The authors evaluated the reproducibility of measurements of ventricular dimensions obtained with cine magnetic resonance (MR) imaging performed on two occasions in 11 healthy subjects. Two reviewers analyzed the studies in a blinded fashion to determine interobserver and interstudy variability of measurements of left ventricular (LV) mass, volume, ejection fraction, and systolic wall stress. LV mass showed good reproducibility between studies, with 3.6% and 3.8% variability for LV end-systolic mass for the two observers. LV end-diastolic volume varied by 5.2% and 3.9%, and LV end-systolic volume, by 9.7% and 0.9%. Variability for LV ejection fraction was 5.0% and 4.9%. The largest interstudy variability was end-systolic wall stress, 11.1% and 13.5%, which was due mostly to changes in arterial pressure between the two studies. It is concluded that anatomic and functional measurements from cine MR images are reproducible between studies. The small interstudy variability is likely related to the fact the measurements are derived directly from cine MR images that encompass the entire heart rather than depend on measurements of only sample images and the use of geometric models.