Reproducibility of ventricular function measurements by contrast angiography

Interobserver reproducibility of single plane contrast angiographic measurements of end-diastolic volume (EDV/m2), end-systolic volume (ESV/m2), stroke volume (SV/m2), and ejection fraction (EF) was studied in two groups of patients. The first group (n = 42) was an unselected, consecutive series of patients in which the calculations of volume were performed manually, representing day-to-day results from a clinical quantitative angiographic laboratory. The second group (n = 31) was selected on the basis of optimal filming technique and volume calculations were performed digitally with the aid of a computer. This group characterizes the reproducibility that could be attained for investigational purposes. In the selected group, interobserver reproducibility of four regional ejection fractions was determined. The correlation coefficient (r), standard error of the estimate (Sy.x), average difference, and 95% confidence limits are given. Measurement variation accounted for 9-15% of the variance in this study for EDV/m2 (consecutive series), indicating that caution is needed when relating contrast angiographic measurements to other (e.g., noninvasive) measures of volume by simple linear regression.