Objectives: In young patients with borderline tumors the fertility-sparing treatment is indicated, thus the preoperative investigation is important. The aim of this study was to perform a comparative assessment of sensitivity and specificity of selected ultrasonographic and clinical parameters for the diagnoses of borderline tumors and ovarian cancers., Methods: We retrospectively analyzed 57 patients who underwent surgical treatment in the Maria Sklodowska-Curie Memorial Cancer Center from Jan 01, 2008 to Dec 31, 2009. Ovarian cancers were diagnosed in 41 patients, and borderline ovarian tumors in 16 patients. Statistical model was developed to determine independent predictive factors that would be useful in preoperative differentiation between both tumors. The model included the following factors: menopausal status, tumor morphology, wall thickness (including outgrowths), septal thickness, echogenicity, resistive index, serum CA-125 level, and free fluid in the peritoneal cavity., Results: Based on the statistical model developed, independent predictive factors in the differentiation between ovarian cancers and borderline tumors included the menopausal status (P=0.005), tumor echogenicity (P=0.047) and the presence of free fluid in the Douglas pouch (P=0.043). With the cutoff value of 13 (with scores below 13 indicating a borderline ovarian tumor, and scores of ≥13 indicating ovarian cancer), sensitivity was 90.2% and specificity was 87%., Conclusions: Our proposed model of preoperative evaluation has a sensitivity of 90% in the differentiation between ovarian cancers and borderline tumors. When combined with intraoperative findings, it allows optimal surgical therapeutic decisions to be made in patients with borderline ovarian tumors.