The present prospective study was designed to evaluate the accuracy of quantitative assessment of mitral regurgitant fraction (MRF) by echocardiography and cardiac magnetic resonance imaging (cMRI) in the modern era using as reference method the blinded multiparametric integrative assessment of mitral regurgitation (MR) severity. 2-Dimensional (2D) and 3-dimensional (3D) MRF by echocardiography (2D echo MRF and 3D echo MRF) were obtained by measuring the difference in left ventricular (LV) total stroke volume (obtained from either 2D or 3D acquisition) and aortic forward stroke volume normalized to LV total stroke volume. MRF was calculated by cMRI using either (1) (LV stroke volume − systolic aortic outflow volume by phase contrast)/LV stroke volume (cMRI MRF [volumetric]) or (2) (mitral inflow volume − systolic aortic outflow volume)/mitral inflow volume (cMRI MRF [phase contrast]). Six patients had 1 + MR, 6 patients had 2 + MR, 12 patients had 3 + MR, and 10 had 4 + MR. A significant correlation was observed between MR grading and 2D echo MRF (r = 0.60, p <0.0001) and 3D echo MRF ( r = 0.79, p <0.0001), cMRI MRF (volumetric) (r = 0.87, p <0.0001), and cMRI MRF (phase contrast r = 0.72, p <0.001). The accuracy of MRF for the diagnosis of MR ≥3+ or 4+ was the highest with cMRI MRF (volumetric) (area under the receiver-operating characteristic curve [AUC] = 0.98), followed by 3D echo MRF (AUC = 0.96), 2D echo MRF (AUC = 0.90), and cMRI MRF (phase contrast; AUC = 0.83). In conclusion, MRF by cMRI (volumetric method) and 3D echo MRF had the highest diagnostic value to detect significant MR, whereas the diagnostic value of 2D echo MRF and cMRI MRF (phase contrast) was lower. Hence, the present study suggests that both cMRI (volumetric method) and 3D echo represent best approaches for calculating MRF.
Mitral regurgitant fraction (MRF) can be calculated by 2-dimensional (2D) or 3-dimensional (3D) echocardiography or using cardiac magnetic resonance imaging (cMRI). Landmark reports have suggested that cMRI may be an appropriate and reproducible tool for the quantification of mitral regurgitation (MR). In contrast, a poor correlation between cMRI and echocardiography for MR grading was recently reported. In this series, 11 of 38 patients (29%) with moderate-to-severe MR on echocardiography and American College of Cardiology/American Heart Association class I or IIa indication of mitral valve replacement/repair, who underwent surgery, had only mild MR using cMRI, questioning the reliability of cMRI to assess the severity of MR. Hence, the present prospective study was designed to assess the reliability of cMRI for the quantification of MR and to compare the quantitative assessment of MRF by echocardiography and cMRI.
Methods
Consecutive patients referred to our echocardiography laboratory for MR evaluation during a 3-month period were prospectively screened. Inclusion criteria were at least mild grade (1+) isolated MR secondary to mitral valve billowing or prolapse. Exclusion criteria were nonsinus cardiac rhythm, unstable clinical condition, inability to maintain breath-hold for at least 6 seconds or to measure left ventricular (LV) forward stroke volume, poor echocardiographic windows, presence of secondary MR, aortic valvular disease, atrial or ventricular septal defects, MRI-incompatible devices or claustrophobia, and unwillingness to provide informed consent as required by our institutional review board.
Two cardiologists (CT and SM), experts in echocardiography and valvular heart diseases, blinded to 3D and cMRI data sets independently performed MR grading using an integrative approach according to the European Association for Cardio-Vascular Imaging recommendations. Echocardiograms were performed on a Vivid E9 ultrasound scanner (BT12; General Electric Healthcare, Horten, Norway) equipped with both 2D (M5S-D) and 3D (4V-D) probes. Two-dimensional data were acquired according to current recommendations. 3D echocardiography full-volume data sets of the LV were obtained in an adjustable volume divided into 6 subvolumes. LV forward stroke volume was calculated as the product of LV outflow tract velocity-time integral by LV outflow tract cross-sectional area. 2D and 3D MRF (2D echo MRF and 3D echo MRF) were obtained off-line (EchoPAC workstation [BT12; General Electric Healthcare]) by measuring the difference in LV total stroke volume (obtained from either 2D or 3D acquisition) and aortic forward stroke volume normalized to LV total stroke volume, by a third investigator, blinded to MR grading of the 2 experts and without access to echocardiography-Doppler data used for the integrative multiparametric approach.
Cardiac MRI was performed using a 3-T scanner General Electrics discovery 750w equipped with 24-channel GE Torso phased array coil for signal reception. Image acquisition was performed during short repetitive end-expiratory breath-holding. LV volumes were assessed using balanced steady state–free precession gradient echo cine imaging with retrospective gating (repetition time 4 ms, echo time 1.8 ms, flip angle 55°, 30 phases per cardiac cycle, spatial resolution 2.2 × 1.8 × 8.0 mm) in short-axis orientation with full ventricular coverage. Velocity mapping was assessed using plane phase-contrast imaging (repetition time 5.3 ms, echo time 3 ms, flip angle 20, 45 phases per cardiac cycle, spatial resolution 2 × 2 × 6 mm 3 , velocity encoding 250 cm/s for aorta, and 120 cm/s for mitral flow, with individual adaptation of velocity encoding to avoid aliasing). Aortic phase contrast was performed 10 mm above the tip of the aortic valve perpendicular to the aorta. Short-axis mitral phase contrast was acquired through plane phase–encoding setup at the mitral annular plane or slightly into the LV using both horizontal and vertical long-axis cine images diastolic frames. LV end-diastolic and end-systolic, hence, stroke volumes were determined from short-axis cine images according to the Simpson method. Aortic outflow volume was derived from quantitative flow measurements. cMRI MRF (volumetric) and cMRI MRF (phase contrast) were calculated as shown in Figure 1 , with image analysis performed using an off-line workstation ADW cardiac VX (General Electrics Healthcare) by a fourth investigator (MT), blinded to all echocardiography and clinical data. Both echocardiography and cMRI examinations were performed during the same half-day session.
Continuous variables were expressed as mean ± SD and categorical variables as absolute numbers and percentages. The relations between MR severity and MRF by echocardiography or cMRI were examined by the Spearman rank-order correlations. Comparisons among ≥3 groups were performed using the Kruskal-Wallis test with post hoc comparisons performed using the Mann-Whitney U test. Receiver-operating characteristic curves were built to evaluate the diagnostic value of each parameter. Cut-off values that maximized sensitivity and specificity were obtained. Inter-reader variability of MR grading was assessed using the weighted Kappa statistic. Analyses and figures were obtained using PASW 18.0 (IBM, Inc.; Bois-Colombes, France), GraphPad Prism (GraphPad Software, La Jolia, California), and MedCalc for Windows, version 12.5.0 (MedCalc Software, Mariakerke, Belgium).
Results
Thirty-four patients (mean age 57 ± 19 years, 71% men) were enrolled in the present study. Mean systolic blood pressure, diastolic blood pressure, and heart rate were 131 ± 14 mm Hg, 70 ± 8 mm Hg, and 69 ± 11 beats/min, respectively. Twelve patients received β blockers, 10 angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, and 6 loop diuretics. Based on the integrative multiparametric approach, 6 patients had 1 + MR (18%), 6 patients 2 + MR (18%), 12 patients 3 + MR (35%), and 10 had 4 + MR (29%). MR jet was eccentric in 24 patients. Mean values of vena contracta width, mitral-to-aortic time velocity integral ratio, E-wave velocity, mitral effective regurgitant orifice area, regurgitant volume, and frequency of systolic flow reversal according MR grading are listed in Table 1 . The Kappa coefficients for the inter-reader concordance in 4 grades (1+/2+/3+/4+) or in 2 categories (1+/2+ vs 3+/4+ patients) were 0.92 and 0.94, respectively. LV end-diastolic and end-systolic volumes were larger with cMRI compared with 3D echo (183 ± 50 vs 157 ± 34 and 75 ± 30 vs 55 ± 15, p <0.0001, respectively) and compared with 2D echo (183 ± 50 vs 162 ± 38 and 75 ± 30 vs 56 ± 18, p <0.0001, respectively). Mean values for MRF by 2D echo, 3D echo, and cMRI are detailed in Table 2 . A significant correlation was observed between MR grading and MRF obtained by 2D echo ( r = 0.60, p <0.0001, Figure 2 ), by 3D echo ( r = 0.79, p <0.0001, Figure 2 ), by cMRI (volumetric, r = 0.87, p <0.0001, Figure 2 ), and by cMRI (phase contrast, r = 0.72, p <0.001, Figure 2 ). Comparisons between individual grades for MRF by cMRI (volumetric) and 3D echo were all significant except for 1+ and 2 + MR, whereas comparisons between individual grades for MRF by 2D echo and cMRI (phase contrast) were significant only for 3+ versus 4 + MR (p = 0.021) and 1+ versus 2 + MR (p = 0.007), respectively. The accuracy of the MRF (area under the receiver-operating characteristic curve) for the diagnosis of MR ≥3+ or 4+ was the highest with cMRI MRF (volumetric) and by 3D echo MRF, followed by 2D echo MRF and cMRI MRF (phase contrast; Table 3 , Figure 3 ). All patients with a 3D echo MRF ≥35% had 3+ or 4 + MR, whereas those with 3D echo MRF ≤30% had 1+ or 2 + MR, with a “gray” overlap zone involving 23% of the study population ( Figure 4 ). All patients with cMRI MRF (volumetric) >35% had severe MR, whereas those with cMRI MRF (volumetric) ≤26% had 1+ or 2 + MR, 15% of the study population being in the gray overlap zone ( Figure 4 ). Comparatively, all patients with 2D echo MRF ≥35% had 3+ or 4 + MR and those with 2D echo MRF ≤20% had 1+ or 2 + MR, with a large “gray” overlap zone (RF between 20% and 35%) involving 44% of the study population ( Figure 4 ). Last 56% of the patients were in the gray overlap zone for cMRI MRF (phase contrast; Figure 4 ).
| Mitral regurgitation severity | Overall P-value | ||||
|---|---|---|---|---|---|
| 1+ (n=6) | 2+ (n=6) | 3+ (n=12) | 4+ (n=10) | ||
| Vena contracta width (mm) | 2 ± 1 | 5 ± 2 | 7 ± 2 | 7 ± 2 | 0.001 |
| Mitral to aortic time velocity integral ratio | 0.96 ± 0.22 | 1.29 ± 0.26 | 1.55 ± 0.45 | 1.72 ± 0.32 | 0.002 |
| E wave (m/s) | 0.62 ± 0.10 | 1.15 ± 0.22 | 1.08 ± 0.20 | 1.16 ± 0.24 | 0.002 |
| Effective regurgitant orifice area (mm 2 ) ∗ | 6 ± 2 | 25 ± 8 | 38 ± 15 | 62 ± 32 | <0.0001 |
| Regurgitant volume (mL) ∗ | 8 ± 4 | 39 ± 7 | 63 ± 10 | 85 ± 32 | <0.0001 |
| Systolic pulmonary flow reversal | 0 (0%) | 1 (17%) | 5 (42%) | 8 (80%) | 0.007 |
∗ Effective regurgitant orifice area and regurgitant volume were obtained by the proximal isovelocity surface area method in 30 on 34 patients (88%), 5 patients with mild mitral regurgitation, 4 with mild to moderate mitral regurgitation, 11 with moderate to severe mitral regurgitation and 10 with severe mitral regurgitation.
| Overall population (n=34) | Mitral regurgitation severity | Overall P-value | ||||
|---|---|---|---|---|---|---|
| 1+ (n=6) | 2+ (n=6) | 3+ (n=12) | 4+ (n=10) | |||
| 2D Echo mitral regurgitant fraction, % | 37 ± 16 | 21 ± 7 | 26 ± 10 | 38 ± 13 | 52 ± 11 ∗ | <0.0001 |
| 3D Echo mitral regurgitant fraction, % | 35 ± 16 | 18 ± 10 | 19 ± 14 | 38 ± 7 † | 52 ± 7 ∗ | <0.0001 |
| cMRI mitral regurgitant fraction (volumetric), % | 36 ± 18 | 14 ± 10 | 19 ± 12 | 41 ± 7 † | 54 ± 8 ∗ | <0.0001 |
| cMRI mitral regurgitant fraction (phase contrast), % | 20 ± 19 | 5 ± 8 | 17 ± 13 ‡ | 17 ± 13 | 38 ± 17 | 0.001 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
