also, long-term clinical outcome and follow-up data such as defibrillator implantation and appropriate shock therapies, ventricular arrhythmias, and sudden cardiac death are needed for these patients to answer questions regarding the efficacy of current Task Force Criteria in directing further management.

A minority (2.6%) of clinically suspected ARVC patients undergoing CMR met the 2010 TFC at our center. While this suggests a low diagnostic yield of CMR for ARVC (and we also noted no additional benefit of fat assessment), we demonstrated a significantly higher yield for potential ARVC mimics (cardiomyopathies, RV volume and pressure overload conditions) at 4.4% 31 32 . Although our results are provocative, a larger, multicenter registry of CMR for ARVC would be preferable to confirm our single-center results.

In summary, application of the 2010 TFC resulted in a reduction of total patients meeting any diagnostic MRI criteria for ARVC from 22.7% to 2.6% versus application of the 1994 TFC. The hypothetical inclusion of myocardial fat as a criterion would not have changed the TFC diagnosis in any patients. CMR identified alternative cardiac diagnoses in 9.1% of patients, and 4.4% of all diagnoses were potential mimics of ARVC.

TFC: Task force criteria; CMR: Cardiovascular magnetic resonance; ARVC: Arrhythmogenic right ventricular cardiomyopathy; RV: Right ventricle; RVEDV: Right ventricle end diastolic volume; RVEF: Right ventricle ejection fraction; LVEDV: Left ventricle end diastolic volume; LVEF: Left ventricle; BSA: Body surface area; A-P: Anterior-posterior; LVNC: Left ventricular non-compaction; ASD: Atrial septal defect; SVC: Superior vena cava.

The authors declare that they have no competing interests.

TL and AP were involved in designing the study and also collecting and analyzing the data, drafting and submitting the manuscript. UCS was involved in designing the study, reviewing cases and drafting the manuscript. YV, DV PD ML AS MJ HL MS HM were involved in collecting the data and revising the manuscript. AML and SD were involved in revising the manuscript. SA, GH UH was involved in overseeing the study. BBG was involved in designing the study, reviewing cases and drafting the manuscript. All authors have read and approved the final manuscript.