Right ventricular function has emerged as a crucial parameter in the early diagnosis and prognostic assessment of cardiovascular disease [1]. For example, this function has a potential predictive role in pulmonary hypertension (PH) with right ventricular involvement, advanced heart failure (HF) with biventricular involvement [2] and global heart involvement, such as myocardial amyloidosis, as well as in patients surgically treated for tetralogy of Fallot (TOF).

Cardiac magnetic resonance (CMR) is currently the gold standard for assessing right ventricular function [3]. Moreover, right ventricular strain analysis using CMR is one of the several methods for assessing right ventricle (RV) systolic function and detects both myocardial deformability and early abnormalities to provide independent prognostic information. Studies have demonstrated the significance of left ventricular strain in predicting the prognosis of cardiovascular disease [4, 5]. Recently, certain studies have established that RV strain is an independent prognostic factor for several cardiovascular diseases [1, 2]; however, its clinical significance is limited because of the small sample size and the small number of endpoint events. Hence, a systematic review and meta-analysis was conducted to evaluate the prognostic value of right ventricular strain in cardiovascular disease.

This meta-analysis confirmed that RV strain is an important independent predictor of adverse outcomes in many cardiovascular diseases and that RVGLS assessed with CMR has significant prognostic significance in patients with various underlying cardiac abnormalities. In addition, RVGLS values decreased with decreasing RVEF values, which indicates a significant correlation between the two. Also, RVGRS and RVGCS appear to have good predictive value although the number of studies in this analysis is too small to confirm this finding. Finally, RV-LAS was observed to be a powerful predictor of cardiovascular disease development.

CMR has high temporal and spatial resolution, can perform gracilis imaging well and is the gold standard for right ventricular structural evaluation in clinical studies [3]. Two-dimensional echocardiography is the most widely used imaging modality in RV assessment; however, its main limitation is that the image quality depends on operator experience and subject characteristics. EF is the most commonly used and key index that shows systolic function in clinical practice. Nevertheless, EF reflects only global volume changes and cannot reflect alterations in myocardial regional motion or impaired early diastolic function. Myocardial strain technique is a non-invasive quantitative analysis of global and regional myocardial systolic and diastolic functions. Speckle tracking echocardiography (STE) is an accurate and simple method to assess myocardial strain. Park et al. [25] showed that RVGLS obtained via two-dimensional STE correlates well with RVEF and longitudinal strain obtained with CMR. Although STE is primarily a post-processing method, it requires a specific frame rate (50–70 frames/s) and high image quality during image acquisition [26]. Moreover, its wide applicability may be hindered by poor acoustic windows. CMR plays an important role in assessing myocardial strain processes; for example, GCS exhibits better repeatability when calculated using CMR [27].

Several methods are available for obtaining information on myocardial deformation using CMR, which are broadly divided into two categories. The first category requires additional scans and includes CMR tagging, displacement encoding with stimulated echoes (DENSE) and strain-encoded imaging (SENC). The second category is CMR Feature tracking (CMR-FT), which records myocardial strain by acquiring retrospective CMR cine images [27]. CMR tagging, a gold standard for measuring myocardial strain, works on the principle of superimposing magnetic tags (black lines and tags) orthogonally onto the myocardium at the beginning of the cine sequence. The deformation of these lines throughout the cardiac cycle is subsequently analysed [28]. Although it is the most effective CMR technique for assessing myocardial strain [29], its use is limited owing to label fading and low spatial resolution, which reduce its accuracy [28]. CMR tagging requires certain additional sequences, prolonged image acquisition and breath-hold time. Furthermore, it is insensitive to through-plane motion [30]. DENSE for displacement coding was first introduced in 1999 [31], and it is a technique for encoding tissue displacement into the phase of an image. DENSE possesses high temporal and spatial resolution as well as superior strain accuracy and reproducibility [32]. However, additional sequences cause prolonged scanning times, thereby limiting its clinical application. SENC was developed based on the concept of CMR tagging [33], thus enabling the quantification of local deformation of tissues. To calculate myocardial strain, SENC uses magnetized tags parallel to the image plane (rather than being orthogonal as in CMR tagging), which allows higher spatial resolution and, hence, better right ventricular endocardial delineation. Therefore, this technique is effective in quantifying the strain in the through-plane. Because T1 relaxation time leads to fading of the tag, SENC cannot be used to assess myocardial deformation throughout the cardiac cycle. CMR-FT is a simple and convenient new technique for determining myocardial deformation, which can quantitatively analyse the systolic and diastolic functions of the global and regional myocardium in a non-invasive manner. This technology uses post-processing software to delineate the relative motion and displacement of voxels on the endocardial and epicardial boundaries of the right ventricle during the cardiac cycle. Moreover, local or global strain and strain rate of the right ventricle are obtained in radial, circumferential and longitudinal directions. Currently, CMR-FT post-processing analysis can be accomplished with several commonly used commercial software, such as TomTec, Circle, Medis and Medviso. In the case of CVI42, RV myocardial strain is estimated by loading images from four-chamber and short-axis slices into the strain analysis module. In all series, endocardial and epicardial contours are manually delineated per slice at end-diastole. Global myocardial strain parameters are obtained automatically [34], while Fig. 5 shows the specific measurement method. However, because the strain is not derived from the full thickness of the myocardial tissue, CMR-FT may be associated with low accuracy and measurement variability. As additional breath-hold scanning sequences are not required and the reduced scanning time improves the imaging efficiency, CMR-FT is an attractive clinical method. Therefore, all our included studies utilized this method for strain analysis. Additionally, RV-LAS refers to the percentage change in length between the LV apical adventitial border and the midpoint of the line between tricuspid annulus at the end of systole and that at diastole. Long-axis strain (LAS) is a rapidly derived new parameter in CMR cine images, which can be evaluated online without additional software tools and has good clinical application value.

Fig. 5.

Three-dimensional FT global longitudinal strain (GLS), global radial strain (GRS) and global circumferential strain (GCS) of RV obtained using CVI42.

Unlike left ventricular measurements, the assessment of right ventricular function is challenging because right ventricular trabecular muscles are significantly more in number, right ventricular wall is significantly thinner and arrangement and movement of cardiomyocytes are different [35]. CMR can reveal the morphology and function of the heart and is the gold standard for assessing the complex geometry of RV. Right ventricular strain obtained using CMR is more sensitive than EF and has a high clinical value in assessing right ventricular myocardial damage in patients with subclinical myocardial damage. The method is also useful in risk stratification and in determining the therapeutic effect. Compared with conventional parameters of RV systolic function, such as RVEF, RV strain can detect subtle changes in RV function without significant wall motion abnormalities or reduced global RV systolic function. Such changes have significantly higher predictive values than conventional parameters. For example, Mahmod et al. [22] found that both RV and LV strain were impaired in patients with HCM despite normal LVEFd. A small decrease was observed in RVEF, but it was within the acceptable normal range. Moreover, Henning et al. [36] found that the percentage of normal myocardium in LV and RV detected using rapid SENC (normal LV and RV myocardial segmental strain $\le$–17%) better identifies asymptomatic patients with subclinical LV dysfunction. This technique may be useful in the early identification of healthy subjects who may be at risk for HF as well as in the monitoring of LV and RV deformation during pharmacological interventions in future studies. RV strain can help in detecting the subclinical stage of PH and in assessing the disease and the associated prognosis [17]. In congenital heart diseases, such as TOF, CMR can assess the efficacy and necessity of surgery. Moon et al. [37] used CMR-FT and found that right ventricular strain values were significantly lower in the TOF group than in the normal group and that right ventricular longitudinal strain was closely related to adverse outcomes in patients with TOF. Early diagnosis and analysis of wall motion abnormalities are possible in patients with ARVC [38]. CMR-FT strain analysis helps to objectively quantify global/regional right ventricular dysfunction and dyssynchrony in patients with ARVC and provides effective diagnostic information.

The cut-off values of RVGLS were obtained in five studies [2, 13, 17, 19, 20]. Although the strain decreased compared with the normal group, the five values were significantly different. This variation could be attributed to the fact that the sources of patients differed in the five included studies [2, 13, 17, 19, 20]. The study by Liu et al. [2] only included patients with HF in stages C and D of cardiac function; hence, RVGLS was relatively worse.

RV fibres are primarily arranged along the longitudinal axis under the epicardium. In healthy individuals, longitudinal shortening largely leads to RV shortening [39]. Twelve studies [2, 10, 11, 12, 13, 14, 16, 18, 19, 20, 21, 22] on RVGLS were included in our meta-analysis, and despite the presence of moderate heterogeneity, the pooled effect values demonstrated the prognostic value of RVGLS. Of these, three articles concluded that right ventricular strain did not have a statistically significant prognostic value in cardiovascular disease. The following could be the reasons for this conclusion: Yang et al. [16] consecutively included patients with CMR findings and did not classify the disease specifically; hence the heterogeneity was relatively high. Bourfiss et al. [19] found that RV and LV strain were no longer significant predictors of persistent ventricular arrhythmias (VA) after adjusting for risk factors, such as RVEF and LVEF. The reason could be that most patients who developed VA already had advanced structural disease, but decreased RV longitudinal and LV circumferential strain were observed in other patients with ARVC and persistent VA during follow-up, thus indicating that strain can reflect changes in ventricular function. Although Mahmod et al. [22] noted that the decrease in RVGLS was not associated with cardiovascular composite events, it had a significant predictive value for the occurrence of non-sustained ventricular tachycardia. Several patients in the included studies had right heart involvement and global involvement, but a certain bias existed. Hence, the additional independent prognostic value of right ventricular strain must be carefully studied while considering conventional cardiac function. In four other studies [2, 13, 16, 18] that included RVGCS, it was shown to have good prognostic ability. However, more research on RVGRS and RVGCS is required because of the small number of available studies. In three studies [12, 16, 20], RVGRS exhibited good predictive power, possibly because strict motion of the basal RV and the thin atrioventricular free wall hindered adequate endocardial tracking on long-axis views [16]. The contribution of circumferential strain is important in patients with PH. Unfortunately, data on RVGCS were not included in the three studies [12, 16, 20] on PH, and thus, the results could not be validated. Yang et al. [23] showed that the predictive ability of RV-LAS was better than that of RVEF and TAPSE for the poor prognosis of patients with HCM. Arenja et al. [24] showed that RV-LAS had the highest diagnostic accuracy in the cohort of patients with NIDCM. Moreover, RV-LAS was an independent marker of MACEs in multivariate analysis. The above studies assert the important role of RV-LAS in prognosis prediction. RV-LAS is defined as the change in length from the LV apex to the tricuspid annulus rather than the RV apex; therefore, it may incorporate both LV and RV longitudinal functions. With the increase in studies on RV-LAS, in the future, the predictive value of RV-LAS can be proved using a meta-analysis.

This systematic review and meta-analysis has certain limitations. First, because studies on right ventricular strain are currently in the developmental stage, the number of studies included in this meta-analysis is relatively small (n = 14). Therefore, publication bias could not be explored. Moreover, only five studies [2, 13, 17, 19, 20] included cut-off values and they dealt with different diseases; hence, a pooled effect size and cut-off value could not be calculated for a specific disease. We are hopeful that with the increase in the number of in-depth studies in this field, we would be able to update and include more studies in the future to provide robust evidence. Furthermore, we believe that we would have a better understanding of the prognostic value of RVGRS and RVGCS. Second, as with meta-analyses of several observational studies, moderate heterogeneity exists among the articles because of differences in the study design, inclusion criteria, follow-up time and vendor software. Although we performed a correlative meta-regression analysis for some measures, we were unable to perform a meta-regression analysis due to the diverse variety of software vendors for the articles we included. it is not well understood whether differences in software suppliers are the source of heterogeneity. Our study demonstrated a significant negative correlation between RVGLS and RVEF. These conclusions have been confirmed in several studies [40, 41]. We understand that strain values may be affected by factors such as magnetic fields and post-processing software. With the increase in the studies on right ventricular strain, in the future, we can analyse the above factors in subgroups. Finally, to reduce the heterogeneity of articles, we excluded studies involving patients with a history of cardiac surgery. Strain is a valuable prognostic indicator for patients with congenital heart disease [42]. With continued technological advancements, a meta-analysis of these patients can be performed in the future.

This systematic review and meta-analysis establishes the prognostic value of right ventricular strain determined using CMR in cardiovascular disease. RVGLS assessed with CMR has significant prognostic implications in patients with different underlying cardiac abnormalities. RVGRS, RVGCS and RV-LAS appear to show good prognostic value, and although the number of studies on these strains is small, they can pave way for future studies and thus validate our conclusions. In the future, the prognostic value of right ventricular strain in different diseases can be investigated to obtain cut-off values that correspond to different diseases, thereby guiding treatment decisions and prognostic stratification.

KB, LX and ZHS are guarantors of the integrity of the entire study; KB and ZZ were responsible for the screening of the literature, extraction of the data; HW and YFG was responsible for the analysis of the data; All results were checked by HKZ and TL; KB prepared the first draft of the manuscript, which was critically revised by LX and ZHS.

Not applicable.

Not applicable.

This study was supported by grants from the National Natural Science Foundation of China (U1908211, 82271986), the Capital’s Funds for Health Improvement and Research Foundation of China (2020-1-1052).

The authors declare no conflict of interest. Zhonghua Sun is serving as one of the Editorial Board members and Guest Editors of this journal. Tong Liu is serving as one of the Guest Editors of this journal. We declare that Zhonghua Sun and Tong Liu had no involvement in the peer review of this article and have no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to Dimitrios Tsartsalis and Constantina Aggeli.