Assessment of left ventricular volumes and primary mitral regurgitation severity by 2D echocardiography and cardiovascular magnetic resonance

The present study concerned 41 prospectively included patients in 2 Belgian centres (CHU Sart Tilman Liège and University of Antwerp Hospital) with at least moderate primary MR (effective regurgitant orifice [ERO] ≥ 20 mm² and/or a RVol ≥ 30 ml) and without overt signs of LV dysfunction or dilatation (LV ejection fraction (EF) ≥60% and LV end-systolic diameter ≤45 mm) after informed consent. Patients with poor echogenicity, history of coronary artery disease, persistent arrhythmias or other significant concomitant valvular heart disease (> mild mitral/aortic stenosis or regurgitation) were excluded. Moreover, we did not enrol patients with CMR-incompatible devices or claustrophobia. The ethical committee of the 2 centres approved the study protocol (Ethisch comité Universitair Ziekenhuis Antwerpen and Comité d'éthique Hospitalo-Facultaire Universitaire de Liège).

We performed a comprehensive 2D TTE in all patients using Vivid 7 or Vivid 9 cardiovascular ultrasound system (GE Healthcare, Little Chalfont, UK). All data obtained by echocardiography were analyzed off-line with an EchoPAC workstation (GE Vingmed Ultrasound AS, Horten, Norway). End-systolic and end-diastolic LV diameters were measured with M-mode in the parasternal long-axis view according to current recommendations[9] with subsequent calculation of the LVEF by the Teichholz formula. LV volumes and LVEF were quantified by modified Simpson’s method in the apical four- and two-chamber view (Figure 1A). The severity of MR was assessed as recommended[3]. The ERO was quantified using the proximal isovelocity surface area (PISA) method (Figure 2A). RVol was assessed by the PISA method, as well as by the Doppler volumetric method. Moderate primary MR was defined as an ERO between 20 mm² and 40 mm³ and/or a RVol between 30 ml and 60 ml while an ERO ≥ 40 mm² and/or a RVol ≥ 60 ml defined severe MR.

CMR imaging was performed using a 1.5-T scanner (Symphony TIM, Siemens, Erlangen, Germany). Breath-hold ECG-gated steady-state free precession sequences in standard long-axis and multiple parallel short-axis slices were used for the assessment of end-systolic and end-diastolic LV dimensions and volumes (Figure 1B). LV stroke volume (end-diastolic LV volume – end-systolic LV volume) and the LVEF ([end-systolic volume/end-diastolic volume] × 100) were calculated. In 22 patients of the study group, we measured the anatomic regurgitant orifice (ARO) by planimetry of the regurgitant orifice in a slice parallel to the valvular plane and perpendicular to the regurgitant jet at mid-systole as previously described by Buchner et al.[8] (Figure 2B). The antegrade LV stroke volume was obtained by phase contrast velocity mapping at the ascending aorta[6]. CMR RVol was calculated as the difference between the LV stroke volume and the antegrade LV stroke volume. All CMR data were assessed by agreement of 2 readers (PJB and LD) who were blinded to echocardiographic data.

An important finding of the present study is that LV volumes are significantly underestimated by 2D TTE in patients with moderate to severe primary MR in comparison with CMR, which is widely accepted as the most accurate non-invasive imaging tool to assess LV shape and sizes[10]. This finding might be of importance since adequate assessment of LV contractility and dimensions is crucial for clinical decision making in these patients. It is well known that LV volumes may be significantly underestimated (even up to 50%) by 2D TTE in comparison with invasive ventriculography or CMR in patients with normal and decreased LV function due to poor image quality, apical foreshortening and geometrical assumptions[5, 11, 12]. However, there are poor data about the degree of underestimation of LV volumes by 2D TTE in patients with primary MR with good echogenicity and without overt LV dysfunction; our data showed a mean difference of 28 ml regarding the LV end-diastolic volume and 20 ml regarding the LV-end systolic volume. Interestingly, there was a strong correlation between measurements of LV volumes by the two imaging methods, which may suggest that 2D TTE may still be useful to evaluate the progression of LV remodelling in a single patient. At present, guidelines use linear LV dimensions (LV end-systolic diameter) to define LV dilatation in primary MR as a criterion for surgery[4, 13]. Our study showed very good agreement between 2D TTE and CMR regarding the measurement of linear LV dimensions. However, a recent CMR study by Schiros et al.[14] showed that the LV end-systolic diameter measured in the parasternal long-axis at the mitral valve leaflet tips may underestimate the LV volume due to spherical mid to apical remodeling. Further studies are needed to determine whether assessment of LV volumes has a benefit over LV linear dimensions for risk stratification in patients with asymptomatic primary MR.

Both Teichholz formula and modified Simpson’s method by 2D TTE seemed to overestimate LVEF in comparison to CMR. Furthermore, there was no significant correlation between both imaging modalities regarding quantification of the LVEF. However, our study population consisted of subjects with a preserved LV function and, as a consequence, a narrow range of LVEF (60-75%). Furthermore, 2D TTE is known to have a higher inter-observer variability than CMR, with a mean percentage of error that can be up to 10-15%[12] which implies that a large number of patients should have been included to detect any correlation between the two imaging modalities. Of note, 33% of our patients with severe MR and preserved LV function (LVEF ≥60%) by 2D TTE Simpson’s method had a mild decreased LVEF (50-59%) by CMR and therefore already have a class I recommendation for surgery[4, 13]. These observations are in line with another small CMR study in patients with moderate to severe primary MR[15] and suggest that more accurate assessment of LVEF by CMR might be indicated in asymptomatic severe MR to determine optimal timing for surgery.

The PISA method with calculation of ERO and RVol is strongly recommended to assess MR severity[3]. However, several studies have used RVol obtained by phase contrast velocity mapping by CMR as a reference method to compare various 2D and 3D echocardiographic techniques to evaluate MR severity[6, 8, 16, 17].

To our knowledge, only one study by Buchner et al. in 35 patients with mitral regurgitation described planimetry of the ARO and found a good agreement with calculation of the ERO by the PISA method[8]. In their study, 66% of patients had less than grade III mitral regurgitation, 66% had primary MR and 43% had a LVEF <50%. Their data showed that ARO was slightly higher than ERO, presumably because ERO represents the narrowest flow stream, which is located distal from the anatomic orifice. Our study confirms good agreement between the two methods in 21 patients with moderate to severe primary MR without overt LV dysfunction. We observed a rather higher ERO than ARO for most ERO values >50 mm², although this finding has little clinical implications since most of these cases were graded as ‘severe’ by both methods. Thus, planimetry of the ARO by CMR seems to be a valuable alternative in patients with primary MR and poor echogenicity or eccentric jets which are difficult to assess with the PISA method.

In our study, RVol derived from CMR showed only moderate correlation with RVol obtained by the PISA method as well as with the Doppler volumetric method. Furthermore, RVol might have been overestimated by the PISA method in comparison to CMR, which is in line with a recent study by Hamada et al.[17]. This might be due to difficult determination of the - often dynamic - radius of the hemispheric contour of the flow convergence zone, especially in eccentric jets. By contrast, other studies found rather an underestimation of RVol by 2D and 3D echocardiographic techniques in comparison with CMR[18, 19], indicating that in any case measurements of RVol by CMR or by 2D TTE are not interchangeable.

We did not perform more advanced echocardiographic techniques like 3D or contrast enhanced echocardiography, which are known to correlate better with CMR regarding LV volumes, LVEF and RVol[12, 17‐20] than 2D TTE. However, notwithstanding the well known benefits of these techniques, 2D TTE remains the most widely used imaging method in daily practice for the assessment of patients with primary MR. Furthermore, there was no comparison with invasive measurements of LV function or MR severity by angiography, but this latter approach is not any more recommended.

In most patients 2DTTE and CMR were not performed on the same date but within 1 month, which might influence the measurement of load-dependent parameters, in particular end-diastolic dimensions. However, we did not include patients with clinical or echocardiographic signs of heart failure and there were no patients with acute onset of symptoms during the study period. Furthermore, underestimation of LV volumes on 2DTTE was also seen in patients in whom both exams were performed on the same date.