Observer variability of radionuclide left ventricular volume determination at rest and during exercise.

The reproducibility expressed as the intra- and interobserver variation in the determination of cardiac left ventricular (LV) volumes by the radionuclide multigated equilibrium technique in the upright position is presented. No systematic difference was found in the reproducibility between LV volumes determined in healthy subjects and cardiac patients or between examinations performed at rest and during exercise. The intra- and interobserver variation were of the same magnitude. SD of the difference was 8 to 9 ml for LV end-diastolic volume, 4 to 7 ml for LV end-systolic volume, and 2 to 5% for LV ejection fraction. Thus, there is a 95% probability that repeat measurements, either by the same observer or by 2 independent observers, will result in the same LV end-diastolic volume within 18 ml, LV end-systolic volume within 11 ml, and LV ejection fraction within 8%. Only 15% of the variation can be ascribed to determination of the attenuation correction factor.