Purpose: To evaluate method- and user effects as well as the required time in the determination of left- and right ventricular ejection fractions in comparing manual and semiautomatic border detection of the endocardium.
Methods: In 22 patients with varying cardiac diseases endocardial borders were detected either manually or semi-automatically (seeded ROI) and subsequently the right and left ventricular volume ejection fractions were computer-calculated. Two investigators defined the endocardial borders for both ventricles using both methods. For statistical analysis two-factorial analyses of variance were performed for the four variables: VEF right ventricle, VEF left ventricle, required time right ventricle, required time left ventricle (SAS, Anova).
Results: There was no user or method effect detectable for both ventricles. The required time for both ventricles also did not reveal a user effect, but there was a dramatic method effect concerning a reduction of the required time for analysis for both ventricles > 50%.
Conclusions: The semiautomatic mode for detecting the endocardial borders in determining right and left ventricular ejection fractions did not demonstrate results differing from the results of the manual method, but is significantly superior concerning the time required for analysing ejection fractions of both ventricles with a reduction of over 50%. This might significantly reduce the expenditure for personnel.