Comparison of magnetic resonance imaging and echocardiography in determination of cardiac dimensions in normal subjects.

No data exist regarding the ability of magnetic resonance imaging to assess cardiac size and performance in human beings. Therefore, measurements of cardiac dimensions by magnetic resonance imaging were compared with those obtained by two-dimensional echocardiography in 21 normal subjects. Magnetic resonance transverse cardiac sections were obtained during electrocardiographic gating using a spin echo pulse sequence. In normal subjects, magnetic resonance imaging yielded a range of values for cardiac dimensions having a similar standard deviation as that of two-dimensional echocardiography. Diastolic measurements of the aorta, left atrium, left ventricle and septum obtained by magnetic resonance imaging correlated well with those obtained by two-dimensional echocardiography (r = 0.82, 0.78, 0.81 and 0.75, respectively). The correlation coefficient of r = 0.35 observed for the posterior wall thickness was not surprising in view of the narrow range of normal values. Only a general correlation (r = 0.53) existed for the right ventricular diastolic dimension; this was probably related to the difficulty in obtaining representative measurements due to the complex geometry of this chamber. Failure of systolic dimension measurements by magnetic resonance imaging to correlate with those obtained by echocardiography is probably related to limitations of electrocardiographic gating, especially of determining the exact end-systolic frame. Although technically complex at present, magnetic resonance imaging does provide an additional noninvasive technique for measurement of cardiac size.