Our study was a 1-year cross-sectional, open-labeled, non-randomized, single cohort study conducted at a single cardiac center in a tertiary care hospital. Investigators weren’t blinded to the study group. The study design was approved by the hospital ethics committee review board, all participants signed written informed consents, study procedures were carried out following the Code of Ethics of the World Medical Association (Declaration of Helsinki), all information/images were anonymized, and the privacy rights of the study participants were observed diligently.

Study participants were patients referred to the cardiology department. They were subjected to history taking and data collection for age, gender, risk factors especially hypertension, diabetes mellitus (DM), chronic kidney disease, coronary artery disease, smoking and hyperlipidemia, nature and course of ischemic, pulmonary congestion and low cardiac output symptoms, and comprehensive clinical examination.

Thirty eligible participants were enrolled, consecutively assigned and allocated in a single cohort, and underwent TTE using Phillips EPIQ 7 ultrasound system for cardiology equipped with xMATRIX transducer (X5-1). Data documented with TTE included jet dimensions, vena contracta, mitral regurgitant volume (RVol) quantification by proximal isovelocity surface area (Echo_RVol), effective regurgitant orifice area, mitral regurgitant fraction (RF) determined by the quantitative Doppler method (Echo_RF), early diastolic mitral E wave velocity (E wave), left atrial (LA) volume, and left ventricular (LV) volume and function by the modified Simpson’s biplane method. MR was categorized based on Echo_RVol and Echo_RF into mild ( 30 mL/beat), moderate (30-44 mL/beat based on Echo_RVol vs 30-39 mL/beat based on Echo_RF), moderate to severe (45-59 mL/beat based on Echo_RVol vs 40-49 mL/beat based on Echo_RF) and severe ($\ge$ 60 mL/beat based on Echo_RVol vs $\ge$ 50 mL/beat based on Echo_RF) grades (Table 1) [5]. Screened participants were enrolled if they had isolated MR and Echo_RVol $\ge$ 30 mL/beat. Screened participants with multiple valvular heart disease, Echo_RVol 30 mL/beat, contraindications for CMR including cardiac pacemaker or mechanical prosthesis, insulin pumps and metallic intraocular objects, incomplete or suboptimal TTE study, or pregnancy were excluded from the study.

The enrolled study participants underwent cardiac magnetic resonance (CMR) within one week for quantification of MR using Siemens Magnetom Aera 1.5T magnetic resonance imaging scanner. Data documented with CMR include indirect quantification of mitral RVol by measuring the difference between left and right ventricle stroke volumes (CMR_RVol${}_1$) and direct quantification of mitral RVol by phase-contrast velocity mapping (CMR_RVol${}_2$) (Table 2) (Figs. 1) [6].

Fig. 1.

Echo and CMR images of patient with severe eccentric MR. Showing severe primary MR with eccentric jet by echo (Rvol 62 mL, EROA 0.48 cm${}^2$) showing excellent correlation with CMR-Rvol${}_1$ (Rvol 59 mL, RF 50%). And moderate correlation with CMR_RVol${}_2$ (Rvol${}_2$ 28.92 mL, RF 32%). From A to D are Transthoracic echo images showing; (A) Parasternal Long axis view showing MV bileaflets prolapse with posteriorly directed eccentric MR jet. (B) short axis view at basal LV. (C) proximal isovelocity surface area (PISA) MR with radius 1.0 cm, Nyquist limit 39.9 cm/s. (D) Continuous wave (CW) doppler showing Vmax and VTI of MR jet. Form E to G are CMR images showing. (E) Short axis slice showing LV end systolic volume. (F) Short axis slice showing LV end diastolic volume. (G) Phase contrast velocity encoding (VENC) showing total forward flow (mitral valve stroke volume = 88.87 mL), flow reversal (CMR_ Rvol${}_2$ = 28.92 mL).

The study evaluated the correlation between Echo_Rvol and CMR_RVol${}_1$, and the correlation between Echo_RVol and CMR_RVol${}_2$. Secondary objectives included assessment of the inter-observer reliability of the MR grade between TTE and CMR.

The echocardiographic and CMR assessment outcomes were coded, and the data was analyzed with the Statistical Package for the Social Sciences software for MAC (SPSS®) version 26. Quantitative (continuous) data was expressed as means and standard deviations, while qualitative (categorial) data was expressed as medians and ranges. Parametrically distributed quantitative variables were compared with the Independent two-tailed t-test and correlated with Pearson’s Correlation Coefficient, respectively. Pearson’s Correlation Coefficient (r) value was interpreted as follows: 1 means perfect correlation, 0.75-1.0 is strong correlation, 0.50-0.75 is moderate correlation, 0.25-0.5 is weak correlation, and 0.25 means no relationship. The intraclass correlation coefficient (ICC), two-way random model and absolute agreement type were used for comparison between quantitative measurements. The scale of ICC ranges from 0 to 1 where 1 represents perfect reliability with no measurement error and 0 represents no reliability. ICC Value 0.5 indicates poor reliability, 0.5-0.75 indicates moderate reliability, 0.75-0.9 indicates good reliability and greater than 0.90 indicates excellent reliability [7]. Bland-Altman plot was used for the evaluation of bias between the means of the two modalities within 95% limits of agreement [8]. For categorial agreement, Cohen’s kappa coefficient ($\kappa$) was used. Kappa values $\le$ 0 means no agreement, 0.01-0.20 means none to slight agreement, 0.21-0.40 means fair agreement, 0.41-0.60 means moderate agreement, 0.61-0.80 means substantial agreement, and 0.81-1.00 means almost perfect agreement [9]. The confidence interval was set to 95% and the margin of error accepted was set to 5%. Any comparison considered statistically significant was at P 0.05 or less.

We recruited 30 patients from one hospital in one country from January 2019 through December 2019. The study group was balanced with regards to baseline characteristics and risk factors. The key sociodemographic feature of enrolled participants was equal gender representation (Mean age 52.7 $\pm$ 19.3 years, 50% males, 50% females). All enrolled participants completed the study and there were no withdrawals. As per TTE assessment, 56.7% had moderate MR, 26.7% had moderate-severe MR, and 16.7% had severe MR. On the other hand, 16.7% had mild MR, 40.0% had moderate MR, 33.3% had moderate to severe MR and 10% had severe MR as per CMR assessment. Primary MR was observed in 12 (40.0%) patients, secondary MR in 18 (60.0%) patients, central jet in 20 (66.7%) patients, and eccentric jet in 10 (33.3%) patients. Twenty (66.7%) patients had normal left ventricular systolic function with ejection fraction $\ge$ 60%, while 10 (33.3%) patients had left ventricular ejection fraction 60% (Table 3).

The mean left ventricle end-diastolic volume (LVEDVi) was 80.2 $\pm$ 33.5 mL/m${}^2$ by TTE vs 103.1 $\pm$ 29.3 mL/m${}^2$ by CMR, the mean left ventricle end-systolic volume (LVESVi) was 44.6 $\pm$ 30.5 mL/m${}^2$ by TTE vs 60.8 $\pm$ 33.4 mL/m${}^2$ by CMR, the mean left ventricle stroke volume (LVSVi) was 34.3 $\pm$ 15.8 mL/m${}^2$ by TTE vs 42 $\pm$ 16.7 mL/m${}^2$ by CMR, the mean left atrial volume (LAVi) was 45.8 $\pm$ 16.1 mL/m${}^2$ by TTE vs 46.8 $\pm$ 18.1 mL/m${}^2$ by CMR, and the mean left ventricle ejection fraction (LVEF) was 47 $\pm$ 19.2 % by TTE vs 44.7 $\pm$ 19.7 % by CMR, respectively. There was a statistically significant good reliability between the mean LVEDVi by TTE and the mean LVEDVi by CMR, the mean LVESVi by TTE and the mean LVESVi by CMR, the mean LVSVi by TTE and the mean LVSVi by CMR, the mean LAVi by TTE and the mean LAVi by CMR, and a statistically significant excellent reliability between the mean LVEF by TTE and the mean LVEF by CMR (ICC = 0.839, 0.888, 0.786, 0.751, and 0.956, respectively, P $\le$ 0.0001) (Table 4).

The mean mitral RVol was 45.0 $\pm$ 12.99 mL/beat by TTE quantification, 42.75 $\pm$ 12.48 mL/beat by indirect CMR quantification, and 34.23 $\pm$ 11.66 mL/beat by direct CMR quantification. There was a statistically significant strong positive correlation between Echo_RVol and CMR_RVol${}_1$ (r = 0.753, n = 30, P 0.001) (Figs. 2) and a statistically significant good reliability between Echo_RVol and CMR_RVol${}_1$ (ICC = 0.859, 95% CI 0.706-0.932, P $\le$ 0.0001) (Table 5), a statistically significant moderate positive correlation between Echo_RVol and CMR_RVol${}_2$ (r = 0.530, n = 30, P 0.003) (Figs. 3), and statistically significant moderate reliability between Echo_RVol and CMR_RVol${}_2$ (ICC 0.557, P$\le$ 0.001). This agreement between Echo_RVol and CMR_RVol${}_1$ was statistically significant with excellent reliability in patients with normal EF (LVEF $\ge$ 60%) (ICC 0.942, 95% CI 0.800-0.983, P $\le$ 0.0001), but non-significant with moderate reliability in patients with low EF (LVEF 60%) (ICC 0.527, 95% CI 0.263-0.823, P = 0.066). Bland-Altman plot demonstrated bias of 2.24 units between the means of Echo_RVol and CMR_RVol${}_1$ within 95% limits of agreement (95% CI 15.12-19.60) (Figs. 4).

Fig. 2.

The correlation between Rvol by echo vs CMR by indirect quantification. Simple Scatter plot with fit line demonstrating statistically significant strong positive correlation between echo mitral regurgitant volume (Echo_RVol) and CMR mitral regurgitant volume by indirect quantification (CMR_RVol${}_1$) (r 0.753, P 0.00001).

Fig. 3.

The correlation between Rvol by Echo vs CMR by direct quantification (velocity mapping). Simple scatter plot with fit line demonstrating positive moderate correlation between Echo-RVol and CMR-RVol by velocity mapping (CMR_RVol${}_2$) (r 0.530, P 0.003).

Fig. 4.

Bland-Altman plot showing bias of 2.24 units between the means of RVol by Echo vs CMR by indirect quantification within 95% limits of agreement. Bland-Altman Plot demonstrating bias of 2.24 units between the means of Echo-RVol vs CMR_RVol${}_1$ within 95% limits of agreement (95% CI 15.12-19.60).

The inter-observer reliability of the MR grade between TTE and CMR was seen in 20 (66.4%) patients showing statistically significant moderate agreement ($\kappa$ = 0.502, P = 0.0001) when the observers used Echo_RVol for grading of MR (Table 6, Figs. 5), while the inter-observer reliability of the MR grade between TTE and CMR was seen in 15 (50%) patients showing statistically significant faint agreement ($\kappa$ = 0.251, P = 0.024) when the observers used Echo_RF for grading of MR (Table 7, Figs. 6).

Fig. 5.

Bar chart showing the patients with the same versus different MR category between echo versus CMR based on Rvol. 1. Blue; Patients with mild MR, 2. Red; Moderate MR, 3. Green; Moderate to severe MR and 4. Orange; Severe MR.

Fig. 6.

Bar chart showing the patients with the same versus different MR category between echo versus CMR based on RF. 1.Blue; Patients with mild MR, 2. Red; Moderate MR, 3. Green; Moderate to severe MR and 4. Orange; Severe MR.

Our study didn’t have missing data allowing robust per-protocol analysis, the investigators who analyzed and reported the CMR assessment outcomes were blinded to the participants’ clinical data and TTE results, and a second investigator blinded to the CMR measurements reported by the first investigator repeated them at a different time interval to evaluate the inter-observer reproducibility. On the other hand, the study has important limitations. It was a single centered study with small sample size. Being a cross-sectional study with a lack of lengthy follow up didn’t allow us to investigate the chronological relationship between the TTE results, the CMR measurements, and the clinically driven outcomes.