Assessing left atrial function in patients with atrial fibrillation and valvular heart disease using cardiovascular magnetic resonance imaging

Abstract Background Atrial fibrillation (AF) is common arrhythmia in valvular heart disease (VHD) and is associated with adverse outcomes. Hypothesis To evaluate the left atrial (LA) function in patients with AF‐VHD by cardiovascular magnetic resonance imaging feature tracking (CMR‐FT) using LA strain (ε s/ε e/ε a) and their corresponding strain rate (SRs/SRe/SRa). Methods This was a retrospective cross‐sectional inter‐reader and intra‐reader reproducibility conducted from July 1, 2020, to January 31, 2021. A total of 39 patients with AF‐VHD (rheumatic heart valvular disease [RHVD] [n = 22], degenerative heart valvular disease [DHVD] [n = 17]) underwent MRI scans performed with drug‐controlled heart rate before correcting the rhythm and valves through maze procedure. Fifteen participants with normal cardiac MRI were included as healthy control. ε s/SRs, ε e/SRe, and ε a/SRa, corresponding to LA reservoir, conduit, and booster‐pump function, were assessed using Feature Tracking software (CVI42 v5.12.1). Results Compared with healthy controls, LA global strain parameters (ε s/ε e/ε a/SRs/SRe/SRa) were significantly decreased (all p < 0.001), while LA size and volume were increased in AF‐VHD group (all p < 0.001). In the subgroup, RHVD group showed lower LA total ejection fraction (LATEF) and strain data than DHVD group (12.6% ± 3.3% vs. 19.4 ± 8.6, p = 0.001). Decreased LATEF was significantly related to altered LA strain and strain rate, especially in ε s, ε e, and SRs (Pearson/Spearman r/ρ = 0.856/0.837/0.562, respectively; all p < 0.001). Interstudy and intrastudy reproducibility were consistent for LA volumetry and strain parameters (intraclass correlation coefficient: 0.88–0.99). Conclusions CMR‐FT can be used to assess the LA strain parameters, and identify LA dysfunction and deformation noninvasively, which could be a helpful functional imaging biomarker in the clinical treatment of AF‐VHD.


| INTRODUCTION
Atrial fibrillation (AF) is the most common human tachyarrhythmia diagnosed clinically; patients with AF are at an increased risk of stroke and heart failure, in addition to a decreased quality of life and lower survival. 1,2 AF is associated with profound structural and functional alterations of the atrial myocardium, 3 promoting a true atrial cardiomyopathy, the severity of which is an important determinant of AF recurrence and response to treatment. [4][5][6] Stroke prevention is a pivotal part of the treatment of patients with AF. 7 Patients with AF and concurrent valvular heart disease (AF-VHD) have an even higher thromboembolic risk than those with AF alone. 7,8 Patients with AF have a 5-fold increased risk of stroke compared with patients without cardiovascular disease, and patients with AF coupled with mitral stenosis have a 20-fold risk of stroke. 9 It is of major clinical interest to have a new imaging biomarker with which to quantify the degree of LA dysfunction and make earlier clinical decisions in patients with AF-VHD in an effort to prevent cardiac events. The LA normal function includes reservoir (collection of pulmonary venous blood during left ventricular [LV] systole), conduit (passage of pulmonary venous blood flow to the LV during LV early-diastole), and booster pump function (augmentation of LV filling during LV late-diastole/atrial systole). 10 It is important to recognize the interplay that exists between these atrial functions and ventricular performance throughout the cardiac cycle. Previous publications investigating LA function have primarily focused on LA size and volume. 11 Due to the LA's complex geometry and intricate fiber orientation and the variable contributions of its appendage and pulmonary veins, these two parameters alone may be insufficient to describe the complexity of the LA function and wall motion. [12][13][14] LA strain assessed by cardiovascular magnetic resonance imaging feature tracking (CMR-FT) has been used in many cardiovascular diseases and enhanced the diagnostic value, [12][13][14] which might be higher and more sensitive than conventional LA volumetric parameters 12 and LV function 13,14 and presented with good intra-observer and interobserver reproducibility in normal volunteers. 15 Due to the very thin LA wall, it is challenging to measure the LA strain. 15 And radial strain has already been noted for its difficulty to be obtained with poor reproducibility. 15,16 Echocardiographic data also showed poor reproducibility of strain rate and radial strain 16 and was limited in the two-dimensional approaches with a semi-quantitative and subjective measure. 17,18 It is clear that studies of LA function provide new insights into the contribution of LA performance to cardiovascular disease and are promising tools for predicting cardiovascular events in a wide range of patient populations. Considerable data also support the use of LAEF for predicting events. 12 Accordingly, LA function indices such as strain and strain rate have been proposed using noninvasive imaging modalities such as echocardiography speckle tracking 19,20 and color tissue doppler. Although speckle tracking is presently the only available reference for LA strain estimation, ultrasound beam direction as well as heart motion relative to the probe may influence measurements and inter-vendor variability, essentially described in the setting of LV function, need to be further investigated. 12,21 CMR-FT is a new method to evaluate myocardial strain and strain rate; it can be applied to routine cine images and has the advantages of high spatial resolution, large field of view, good reproducibility, and it can more sensitively reflect the functional characteristics of myocardial tissue. 22 As a new technique, CMR-FT has been mainly used in the study of LV strain in recent years, 23 but has rarely been applied to analysis of LA. The aim of this study was to evaluate LA strain and strain rate using CMR-FT before valve replacement surgery, assess the feasibility and reproducibility of CMR-FT for the quantification of global LA function in patients with AF-VHD, and review the clinical application value of CMR-FT. In addition, we compared global LA function between patients with AF-VHD and those without cardiac disease.

| MATERIAL AND METHODS
This study was approved by the local ethics committee. All patients or their families provided written informed consent for the examination.

| Patient selection
From July 1, 2020, to January 31, 2021, 39 consecutive patients with AF-VHD who were admitted to the department of cardiac surgery in our hospital, and performed MRI to assess the left atrium function before valve replacement surgery only or valve replacement surgery with Maze procedure to correct valves and rhythm were included in this study. All patients were with persistent AF and mitral stenosis.
The clinical characteristics, echocardiography findings on admission, and cardiac MRI data were retrospectively collected.
The inclusion criteria 24 were as follows 1 : Patients' age >18 years with AF who were planning to undergo cardiac surgery to correct the rhythm and valves 2 ; AF of more than 30 s as recorded by electrocardiogram (ECG) or dynamic ECG; and 3 patients who were treated according to the American Heart Association/American College of Cardiology/Heart Rhythm Society (AHA/ACC/HRS) AF guidelines for the management of AF 9 and VHD. 4,25 According to etiological analysis, the patients divided into two groups, degenerative heart valvular disease (DHVD) group and rheumatic heart valvular disease (RHVD) group. Exclusion criteria included patients who had undergone previous valvular surgery or ablation for AF.
At the same time, 15 healthy participants with normal cardiac MRI in our hospital were included as the healthy control group.

| Imaging analysis
Images were analyzed using CVI42 (Circle, version 5.12.1).  (4) The intraobserver and interobserver variability for the LA parameters, measurements were assessed by the intraclass correlation coefficient (ICC). 28 Intraobserver reproducibility was established by the same observer (J. H., 5-year experience in CMR diagnosis) who reanalyzed the same subjects after 1 month.
Interobserver reproducibility was assessed by a second independent observer (Y. S., 3-year experience in CMR diagnosis) who was blinded to the first observer's results.
The left atrial (LA) strain and strain rate curve. Global endocardial LA strain and strain rate values (yellow line) were recorded. SRa, peak late negative strain rate; SRe, peak early negative rate; SRs, peak positive strain rate; ε s , total strain; ε a , active strain; ε e , ε s −ε a , passive strain. ε s and SRs, corresponding to LA reservoir function; ε e and SRe, corresponding to LA conduit function; ε a and SRa, corresponding to LA booster pump function

| Statistical analysis
Data were analyzed using SPSS, version 20.0 (SPSS Statistics, IBM Corporation). Mean ± SD or median (quartiles) was used to express measurement data in accordance with normal distribution, and the continuous variables were analyzed by independent sample t-test or Mann-Whitney U test. χ2 or Fisher exact test was used to assess categorical data. Pearson or Spearman correlation was performed to investigate the potential relationship between LA strain parameters and LA function. Moreover, we assessed the ICC to evaluate the accuracy and the precision of the method to measure each LA parameters. ICC was scored as follows: poor reliability, ICC < 0.50; moderate reliability, ICC: 0.50-0.75; good reliability, 0.75-0.9; excellent reliability, ICC > 0.9. 28 p-value < 0.05 was considered statistically significant.

| Reproducibility
Intraobserver and interobserver reproducibility of global LA strain, strain rate, and volumetric parameters using CMR are shown in Table 3. There were excellent and good intra-observer and interobserver reproducibility, respectively. For intra-observer repro-

| DISCUSSION
Our study evaluated the LA function in patients with AF-VHD using CMR-FT. We found that LA function in AF -VHD was lower than the Control healthy participants. AF-VHD patients with reduced reservoir and conduit function, reduced or absent booster pump function, which showed in strain parameters. LATEF has a linear correlation with LA strain parameters. LA enlargement was also observed in AF-VHD, which means LA remodeling. In the subgroup, DHVD had higher LV size, volume, mass, and stain parameters than those in RHVD. And LA strain and volumetric parameters showed good reproducibility.
LA remodeling consists of both structural and functional changes; enlargement of the LA and fibrosis of the atrial muscle promotes the persistence of AF, 29 significantly affecting the LA function. The pathophysiology of AF is complex and incompletely understood. 2 Currently, it is believed that AF-induced electrical alterations occur within the atrial myocardium (electrical remodeling), which may also promote or accelerate myocardial apoptosis and fibrosis (anatomical remodeling), 2,30 then the process becomes self-perpetuating. 31 LA size by volumetric index is widely accepted as a significant prognostic marker of mortality and outcomes in many cardiovascular diseases. 11,32 Le Tourneau T et al. 33     0.001* Note: "*" indicates statistical significance. "a"/"b" indicates statistical significance between the control group and RHVD group/DHVD group.

| LIMITATION
This was a single-center study with a relatively modest sample size.
We only assessed global longitudinal strain and did not assess radial strain. Radial strain has been noted as a parameter that is difficult to obtain and reproduce consistently. 16 At present, most research investigates the longitudinal LA strain on the LA long axis, and most of the studies have obtained positive results. It is believed that with the application of three-dimensional strain analysis technology in the future, its accuracy will be further improved.
F I G U R E 2 Scatter plots showing correlations of left atrial total emptying fraction (LATEF) and ε s , ε e , ε a , SRs, SRe, and SRa. SRa, peak late negative strain rate; SRe, peak early negative rate; SRs, peak positive strain rate; ε a , active strain; ε e , passive strain; ε s , total strain T A B L E 3 Intraobserver and interobserver repeatability of LA strain and strain rate Abbreviations: ICC, intraclass correlation coefficient; LA, left atrial; LATEF, left atrial total ejection fraction; LAV max / min , the maximum/ minimum volume of left atrium; SRa, peak late negative strain rate; SRe, peak early negative rate; SRs, peak positive strain rate; ε a , active strain; ε e , passive strain; ε s , total strain.