Magnetic resonance imaging (MRI) has been used to measure left ventricular (LV) mass in animals with superior accuracy. However, its use in cardiac patients has been limited by the long total scan times necessitated by imaging the heart at end-diastole at each of 8 to 10 slice locations. Recent canine studies showed that LV mass may be determined accurately, with considerable timesavings, by use of sequential images throughout the cardiac cycle (single-phase MRI). Twenty normal subjects underwent spin-echo MRI to determine the relationship between LV mass computed from single-phase MRI and results obtained from the more time-consuming end-diastolic MRI (which was used as the reference standard for this study). The left ventricle was spanned with 2 interleaved series of 5 short-axis 1 cm thick slices. 5 images, evenly spaced throughout the cardiac cycle, were obtained at each slice location in all subjects. LV mass ranged from 86 to 198 g. Although end-diastolic LV mass exceeded single-phase results by an average of 5 g (p less than 0.002), there was a close correlation between the 2 (slope = 0.99; r = 0.96). Although LV mass derived from end-diastolic images exceeded single-phase results, this difference is unlikely to be clinically significant and is small compared with the standard error of echocardiographic methods. Furthermore, when the order in which single-phase images were selected was reversed, there was improved agreement with end-diastolic MRI.(ABSTRACT TRUNCATED AT 250 WORDS)