Right ventricular outflow tract dimensions in arrhythmogenic right ventricular cardiomyopathy/dysplasia—a multicentre study comparing echocardiography and cardiovascular magnetic resonance

Abstract

Aims

Right ventricular outflow tract (RVOT) dilation is one of the echocardiographic criteria in the 2010 revised Task Force Criteria (TFC) of arrhythmogenic right ventricular cardiomyopathy/dysplasia (ARVC/D). However, studies comparing cardiac magnetic resonance (CMR) and transthoracic echocardiography (TTE) suggest a lower diagnostic accuracy of TTE due to its operator dependence and limited reproducibility. The goal of this study was to compare the 2010 TFC measures of RVOT dilation with three alternative measures for improving the echocardiographic assessment of RVOT in patients with ARVC/D.

Methods and results

In this multicentre study, CMR and TTE were performed in 38 patients with a definite, borderline, or possible ARVC/D diagnosis and in 10 healthy controls. Besides the echocardiographic RVOT measurements listed by the 2010 TFC, we assessed three additional end-diastolic RVOT diameters. These included the RVOT diameter defined by the parasternal long axis M-mode of the aortic sinus portion (RVOT3), that defined by the parasternal long axis M-mode of the left ventricle (RVOT4), and that obtained by the parasternal short axis view of the distal RVOT proximal to the pulmonary valve (RVOT5). RVOT4 provided the best correlation between CMR and TTE (r = 0.92, [95% confidence interval (CI): 0.84–0.96; P < 0.0001]) and enhanced diagnostic accuracy for diagnosing ARVC/D (area under the curve 0.92 [95% CI, 0.78–0.98]).

Conclusion

Among all RVOT diameters examined, that defined by the parasternal long axis M-mode of the left ventricle (RVOT4) provides the best agreement between CMR and TTE and exhibits the best diagnostic accuracy for ARVC/D. This novel RVOT4 measurement carries the potential for improving the echocardiographic diagnosis of ARVC/D.

Introduction

Arrhythmogenic right ventricular cardiomyopathy/dysplasia (ARVC/D) is a disease of the cardiac muscle with autosomal dominant inheritance causing fibro-fatty infiltration of the right and at later stages the left ventricle.^1–4 It is associated with an increased cardiac morbidity and mortality mainly due to ventricular arrhythmias and heart failure.^5,6 The disease seems to become manifest earlier and more pronounced in physically highly active young adults and accounts for up to 22% of exercise related sudden cardiac deaths in those cohorts.^7–11 Hence, an early and reliable diagnosis of ARVC/D is essential to prevent fatal outcome.¹²

A recent study by Borgquist et al.¹³ evaluated the diagnostic performance of transthoracic echocardiography (TTE) and cardiovascular magnetic resonance (CMR) in ARVC/D based on the 2010 revised Task Force Criteria (2010 TFC)¹⁴ and found that TTE may be unreliable for detecting subtle changes of the right ventricular (RV) dimensions and function. The authors concluded that TTE misses a significant number of patients with imaging-positive ARVC/D by CMR and postulate that the better diagnostic performance of CMR should be appropriately reflected in the guidelines. Although CMR is the most accurate imaging modality in the advanced diagnosis of ARVC/D, the vast majority of patients undergo TTE as initial examination due to the limited availability of CMR. Thus, increasing the diagnostic accuracy of TTE seems essential for improving timely diagnosis in a considerable number of patients with suspected ARVC/D.

The present study aimed at comparing the 2010 TFC diameters defining right ventricular outflow tract (RVOT) dilation and three additional echocardiographic RVOT diameters regarding their correlation with the corresponding CMR diameters as well as CMR-based parameters of RV dilation. We hypothesized that highly standardized echocardiographic projections provide better correlation with CMR as well as higher reproducibility, and may thereby increase the diagnostic accuracy for detecting patients with ARVC/D by echocardiography.

Methods

Patient population

For this cross-sectional study, 44 patients were recruited from the Zurich ARVC/D Program and three international centres (Leiden University Medical Center, The Netherlands; University Hospital RWTH Aachen, Germany; Federal Almazov North-West Medical Research Centre, Saint Petersburg, Russia). The enrolled patients had a definite (n = 24), borderline (n = 9) or possible (n = 11) ARVC/D diagnosis according to the 2010 TFC. The ARVC/D patients and 10 additional healthy controls, recruited at the University Hospital Zurich, Switzerland received an echocardiographic assessment as well as a CMR examination. A total of 90 TTE examinations of control subjects without cardiovascular disease and normal echocardiographic exams were evaluated at the University Hospital Zurich, Switzerland to define the major and minor criteria for the best performing RVOT diameter in accordance with the 2010 TFC. This study was approved by the local and institutional authorities. For retrospective analysis of clinically acquired data the Dutch Central Committee on Human-related Research (CCMO) allows the use of anonymous data without prior approval of an institutional review board provided that the data are acquired for routine patient care. All data used for this study were acquired for clinical purposes and handled anonymously. The study protocol conforms to the 1964 Declaration of Helsinki and its later amendments.

Transthoracic echocardiography

In all patients, standard 2D TTE was performed by experienced investigators trained in the analysis of the anatomy of the RV as previously described.^3,15 The RVOT1 (RVOT-PLAX-Prox), RVOT2 (RVOT-PSAX-Prox) and RVOT5 (RVOT-PSAX-Distal) dimensions were measured in accordance with current guidelines^16,17 (Figure 1A and B). RVOT3 was determined in the axis defined by the parasternal long axis M-mode of the aortic sinus portion, and RVOT4 in the axis defined by the parasternal long axis M-mode of the left ventricle (Figure 1A). All measurements were acquired in the respective 2D images in end-diastole from inner edge-to-inner edge. All patients were in sinus rhythm during the examination. Each parameter was averaged over at least two heart cycles. Dimensions and function of the RV were assessed according to established guidelines for 2D TTE.¹⁷

Cardiovascular magnetic resonance

CMR was performed on clinical 1.5T or 3.0T MRI scanners of different vendors. For the assessment of RVOT dimensions, cine images with a minimum frame rate of 25/heart cycle were acquired. RVOT1, RVOT3 and RVOT4 were measured in the three chamber view, as an anatomical correlate of the echocardiographic parasternal long axis view (PLAX, Figure 1C). In accordance with echocardiographic guidelines, RVOT1 (RVOT-PLAX-Prox) was measured from the interventricular septum at the level of the annulus fibrosus to the anterior RVOT wall in alignment with the angle bisector between the interventricular septum and the ascending aorta. RVOT3 was measured in the axis defined by the echocardiographic equivalent of the parasternal long axis M-mode of the aortic sinus portion. RVOT4 was measured in the axis defined by the echocardiographic equivalent of the parasternal long axis M-mode of the left ventricle. In the short axis view of the aortic valve, RVOT2 (RVOT-PSAX-Prox) was measured perpendicular to the direction of the RVOT from the centre of the aortic valve to the anterior RVOT wall. The short axis view of the aortic valve was oriented parallel to the aortic annulus plane. We acquired five subsequent slices in this orientation and evaluated RVOT2 in the slice that was located in the centre of the aortic valve at end-diastole (Figure 1D). RVOT5 (RVOT-PSAX-Distal) was measured perpendicular to the direction of the RVOT in the RVOT-view¹⁸ just proximal to the pulmonary valve (Figure 1E). All measurements were acquired from inner edge-to-inner edge at end-diastole. In addition, standard measurements of RV and LV volumes and function including RV end-diastolic volume (RV EDV) and RV ejection fraction (RVEF) as well as left ventricular ejection fraction (LVEF) were obtained. All CMR measurements were performed with GTVolume (GyroTools LLC, Zurich, Switzerland).

Reproducibility

To assess the reproducibility of RVOT1-5, intra- and interobserver reproducibility were determined for TTE and CMR. For intraobserver reproducibility, 20 patients and 5 control datasets were randomly selected and reanalysed by the same observer. For interobserver reproducibility, measurements of those 25 datasets were repeated by an independent observer.

Statistical analysis

Continuous variables are presented as mean ± standard deviation (SD) for normally distributed data or median and interquartile ranges (IQR) for not normally distributed data. Normal distribution of data was assessed using the Shapiro–Wilk test. Categorical variables are reported as frequency (percentage). Comparisons between the patient and control groups were performed by the 2-sided unpaired Student’s t-test for normally distributed, the Mann–Whitney U-test for not normally distributed, and the Fisher exact test for categorical variables.

Pearson correlation was performed to analyse correlations between TTE and CMR parameters. Additionally, we calculated receiver operating characteristic curves and derived sensitivity and specificity for the determined cut-off values. The optimal criteria (OC) for RVOT measurements were determined as OC = max (sensitivity(x) + specificity(x) – 1), where x ranges over all possible criterion values. Intraclass correlation (ICC) was used to quantify intra- and interobserver variability. Statistical analysis was conducted with MedCalc Software v 15.6.1 (MedCalc Software bvba, Ostend, Belgium). A P-value of <0.05 was considered statistically significant.

Results

Patient population

Recruitment for this study was performed between December 2008 and September 2015. Data acquisition and evaluation was successful in 38 out of the 44 enrolled patients (86%). Three patients were excluded because no CMR scan could be performed and three patients had to be excluded due to insufficient image quality of at least one imaging modality. The remaining 38 ARVC/D patients (39 ± 18 years, 10 females) were age- and sex-matched with the 10 healthy control subjects (33 ± 6 years, 3 females). Both examinations were performed within 37 [IQR: 8–89] days. According to the 2010 TFC, n = 20 included patients had a definite, while n = 7 had a borderline and n = 11 had a possible ARVC/D diagnosis. On average, patients had 2 major criteria (range 0–5) and 1 minor criterion (range 0–3). Clinical baseline characteristics are summarized in Table 1, imaging baseline characteristics are displayed in Table 2. See Supplemental data for the demographic data of the 90 control subjects that received only a TTE examination.

Agreement between TTE and CMR based RVOT diameters

Figure 2 illustrates the results of the correlation (first row) and agreement (second row) between the five TTE- and CMR-based RVOT diameters, as well as the correlation between indexed TTE-based RVOT diameter and indexed CMR-based end-diastolic RV volume (third row) in all ARVC/D patients.

Regarding the two 2010 TFC recommended echocardiographic diameters, the RVOT1 dimension on TTE exhibited a good correlation with the corresponding CMR diameter with r = 0.87, [95% CI: 0.75–0.93; P < 0.0001] and a good absolute agreement of Δ_CMR-TTE = −0.2 mm [± 1.96 standard deviation (SD): −7.5/7.1 mm]. Indexed RVOT1 also exhibited a fair correlation with CMR-based RV-EDVI (end-diastolic volume index) with r = 0.43 [95% CI: 0.13–0.66; P = 0.007]. In contrast, the 2010 TFC recommended RVOT2 dimension showed a weaker correlation with r = 0.72, [95% CI: 0.42–0.87; P = 0.0002] and worse agreement of Δ_CMR-TTE = −2.6 mm [± 1.96 SD: −13.4/8.1 mm], but a similar correlation with RV-EDVI with r = 0.42 [95% CI: 0.10–0.66; P = 0.012]. Among all the parameters examined, RVOT5 (RVOT-PSAX-Distal) exhibited the poorest correlation with r = 0.50, [95% CI: 0.14–0.75; P = 0.009] and the worst agreement with Δ_CMR-TTE = 3.4 mm [± 1.96 SD: −4.8/11.5 mm], and no significant correlation with CMR based RV-EDVI with r = 0.15 [95% CI: −0.24 to 0.50; P = 0.45].

The echocardiographic RVOT diameter defined by the parasternal long axis M-mode of the left ventricle (RVOT 4) produced the best results in all categories with a correlation r = 0.92, [95% CI: 0.84–0.96; P < 0.0001], an agreement Δ_CMR-TTE = −0.2 mm [± 1.96 SD: −5.4/5.0 mm], and the highest correlation with CMR-based RV-EDVI with r = 0.47 [95% CI: 0.18–0.69; P = 0.003]. RVOT3 did not provide any advantage over the 2010 TFC recommended diameters: correlation r = 0.75, [95% CI: 0.57–0.87; P < 0.0001], agreement Δ_CMR-TTE = −0.8 mm [± 1.96 SD: −10.5/8.9 mm], and correlation with RV-EDVI r = 0.45 [95% CI: 0.15–0.67; P = 0.005]. In general, the correlation between echocardiographic RVOT diameters and RV-EDVI determined by CMR was modest (Figure 2).

Reproducibility of RVOT diameters

The echocardiographic RVOT 1–4 diameters exhibited high reproducibility with ICC values > 0.91 for inter- and intraobserver correlation. The RVOT 5 dimension, however, displayed lower ICC values equalling 0.68 (95% CI: 0.37–0.85) for interobserver and 0.78 (95% CI: 0.52–0.90) for intraobserver correlation. CMR showed excellent inter- and intraobserver reproducibility with ICC values > 0.95 for all RVOT measurements. See Table 3 for a detailed list of all reproducibility data.

Diagnostic accuracy of new RVOT vs. 2010 TFC RVOT diameters

Sensitivity, specificity and diagnostic accuracy for detecting definite or borderline ARVC/D compared to healthy controls are given in Table 4. The best performing echocardiographic parameter was RVOT4 with a diagnostic accuracy of 92% [area under the curve (AUC) 0.92 (95% CI, 0.78–0.98)] for the absolute RVOT diameter. When corrected for body surface area (BSA), RVOT4/BSA (AUC 0.89 [95% CI, 0.74–0.97]) displayed a non-significant trend to higher diagnostic accuracy compared with RVOT1/BSA (AUC 0.84 [95% CI, 0.68–0.94]; P = 0.13) and RVOT2/BSA (AUC 0.81, [95% CI, 0.64–0.92]; P = 0.21). RVOT4 provided an 8% increase in sensitivity (78% vs. 70%) with the same specificity as RVOT1. In CMR, RVOT4 significantly improved diagnostic accuracy when compared with the CMR based measurement of RVOT 1 (RVOT4_CMR: AUC 0.87 [95% CI, 0.71–0.96] vs. RVOT1_CMR: AUC 0.73 [95% CI, 0.55–0.86], P < 0.001). RVOT3 (AUC 0.82 [95% CI, 0.66–0.92]), exhibited a non-significant trend to lower diagnostic accuracy when compared with the 2010 TFC established RVOT1 and RVOT2 diameters. Similarly, RVOT5 (RVOT PSAX-Distal) provided no advantage compared to the 2010 TFC diameters. The TTE-based RVOT5 dimension had a significantly lower diagnostic accuracy compared to RVOT2 (RVOT5_TTE: AUC 0.66 [95% CI, 0.46–0.83] vs. RVOT2_TTE: AUC 0.86 [95% CI, 0.70–0.95], P = 0.03). When corrected for body size, RVOT5/BSA had significantly lower diagnostic accuracy compared to RVOT1/BSA (P < 0.01) and RVOT2/BSA (P < 0.01). The 2010 TFC require a specificity of 95% for major criteria and equal sensitivity and specificity for minor criteria. As RVOT4 exhibited the best diagnostic performance, we evaluated the conditions consistent with a major or minor criterion in accordance to the 2010 TFC. Thereby, we found that a major criterion is fulfilled with RVOT4 ≥ 32 mm and RVOT4/BSA ≥ 17.5 mm/m², whereas a minor criterion is fulfilled with RVOT4 ≥ 28 mm and RVOT4/BSA ≥ 15.5 mm/m² (see Supplemental data for detailed information).

Discussion

This multicentre study is the first to report on the comparative assessment of various RVOT diameters between TTE and CMR in patients with ARVC/D. We observed that measuring the echocardiographic RVOT diameter as defined by the parasternal long axis M-mode of the left ventricle (RVOT4) yielded the highest diagnostic accuracy for the diagnosis of ARVC/D and was superior to the diameters recommended in the 2010 TFC as well as current echocardiography guidelines. Furthermore, echocardiographic RVOT4 displayed the best correlation with the corresponding RVOT diameter as well as RV-EDVI by CMR. Intra-/interobserver correlation was generally very good and comparable for RVOT1-4. Thus, these findings indicate an improved diagnostic potential for the echocardiographic assessment of the RVOT diameter as defined by the parasternal long axis M-mode for assessing left ventricular function and dimensions according to the Teichholz formula.

CMR provides an unparalleled accuracy in characterizing and monitoring non-ischaemic cardiomyopathies and has prognostic value in such patients.^19,20 In particular, in the workup of ARVC/D, CMR has emerged as an excellent imaging technique, as it delivers very accurate information on cardiac morphology, function and tissue characterization in a single investigation.²¹

However, due to its availability, rapid application and relatively low cost, TTE is the most widely used imaging modality in the initial assessment of patients suspected of having ARVC/D. Since CMR permits an accurate and reproducible evaluation of RV morphology, this imaging modality was applied for improving the echocardiographic assessment of RVOT in patients with ARVC/D. The echocardiographic RVOT diameters recommended in the 2010 TFC¹⁴ as well as three novel RVOT diameters were examined regarding their agreement between the two imaging modalities, reproducibility, and diagnostic accuracy. The correlation between RVOT diameters and RV-EDVI were generally modest, indicating that RVOT dimensions do not constitute a good surrogate for RV dilation. As expected, ICC were higher for CMR measurements as compared to echocardiographic measurements, which is in line with previous data.^22,23

The RVOT diameter defined by the parasternal long axis M-mode of the left ventricle (RVOT4) provided the highest reproducibility between CMR and TTE, and highest diagnostic accuracy to diagnose ARVC/D. In accordance with the recent 2010 TFC, we determined cut-off values for major and minor criteria for the new RVOT4 compared to 90 control subjects without cardiac disease and normal echocardiographic exams. Those cut-off values for RVOT4 are well in line with suggested cut-off values for RVOT1 (PLAX) listed in the 2010 TFC. RVOT4 also exhibited the best correlation between indexed (for BSA) TTE-based RVOT diameter and indexed (for BSA) CMR-based RV-EDVI, implying a potential superiority as a surrogate parameter for RV dilation. Different reasons may account for the high diagnostic value of RVOT4. An anatomical reason is related to the largely perpendicular orientation of RVOT4 on the lower RVOT in most patients; this orientation reduces the variability of measurements as compared to the other diameters, in particular RVOT3, which exhibits an oblique and in some patients even tangential orientation in the RVOT. A methodological reason consists in the clear definition of the orientation of RVOT4 as given by left ventricular anatomical landmarks; the latter promote intraobserver and interobserver reproducibility.

Compared to the PLAX measurements of RVOT (RVOT 1, 3 and 4), the diameters taken in PSAX orientation (RVOT 2 and 5) had a lower agreement between CMR and TTE. While TTE tended to overestimate the proximal RVOT in PSAX (RVOT2: Δ_CMR-TTE = −2.6 mm, P = 0.05), the distal RVOT was significantly underestimated by TTE (RVOT5: Δ_CMR-TTE = 3.4 mm, P < 0.01). These discrepancies may be explained by the anatomic difficulties in obtaining RVOT images from the short axis. While PLAX views can be oriented to several anatomic landmarks, it is more difficult to control the orientation of the imaging plane in the short axis. The aortic valve is not a perfect anatomic landmark indeed, because the leaflets do not insert in the aortic wall in a defined plane and because the valve undergoes a translational motion during the cardiac cycle.²⁴ These difficulties may compromise the accuracy of TTE PSAX measurements of the RVOT by rendering the short axis plane more likely to pass obliquely through the RVOT.²⁵ The 2010 TFC recommended TTE-based RVOT2, however, exhibited a fair correlation with the corresponding CMR measurements, despite the high absolute difference. Additionally, RVOT2 had the highest sensitivity (84%) of all diameters in our study to identify definitive or borderline ARVC/D, thereby corroborating its nomination as a 2010 TFC criterion. In contrast, the distal RVOT in PSAX (RVOT5), which is recommended to assess RVOT dimensions according to the guidelines for RV chamber quantification,¹⁷ exhibited only poor correlation with the corresponding CMR measurements, no significant correlation with RV dilation, and the weakest intra- and interobserver reproducibility. In most patients, it is challenging to generate an echocardiographic view centring on the pulmonary valve while simultaneously providing sufficient endocardial definition of the RVOT anterior wall. Compromises in achieving one of these goals account for the poor diagnostic performance of RVOT5.

Limitations

A limitation of the present study is the small sample size of ARVC/D patients with different stages of disease expression, although we made an effort to increase patient numbers in a multicentre approach. In addition, the control cohort did not include endurance athletes, who are known to have a physiological dilation of the RVOT. Therefore, the inclusion of endurance athletes in the control cohort may have resulted in lower diagnostic specificity values of RVOT diameters.

Conclusions

In summary, the comparative assessment of RVOT dimensions by CMR and TTE has shown that measuring the RVOT as defined by the parasternal long axis M-mode of the left ventricle (RVOT4) provides the highest agreement between CMR and TTE as well as the best correlation with CMR-based RV-EDVI. In addition, this diameter yielded the highest diagnostic accuracy for detecting ARVC/D. If our results can be validated in larger cohorts, the novel RVOT4 measure has the potential for improving echocardiographic diagnosis of ARVC/D.

Supplementary data

Supplementary data are available at European Heart Journal—Cardiovascular Imaging online.

Conflict of interest: The department of Cardiology of the Leiden University Medical Center receives research grants from Medtronic, Biotronik, Boston Scientific and Edwards Lifesciences. Victoria Delgado received speaker fees from Abbott Vascular.

Funding

This study was supported by the Georg und Bertha Schwyzer-Winiker Foundation and the Baugarten Foundation, both in Zurich, Switzerland.

References

Author notes