Comparison of different geometric assumptions of the mitral annulus and their impact on echocardiographic assessment

Abstract Introduction Accurate measurement of the mitral valve (MV) annulus is important for echocardiographic assessment of left ventricular stroke volume and, ultimately, essential for evaluating valvular disease. Current echocardiographic guidelines recommend estimating cross-sectional area (CSA) using the apical 4-chamber (A4C) view [1]. However, these estimations rely on geometric assumptions that can introduce error [2,3]. Additionally, inherent interobserver variability in echocardiography poses further challenges to measurement reliability. Purpose This study aims to investigate the impact of MV geometric assumptions on CSA estimation and assess the robustness of these measurements in the context of interobserver variability, using annotations from a large cohort of experienced sonographers. Methods MV diameters were estimated from manual MV annulus delineations at end-diastole (ED) in A4C and apical 2-chamber (A2C) sequences on a cohort of 10 patients (median age: 72 y, IQR: 52-77; 70% male) by 35 experienced sonographers (mean expertise 5.3 ± 2.9 y). Ground truth delineations and ED frames were defined by a clinical committee of experts. CSA was calculated based on a circular assumption using A4C (CSA_4C) and A2C (CSA_2C) views, as well as an elliptical assumption integrating both of them (CSA_4C2C) (Figure 1). Agreement and consistency across CSA estimates were evaluated via Pearson’s correlation, linear regression and Bland-Altman analysis. Interobserver variability was quantified using the coefficient of variation. Results MV diameters in A2C were significantly larger than in A4C (mean difference 0.27 ± 0.69 cm; p = 0.03), in contrast with prior literature [2,4]. The poor correlation between these measures (R = 0.33) further underscores their inconsistency. Similarly, CSA_2C and CSA_4C were weakly correlated (R = 0.35) and poorly agreed (mean difference: -1.2 cm2; 95% LoA: -7.4 to 5.0 cm2), while CSA_4C2C provided intermediate values, as expected. These results revealed substantial discrepancies between geometric assumptions and confirmed the presence of a systematic bias when relying on single-view area estimations. Regarding interobserver variability in CSA estimations, CSA_4C showed the highest variability (38.0 ± 9.8%), while CSA_4C2C notably evidenced a reduced variability (29.5 ± 6.8%) (Figure 2). This reduction highlighted the added robustness of combining views, rather than relying on a single-plane measurement. Conclusion This study demonstrates that: (1) CSA estimates of the MV annulus are substantially influenced by geometric assumptions; and (2) the elliptical approach, incorporating both A4C and A2C views, enhances measurement consistency and substantially reduces interobserver variability while maintaining strong correlation with the traditional A4C assessment, preferred in clinical guidelines. Altogether supports the adoption of the elliptical approach for MV annular area quantification in echocardiography.Figure 1 Figure 2