To assess the impact of aortic root asymmetry on the relationship between aortic dimensions derived from two-dimensional transthoracic echocardiography (TTE) as compared with cross-sectional cardiovascular magnetic resonance (CMR) imaging in adults with a bicuspid aortic valve (BAV). Maximal CMR cross-sectional aortic measurements at the level of the sinuses of Valsalva, including cusp-commissure, cusp-cusp diameters and aortic root areas, from 68 consecutive patients (65% male) were retrospectively analyzed. The degree of aortic root asymmetry on CMR was expressed using the coefficient of variance of the root diameters in each dimension for an individual (CoeffVi) as compared with the median of the entire population (CoeffVp) and asymmetry was defined as CoeffVi > CoeffVp. Values obtained from CMR were compared with standard root measurements using TTE from contemporary studies (48 patients, 71%). Reproducibility of CMR measurements was assessed using the intra-class correlation coefficient (ICC). Echocardiography systematically underestimated aortic root dimensions in comparison with CMR, particularly in asymmetric roots with cusp-cusp measurements in systole (bias: -4.9 mm). Best agreement between modalities existed in symmetric roots with cusp-commissure measurements in diastole (bias: -0.01 mm). CMR measurements showed excellent intra-reader (ICC ≥ 0.98) and moderate inter-reader (ICC range 0.37-0.95) reproducibility, particularly aortic root area (inter/intra-reader ICC ≥ 0.94). In comparison to cross-sectional CMR diameters, standard TTE measurements consistently underestimates maximum aortic root diameter in adults with a BAV and aortic root asymmetry further decreases the agreement between CMR and TTE. CMR-derived aortic root measurements are reproducible and aortic root area showed the best reproducibility.