Introduction/Background Currently, the diagnosis of borderline ovarian tumors (BOT) by ultrasound is a challenge. Our study proposes how to differentiate between true positives (TP) and false positives (FP) in the diagnosis of BOT

Methodology A retrospective study of patients who underwent transvaginal ultrasound with a diagnosis of adnexal mass and a subsequent surgery between 2012 and 2022 at a tertiary referral university hospital with a gynaecological oncology unit. The IOTA ADNEX model was used to estimate the probability of benignancy or malignancy, with a cut-off of 10% for malignancy. Pathology was the reference standard used. Descriptive statistics and bivariate analysis were performed.

Results For a total of 757 patients, the IOTA ADNEX model identified 171 (22.4%) masses as malignant, of which 64 (37.4%) were classified as BOT. 18 (28%) masses were confirmed as TP by histopathology and 46 (72%) as FP. Its usual ultrasound characteristics were similar with the particularity of having a greater number of papillae and blood vessels in papillary projections in the TP. Therefore, additional sonographic features, such as the presence of a low level of echogenicity content and a pattern of microcystic tissue resembling a cluster of tiny bubbles, were evaluated. These characteristics were relevant in most of the TP. Regarding the FP, histopathology showed that 40% were serous cystadenofibrome, 20% ovarian tube abscess, 10% Low-grade serous carcinomas, among others.

Conclusion Our findings show how some usual and novel features in ultrasound are necessary for an adequate differentiation of BOT.