End-diastolic volume and end-systolic volume were measured in 35 consecutive patients with cardiomyopathy using 2-dimensional (2-D) and 3-dimensional (3-D) echocardiography (2, 4, and 8 planes) and cardiac magnetic resonance imaging. Three-dimensional echocardiography correlates better with magnetic resonance imaging than does 2-D echocardiography. Its accuracy improves with the increase in the number of planes used. Two-dimensional echocardiography underestimates volumes, mainly in the subgroup with an ejection fraction of <50%, whereas 3-D echocardiography does not, if enough planes are used. However, in patients with an end-diastolic volume > or =150 ml, the underestimation of 3-D echocardiography is statistically significant. Increasing the number of planes to 8 reduces this bias. Conversely, patients with an end-diastolic volume <150 ml are accurately studied with just 4 planes.