Validation of REM score to predict endometrial cancer in patients with ultrasound endometrial abnormalities: results of a new independent dataset.

The risk of endometrial malignancy (REM) score is a model formulated in a previous single-center validation study, which has been shown to predict endometrial cancer in women with ultrasound endometrial abnormalities based on multiple features (clinical, ultrasound and laboratorial). The purpose of this study was to validate the performance of REM score in an external validation setting. A population-based database with patients, who underwent elective hysteroscopy for ultrasound endometrial abnormalities between 2013 and 2016 at Department of Obstetrics and Gynecology of Campus Bio-medico of Rome, was used. Starting from January 2013 to June 2016, 330 patients were enrolled for hysteroscopy. Thirty-two patients were excluded due to Asherman syndrome or cervical stenosis. Therefore, a total of 298 patients were considered for the analysis. Based on pathologic examination, 102 patients were found to have endometrial cancer, and 196 had benign endometrial disease. Using the predefined cutoff of 0.3185, identified in the previous publication, in this independent cohort of patients we correctly classified 93/102 patients with endometrial cancer and 187/196 with benign disease, reporting an overall sensitivity and specificity of 93.9 and 95.4% (PPV = 0.91, NPV = 0.95), respectively. REM score showed a high positive predictive value for endometrial cancer prediction. However, before REM score can be applied in daily clinical practice, data from randomized controlled trials are needed.