Atrioventricular plane displacement versus mitral and tricuspid annular plane systolic excursion: A comparison between cardiac magnetic resonance and M‐mode echocardiography

Both echocardiography and CMR imaging are used to quantify longitudinal function. Inter‐method variability for mitral (MAPSE) and tricuspid (TAPSE) annular plane systolic excursion, and variability between directly measured MAPSE and TAPSE and as based on atrioventricular plane displacement (AVPD) analysis by CMR, are, however, not known. This study, therefore, assessed inter‐method variability and variability between annular plane systolic excursion and AVPD‐based values in a healthy adult population.


| INTRODUC TI ON
Ventricular systolic function is dependent on both longitudinal pumping, that is the base-to-apical movement of the valve plane, and radial pumping, that is the squeezing motion of the ventricle. Longitudinal pumping is the main contributor to stroke volume on both the left and right cardiac side (Carlsson et al., 2007a(Carlsson et al., , 2007bSteding-Ehrenborg et al., 2015). Measures of longitudinal function are of clinical importance as a decreased longitudinal function has prognostic implications (Romano et al., 2019). Mitral annular plane systolic excursion (MAPSE) and tricuspid annular plane systolic excursion (TAPSE) as measures of systolic longitudinal function can be assessed by echocardiography (Dutta & Aronow, 2017;Feigenbaum et al., 1967;Kaul et al., 1984;Medvedofsky et al., 2015). Further, cardiovascular magnetic resonance (CMR) can assess both MAPSE and TAPSE and the atrioventricular plane displacement (AVPD) (Seemann et al., 2017).
Measurement points for AVPD are placed on the basal top of the compact myocardium in the respective long-axis. This may differ from where the mitral and tricuspid annuli and valve hinge points are visualized, and it may be hypothesized that MAPSE and TAPSE according to clinical routine and as based on tracked data from AVPD analysis may result in different values. Also, inter-method variability for MAPSE and TAPSE, and variability between MAPSE/TAPSE by echocardiography and based on AVPD by CMR are not known.
The aims of this study were therefore to compare MAPSE and TAPSE by echocardiography with values by semi-automatic CMR analysis from both direct measurements and based on tracked data from AVPD and to determine reproducibility of both methods in a healthy adult population.

| Study design
The Ethics Committee of Hospital Clinic and Hospital Sant Joan de Deu approved the study protocol, and the study was performed in accordance with the Helsinki Declaration. Healthy subjects between age 25 and 40 years were invited to participate and underwent a complete cardiovascular risk assessment. Exclusion criteria were cardiovascular or renal disease, diabetes mellitus, autoimmune disease or contraindications to CMR. All subjects provided written informed consent before participation and underwent comprehensive echocardiography and CMR studies.

| Echocardiography
All subjects were evaluated according to clinical routine in the left lateral decubitus position performed by an experienced cardiologist from the Cardiology Unit of Hospital Clínic, Barcelona, with more than ten years of clinical practice. A Vivid E9 ultrasound machine (GE Healthcare, Horten, Norway) with a 3.5 MHz (M5S) transducer was used for all examinations with image acquisition synchronized to ECG. Cine loops of the apical four-chamber view and M-mode images acquired in free breathing were stored for offline analyses.
An experienced observer (A.S-M.) quantified MAPSE and TAPSE off-line using the software Echopac® (GE Healthcare, Milwaukee, WI, USA). In short, 2D-mode cines with frame rates of 60-70 frames per second were used. An M-mode line was applied across the lateral mitral and tricuspid annulus in the four-chamber view for assessment of MAPSE and TAPSE, respectively ( Figure 1). Two to three echocardiography cine data sets were acquired in each patient, and the clip number was not known for the second observation.  (Heiberg et al., 2010;Seemann et al., 2017).  F I G U R E 2 Mitral and tricuspid annular plane systolic excursion reference markers at end-diastole for semi-automatic measurement using cardiac magnetic resonance imaging. The reference markers at end-diastole for MAPSE (1) and TAPSE (2) are denoted

| Intra-and inter-observer analysis
A subset of 20 subjects was randomly selected for intra-and interobserver variability of MAPSE and TAPSE for echocardiography and CMR. One blinded observer (A.S-M.) analysed echocardiography and CMR measurements one week apart for intra-observer variability for both methods. Two other blinded observers provided analyses of echocardiography (B.V-A.) and CMR (K.S-E.) for inter-observer variability.

| Statistical analyses
The D'Agostino test was performed to test for normal distribution.
Variables were expressed as mean ± SD or median [IQR], and statistical differences were tested using Student's t-test. Categorical variables were expressed as n (%).
For comparisons between echocardiography and CMR, Deming regression was used to estimate the fitted up-slope for overcoming the assumptions of a classic linear regression considering measurement errors present in both methods (Cornbleet & Gochman, 1979). For 95% confidence interval (CI) determination, the Jackknife method was used. The coefficient of variation (CoV), as a measure of relative variability, was estimated as a ratio of the standard deviation to the mean. Bias and 95% limits of agreement for MAPSE and TAPSE were calculated and presented in Bland-Altman plots (Bland & Altman, 1986).
Intra-and inter-observer reproducibility was calculated as the concordance correlation coefficient (CCC) (Lin, 1989), with the CCC agreement scale as described by McBride (McBride, 2005). Bias and limits of agreement were calculated and presented using Bland-Altman plots.
To determine the impact of time between echocardiography and CMR on reproducibility, a quantile regression test was performed, considering the time in days between examinations by echocardiography and CMR as the independent variable and the absolute inter-method difference as the dependent variable.
Time between examinations was divided in four intervals, based on the largest difference between studies (<26 days, 27-52 days, 53-79 days, >80 days), and a trend analysis with the Jonckheere-Terpstra test performed, considering the absolute time between examinations.
All statistical analyses were performed in Stata 14.2 (Statacorp, College Station, Texas, USA). A two-sided P-value < 0.05 was considered to show significant differences.

| RE SULTS
One hundred and eleven healthy subjects were included in the current study. Median time between echocardiography and CMR was 41 [20-103] days. Quantile regression for MAPSE and TAPSE demonstrated no correlation between measurements and time between examinations (r = 0.0; p = 0.9, and r = 0.002; p = 0.6, respectively).
Trend analysis confirmed these results, with no trend found after considering time between examinations (p = 0.4 and p = 0.6 for MAPSE and TAPSE, respectively). Analyses were, therefore, based on all 111 subjects included in the current study. Participants' characteristics are presented in Table 1. There was no difference between heart rate at echocardiography (68 ± 12 bpm) and at CMR (66 ± 11 bpm; p = 0.1).

| Comparison of echocardiography and CMR
Mitral annular plane systolic excursion by manual echocardiography measurements was 17 ± 2 mm vs. 17 ± 2 mm by semi-automatic CMR (p = 0.16). Tricuspid annular plane systolic excursion by manual echocardiography measurements was 25 ± 3 mm vs. 25 ± 3 mm by semi-automatic CMR (p = 0.7). Deming regression analyses are F I G U R E 3 Mitral and tricuspid annular plane systolic excursion measurements by the semi-automatic time-resolved algorithm using cardiac magnetic resonance imaging. The outer area of the ventricles is delineated at end-diastole (dotted line) and end-systole (solid line) and reference markers for tracking denoted with closed circles. Two-sided arrows represent peak mitral (1) and tricuspid (2) annular plane systolic excursion tracking. Note how the mitral and tricuspid annulus points differ in position compared with the atrioventricular plane displacement shown in Table 2. Coefficient of variation was lower for MAPSE (3.0%) than TAPSE (13.3%). Fitted up-slopes demonstrated a high inter-method correlation for both MAPSE and TAPSE. Figure 4a,b presents Bland-Altman and scatter plots for both measurements.
Fitted up-slopes demonstrated a high inter-method correlation for both MAPSE and TAPSE (Table 2).

| D ISCUSS I ON
This study showed that both MAPSE and TAPSE measurements by echocardiography and semi-automatic CMR analysis have low variability for each method in a population of healthy adult subjects.   (Ochs et al., 2017). These previous results were thus, despite assessed the same day, of larger variability than the current results. Interestingly, the coefficient of variation between methods was larger between direct measurements by CMR and echocardiography, than between measurements based on AVPD analysis by CMR and echocardiography. However, as indicated by both bias and intercept by Deming regression, the statistically significant differences between methods and modalities are likely not of clinical significance.

TA B L E 1 Demographic and anthropometric characteristics of the study population
Regarding the semi-automatic CMR method used in the current study, it is important to acknowledge that in the original protocol by Seemann et al., (2017), a cine acquisition with 30 images per cardiac cycle was used, whereas, the current protocol used 25 images per cardiac cycle. This is related to scanner settings for clinical routine and is still used in many centres. Nevertheless, this relatively lower setting for temporal resolution could be associated with that the real maximum MAPSE and TAPSE are missed. As echocardiography has a substantially higher temporal resolution this might have affected the comparison between methods. Seemann et al., (2017) also showed correlation between CMR and echocardiography for lateral e´ and E/e´ (r = 0.76; p < 0.0001 and r = 0.85; p < 0.0001) for diastolic function in 59 patients. No comparison was, however, performed for longitudinal systolic function.
There are technical differences between 2D echocardiography and CMR. Despite that echocardiography has a higher temporal resolution than CMR, it may have acoustic window limitations, mainly in extremely thin or obese patients, which could affect acquisition  (Medvedofsky et al., 2015). Foreshortening and off-angle slice positions may affect both echocardiography and CMR, particularly on the right side, and there is thus a risk that TAPSE is measured in a different location by echocardiography and CMR in the current study.
This may in part explain the higher correlation between methods for MAPSE than TAPSE in the current study. Outliers in Figure 4, for ex-

| Limitations
As echocardiography and CMR were not performed the same day, physiological variation may have had an impact on comparisons. To minimize this, subjects were advised to not exercise before examinations, and there is no obvious bias in the results indicating significant change in cardiac function between assessments. Time between echocardiography and CMR could be a potential bias in itself. Quantile regression and trend analysis, however, showed no impact of time between examinations on measurements. The CMR protocol used a lower temporal resolution for cine acquisition than used in the original publication, possibly affecting the correct estimation of maximum excursion of the mitral and tricuspid annuli by CMR. Although increased temporal resolution is advocated, there is no significant systematic bias in the current results indicating that this is crucial for the comparison between methods. Only the apical four-chamber view was used, whereas MAPSE can also be performed using the apical two-chamber view. In the larger inclusion study, however, only the 4-chamber view was assessed. TA B L E 3 Intra-and inter-observer reproducibility of MAPSE and TAPSE by echocardiography and cardiac magnetic resonance (n = 20)

| Consent for publication
All participants provided written consent for publication of any data obtained during the study.

ACK N OWLED G M ENTS
This project has been co-funded with support of the Erasmus + Programme of the European Union (Framework Agreement number: 2013-0040). This publication reflects the views only of the authors, and the Commission cannot be held responsible for any use that may be made of the information contained therein.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R S' CO NTR I B UTI O N S
ASM, FC and EH contributed substantially to the study design. ASM drafted the manuscript. ASM, KSE, BVA and EH analysed data. All authors interpreted data and revised the manuscript critically for important intellectual content, have provided final approval of the version to be published and have agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated and analysed during the current study are not publicly available due to the Spanish law of data protection in research but are available from the corresponding author on reasonable request.