P11.05: Subjective ultrasound assessment and the ADNEX model to differentiate between benign and malignant ovarian tumours

Abstract

To compare subjective ultrasound assessment and the ADNEX model to differentiate benign and malignant ovarian tumours. This was a retrospective study including women with ovarian tumours who underwent surgery for these lesions. All patients underwent both transabdominal and transvaginal examination to assess tumour morphology and extent of the disease. Two expert sonographers performed all the scans. The ultrasound examiner assessed tumour morphology according to the IOTA protocol. Finally, the sonographer predicted the tumour as benign or malignant. This assessment was based on subjective evaluation based on knowledge from previously published IOTA studies and was given without knowledge of the results of the ADNEX model. Accuracy of the ADNEX was calculated for the cut-offs of 3% (ADNEX-3%) and 10% (ADNEX-10%) total risk of malignancy. The McNemar's test with the Yates continuity correction was used to compare the accuracy of subjective evaluation with ADNEX-3% and ADNEX-10%, respectively to compare subjective ultrasound assessment and the ADNEX model to differentiate benign and malignant ovarian tumours. A total of 207 women were included in the study with a median (range) age of 52 (15-81) years. Subjective ultrasound evaluation had a significantly higher accuracy (91.8%) in discriminating benign and malignant ovarian tumours in comparison with ADNEX-3% (77.3%; p<0.001) and ADNEX-10% (81.6%; p<0.001). Sensitivity and sensibility of subjective ultrasound evaluation, ADNEX-3% and ADNEX-10% were respectively: 92.2% and 92.4%, 94.0% and 68.9%, 99.1% and 50.0%. This study shows that subjective ultrasound assessment performed by an expert sonographer is more accurate than ADNEX model using two different cut-off points in differentiating benign and malignant ovarian tumours. Future prospective studies with larger population of patients should confirm these preliminary findings.