Reproducibility study of Ovarian-Adnexal Reporting and Data System in describing adnexal masses / 中华超声影像学杂志

ABSTRACT

Objective:To investigate the intra- and inter-observer agreements of different experiencers using the Ovaria-adnexal Reporting and Data System (O-RADS) in the evaluation of adnexal masses.Methods:Totally 48 patients with adnexal masses (48 masses, mean size 9.5±4.7 cm, range 2.3-18.6 cm) found by ultrasound examination in the Third Affiliated Hospital of Sun Yat-sen University, from May 2019 to March 2020 were retrospectively analyzed. All the masses were confirmed by pathology or surgery. Four observers were divided into 2 senior doctors (Doctor 1 and Doctor 2) and 2 junior doctors (Doctor 3 and Doctor 4). Each observer independently evaluated adnexal masses twice using ultrasound O-RADS before and after systematic training, with an interval of 60 days. The intra-observer and inter-observer agreements were analyzed before and after training.Results:The inter-observer agreement between senior doctors were both excellent before and after systematic training (weight Kappa 0.833 vs 0.802, percentage of agreement 83.3% vs 81.3%). Whereas there was difference in the inter-observer agreement between non-experienced observers before and after training (weight Kappa 0.399 vs 0.824, percentage of agreement 50.0% vs 77.1%). After training, inter-observer agreement between junior doctors was significantly improved and comparable to senior (weight Kappa 0.824 vs 0.802, percentage of agreement 77.1% vs 81.3%). Before and after systematic training, the intra-observer agreements of the same doctor, the senior physicians were better than the junior (weight Kappa 0.882 and 0.843 vs 0.440 and 0.605; percentage of agreement 87.5% and 83.3% vs 58.3% and 54.2%).Conclusions:O-RADS risk classification system is a highly reproducible method in the subjective assessment of an adnexal mass among observers with varying levels of expertise. However, systematic training before clinical application is necessary and effective for non-experienced observers.