Identifying the optimal regional predictor of right ventricular global function: a high‐resolution three‐dimensional cardiac magnetic resonance study†

Summary Right ventricular (RV) function has prognostic value in acute, chronic and peri‐operative disease, although the complex RV contractile pattern makes rapid assessment difficult. Several two‐dimensional (2D) regional measures estimate RV function, however the optimal measure is not known. High‐resolution three‐dimensional (3D) cardiac magnetic resonance cine imaging was acquired in 300 healthy volunteers and a computational model of RV motion created. Points where regional function was significantly associated with global function were identified and a 2D, optimised single‐point marker (SPM‐O) of global function developed. This marker was prospectively compared with tricuspid annular plane systolic excursion (TAPSE), septum‐freewall displacement (SFD) and their fractional change (TAPSE‐F, SFD‐F) in a test cohort of 300 patients in the prediction of RV ejection fraction. RV ejection fraction was significantly associated with systolic function in a contiguous 7.3 cm2 patch of the basal RV freewall combining transverse (38%), longitudinal (35%) and circumferential (27%) contraction and coinciding with the four‐chamber view. In the test cohort, all single‐point surrogates correlated with RV ejection fraction (p < 0.010), but correlation (R) was higher for SPM‐O (R = 0.44, p < 0.001) than TAPSE (R = 0.24, p < 0.001) and SFD (R = 0.22, p < 0.001), and non‐significantly higher than TAPSE‐F (R = 0.40, p < 0.001) and SFD‐F (R = 0.43, p < 0.001). SPM‐O explained more of the observed variance in RV ejection fraction (19%) and predicted it more accurately than any other 2D marker (median error 2.8 ml vs 3.6 ml, p < 0.001). We conclude that systolic motion of the basal RV freewall predicts global function more accurately than other 2D estimators. However, no markers summarise 3D contractile patterns, limiting their predictive accuracy.


Introduction
Right ventricular (RV) function has prognostic value in acute [1,2] and chronic [3][4][5][6] cardiorespiratory disease and in the peri-operative period [7][8][9], although rapid and accurate assessment is challenging due to the right ventricle's complex geometry and motion [10,11]. Several regional surrogate measures of global RV function have been investigated, most commonly targeting the contributions of longitudinal and transverse motion, because these are thought to predominate [12]. In normal subjects, longitudinal shortening may account for the majority of RV function [13] and can be measured by tricuspid annular plane systolic excursion (TAPSE).
However, TAPSE may be unreliable when assessed by less experienced operators [14] and in mild and moderate RV dysfunction [15]. Prospective evaluations suggest TAPSE's prognostic value is limited [16,17].
Identifying the optimal two-dimensional (2D) index of systolic function is difficult as some regions are easy to identify on imaging, some are influential on ventricular function and some are affected by dysfunction. These three areas do not necessarily coincide [18] and may change because of disease processes [19]. Determining the individual contribution made to global function by the excursion of each point of the right ventricle requires a computational model of cardiac motion. Cardiac magnetic resonance (CMR) is considered the reference standard for RV volumetry and, recently, high-resolution three-dimensional cardiac magnetic resonance (3D-CMR) has been used for accurate segmentation of whole-heart anatomy [20][21][22]. The variation in regional motion within a population can then be explored using atlas-based segmentation techniques, where each subject's images are co-registered [23,24]. We decided to apply these techniques in order to determine which area of the right ventricle best reflects global function in the general population, and tested its predictive performance against conventional indices such as TAPSE and septum-freewall displacement (SFD).

Methods
This study was conducted at a single centre and was approved by the Hospital's research ethics committee. Test-retest reproducibility was assessed using bias (mean and standard deviation) and intra-class correlation coefficient (ICC) with a two-way random model for absolute agreement [29]. Regional excursion was assessed for association with global systolic function using pointwise bootstrapped linear regression adjusted for age, sex, race and body surface area (BSA). Data were centred and scaled before analysis. Significance was calculated using permutation testing (10,000 permutations) with correction for multiple testing by false-discovery rate.
Adjacent points on the RV wall are likely to have highly In model validation, each of the five 2D measures was tested by correlation, linear regression and accuracy of RVEF prediction. Correlation with RVEF was assessed using the Pearson product-moment coefficient. Difference in correlation coefficients was assessed by a method described by Steiger [30]. The association of each 2D measure with RVEF was assessed by linear regression with these variables, data were centred, scaled and coefficients bootstrapped (10,000 replications) with RVEF as the dependent variable and age, sex, race, BSA and the 2D measure of interest as independent variables (see also Supporting Information Table S1). Difference in regression models was assessed by treating the model type as an additional covariate. Right VEF prediction was tested using a leave-one-out analysis where RVEF for a single subject was predicted from the remaining 299 subjects and the process was repeated for all 300 validation subjects. Median absolute prediction errors were compared by Kruskal-Wallis rank sum test (v 2 statistic). A p value of < 0.05 was considered statistically significant.

Results
Six hundred and forty adult volunteers successfully completed the imaging protocol. All datasets were used for analysis and there were no failures of the segmentation algorithm. Baseline characteristics of participants included in the study are shown in Table 1 Bootstrapped linear regression identified a distinct, contiguous 7.3 cm 2 patch in the basal freewall where excursion was significantly associated with RVEF (Fig. 2).
The patch was located between 4.1 cm and 6.9 cm from the apex and represented 9% of the total RV surface area. We   Table 2.
Freewall function increased from apex to base and this was mostly due to an increase in longitudinal function (see also Supporting Information Figure S1). Patch function Information Table S3 and Figure S2).

Discussion
Several parameters have been proposed to estimate global RV function, but this is the first report of a systematic computational approach to search for the optimal measurement. Using 3D-CMR, we found that systolic displacement of the basal RV freewall is the strongest predictor of global function, outperforming TAPSE and other related indices in predicting RVEF. However, the 3D contractile pattern of the right ventricle limits the accuracy of estimators derived from 2D imaging planes. Clinicians  Right ventricular dysfunction is increasingly recognised as a significant predictor of morbidity and mortality in patients in the acute [1,2], chronic [3][4][5][6] and peri-operative [7][8][9] settings. The complexity of RV function has prompted efforts to summarise function by markers of longitudinal [14,15,18] and transverse [31] function, though these markers have weaknesses [17,32]. Measures which combine both, such as fractional area change, appear to perform best but rely on accurate endocardial delineation and require suitable analysis software. These indices of RV function are derived from standard 2D imaging planes despite evidence that RV myocardial architecture is oblique to these planes, incorporates 3D contraction patterns and shows regional variation [33,34]. Furthermore, global function is affected by age, sex, race, BSA and interactions between these factors [35][36][37].
Our data show that the complex 3D contractile pattern of the right ventricle can be accurately assessed using CMR coupled with segmentation techniques, allowing consistent comparisons to be made between anatomical points across a population. This method allows modelling of the 3D contractile pattern of the right ventricle, which originates from its layered architecture with deep, longitudinal myofibrils [11] contributing to long-axis excursion [38,39] and superficial myofibrils contributing to transverse and circumferential contraction [11]. Applying these approaches to a cohort of healthy subjects allows the contributions of these components to be assessed in the normal ventricle [18,19]. We suggest that although these contributions may vary in disease processes, and particularly in response to elevated RV pressure [19,[40][41][42], they are still useful for clinicians involved in the assessment of the right ventricle.   Our study has several limitations. Work was confined to healthy volunteers with a presumed low incidence of regional wall motion abnormalities which are known to impair the performance of regional markers of global function [43]. Regional markers based on excursion take no account of abnormal diastolic function or ventricular coordination, which may exist without systolic dysfunction [44]. We evaluated this marker's performance using CMR in

Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article.