e17569 Background: The use of a single and common language in the diagnosis of ovarian tumors is a multidisciplinary effort spanning several decades that has allowed the global adoption of the Ovarian-Adnexal Reporting & Data System (O-RADS) and the International Ovarian Tumor Analysis (IOTA). The objective is to determine the diagnostic accuracy of O-RADS versus modified benign descriptors and IOTA ADNEX in the classification of adnexal masses. Methods: Ambispective diagnostic accuracy study based on ultrasound data collected at a university hospital with a gynecologic oncology unit between 2012 and 2022. Two IOTA-certified level III sonographers classified adnexal masses using modified benign descriptors (MBD). The main outcome was the classification of the masses as benign or malignant for each strategy. Histopathology was used as the diagnostic reference standard. A diagnostic test analysis was performed. Results: 876 patients with 902 adnexal masses were included in the study for both analysis strategies. For the two-step analysis, 573 (64%) met the classification criteria for their evaluation based on MBD (first step) and 329 (36%) were assessed using IOTA ADNEX. 99% (567/573) of the masses classified according to the MBD, were benign by histopathology. The diagnostic performance of IOTA ADNEX (second step) had a sensibility of 93%, specificity of 64%, positive predictive value of 48%, negative predictive value of 96%, AUC (0,880). Regarding the analysis of the O-RADS system to stratify malignancy risk, we observed a prevalence of 58% for O-RADS 2, 22% for O-RADS 3, 11% for O-RADS 4, and 9% for O-RADS 5. Sensibility was 92%, specificity 88%, positive predictive value 48%, negative predictive value del 99%, AUC (0,886). Conclusions: In our experience, the diagnostic accuracy of the two-step strategy or the O-RADS system is similar. However, the ORADS system had a lower proportion of false negatives.