MUGA processing: intra and interoperator variability impact using manual and automated methods

Abstract

Introduction – Multigated acquisition (MUGA) scan is mainly used for the assessment of left ventricular ejection fraction (LVEF) in patients who undergo cardiotoxic chemotherapy drugs. When applying automatic (A) or manual (M) processing methods, some biases in the quantitative metrics can be obtained. The aim of this study is to evaluate the influence of A and M methods, specifically, the inter and intraoperative variability in accordance with the professional experience. Methods – A retrospective study was performed with 14 MUGA exams available in ESTeSL’s Xeleris™ Functional Imaging Workstation v. 1.0628 database. Three operators (OP) with no professional experience and two with more than 10 years of experience, processed every study five times for each method, using the EF Analysis™ and the Peak Filling Rate™. To perform the multiple comparisons, the Repeated Measures ANOVA, Friedman, t-test and Wilcoxon tests were used, considering α=0.05. Results – Four of the OP presented statistically significant differences between methods in one or more parameters; similar values between experienced OP and between the non-experienced were observed when the A method was applied, and higher discrepancies were present for all parameters obtained by the M mode; higher LVEF, peak filling rate, and peak empying rate values were observed for the M method. Conclusion – Variability was found when comparing M and A processing methods, as well as interoperator variability associated with their level of experience. Despite that, there was a trend of less variability between the two experienced OP and in the A method.