ObjectiveTo establish a prognostic model for endometrial cancer (EC) that individualizes a risk and management plan per patient and disease characteristics. MethodsA multicenter retrospective study conducted in nine European gynecologic cancer centers. Women with confirmed EC between January 2008 to December 2015 were included. Demographics, disease characteristics, management, and follow-up information were collected. Cancer-specific survival (CSS) and disease-free survival (DFS) at 3 and 5 years comprise the primary outcomes of the study. Machine learning algorithms were applied to patient and disease characteristics. Model I: pretreatment model. Calculated probability was added to management variables (model II: treatment model), and the second calculated probability was added to perioperative and postoperative variables (model III). ResultsOf 1150 women, 1144 were eligible for 3-year survival analysis and 860 for 5-year survival analysis. Model I, II, and III accuracies of prediction of 5-year CSS were 84.88%/85.47% (in train and test sets), 85.47%/84.88%, and 87.35%/86.05%, respectively. Model I predicted 3-year CSS at an accuracy of 91.34%/87.02%. Accuracies of models I, II, and III in predicting 5-year DFS were 74.63%/76.72%, 77.03%/76.72%, and 80.61%/77.78%, respectively. ConclusionThe Endometrial Cancer Individualized Scoring System (ECISS) is a novel machine learning tool assessing patient-specific survival probability with high accuracy.

Background: The O-RADS system is a new proposal for establishing the risk of malignancy of adnexal masses using ultrasound. The objective of this study is to assess the agreement and diagnostic performance of O-RADS when using the IOTA lexicon or ADNEX model for assigning the O-RADS risk group. Methods: Retrospective analysis of prospectively collected data. All women diagnosed as having an adnexal mass underwent transvaginal/transabdominal ultrasound. Adnexal masses were classified according to the O-RADS classification, using the criterion of the IOTA lexicon and according to the risk of malignancy determined by the ADNEX model. The agreement between both methods for assigning the O-RADS group was estimated using weighted Kappa and the percentage of agreement. The sensitivity and specificity of both approaches were calculated. Results: 454 adnexal masses in 412 women were evaluated during the study period. There were 64 malignant masses. The agreement between the two approaches was moderate (Kappa: 0.47), and the percentage of agreement was 46%. Most disagreements occurred for the groups O-RADS 2 and 3 and for groups O-RADS 3 and 4. The sensitivity and specificity for O-RADS using the IOTA lexicon and O-RADS using the ADNEX model were 92.2% and 86.1%, and 85.9% and 87.4%, respectively. Conclusion: The diagnostic performance of O-RADS classification using the IOTA lexicon as opposed to the IOTA ADNEX model is similar. However, O-RADS group assignment varies significantly, depending on the use of the IOTA lexicon or the risk estimation using the ADNEX model. This fact might be clinically relevant and deserves further research.