Head-to-head comparison of cardiac magnetic resonance imaging and transthoracic echocardiography in the general population (MATCH)

Background Comparing transthoracic echocardiography (TTE) and cardiac magnetic resonance imaging (CMR) is crucial for cardiac assessment. This study aims to clarify their comparability in a large population sample. Methods CMR and two- (2D) and three-dimensional (3D) TTE were used to quantify left and right heart dimensions in participants of the Hamburg City Health Study. Intertechnique agreement was evaluated using Bland-Altman analyses, Pearson correlation coefficients, and Cohen's Kappa. Results Data from 2126 participants (median age 67 [IQR: 59-72] years, 897 (42.2%) female) were analyzed. Left ventricular (LV) diastolic volumes were similar (CMR: 117.0 [96.2, 138.0] ml, 2D-TTE: 111.8 [93.6, 134.3] ml, r = 0.7, p < 0.001), while systolic volumes were lower with CMR (CMR: 36.0 [26.9, 46.0] ml, 2D-TTE: 46.7 [37.9, 57.5] ml, r = 0.67, p < 0.001). CMR LV ejection fraction (LVEF) was 10% higher than 2D-TTE (CMR = 69.0 [64.0, 74.0]%, 2D-TTE = 58.3 [55.5, 61.7]%, p < 0.001; r = 0.40, p < 0.001). Left atrial volumes correlated moderately with low bias (CMR: 53.0 [40.0, 68.0] ml, 2D-TTE: 51.6 [41.5, 64.0] ml, r = 0.63, p < 0.001). LV mass showed good correlation but was higher using 2D-TTE (r = 0.74, p < 0.001). Right ventricular (RV) volumes showed the largest differences, with CMR demonstrating lower interobserver variability (ICC = 0.97 vs. 0.61 for 2D-TTE) and markedly larger volumes (RVEDV mean bias = 74.8 ml, r = 0.50, p < 0.001). Conclusion In a large general population, CMR quantifies cardiac function and dimensions more reliably than TTE. Both modalities provide significantly different absolute values, limiting intertechnique transferability.