Predictive factors of malignancy in patients with adnexal masses.

INTRODUCTION: Good preoperative tumor triage is essential for choosing the appropriate approach. OBJECTIVE: The study aim was to identify factors from standard preoperatively collected data, which could predict the nature of adnexal masses prior surgery. MATERIAL AND METHODS: The study involved all women treated in the Clinic for Gynecology and Obstetrics Clinical Center of Serbia for adnexal tumors throughout a period of 18 months. On admission, detailed anamnestical and laboratory data were obtained and ultrasound scans were performed. Obtained data were compared with hystopathological findings of tumors. Methods of correlation and logistic regression were applied to create association models. RESULTS: Three new models for predicting tumor nature were achieved from anamnestical data, characteristics of women and tumors, and laboratory analyses. Two statistically significant (p = 0.000) equations were obtained for anamnestical data and characteristics of women and tumors, while three were made for laboratory analyses. Sensitivity of anamnestical malignancy index (AMI) was 73.33%, specificity 72.87%, positive predictive value (PPV) 39.49% and negative predictive value (NPV) 91.88%. Sensitivity of characteristic malignancy index (CMI) was 92.38%, specificity 67.36%, PPV 40.59% and NPV 97.34%. Sensitivity of laboratory malignancy index (LMI) was 56.45%, specificity 90.24%, PPV 68.63%, and NPV 84.57%. CONCLUSIONS: The best predictors of malignancy are menopausal status, body mass index (BMI), age, metastases, ascites, tumor marker CEA level, and erythrocyte sedimentation rate (ESR). Along with the risk of malignancy index (RMI), for more reliable triage and preoperative tumor evaluation the authors propose introduction of another three indexes (AMI, CMI, LMI) in clinical practice.