Borderline ovarian tumors (BOTs) represent a subset of epithelial ovarian neoplasms characterized by atypical epithelial proliferation without stromal invasion. Accurate intraoperative diagnosis via frozen section is crucial for guiding surgical decisions, particularly in fertility-sparing cases. However, the reliability of frozen section diagnosis in differentiating benign, borderline, and malignant ovarian tumors remains variable.

This study aims to evaluate the correlation between frozen section and final histopathological diagnosis of BOTs in a single-center retrospective analysis.

A retrospective review was conducted on 68 patients diagnosed with BOTs on frozen section at a tertiary care hospital from 2019 to 2023. Demographic, clinical, and histopathological data were collected. Final histopathology reports were compared with frozen section diagnoses to assess diagnostic accuracy. Univariate and multivariable logistic regression models were used to identify predictors of misdiagnosis. Agreement between frozen section and final histopathology was measured using Cohen’s Kappa.

Among 68 cases, the mean patient age was 43.75 years, with 76.4% premenopausal. The most common histologic types on frozen section were mucinous BOT (44.1%) and serous BOT (45.6%). Final histopathology revealed 33.8% benign tumors, 41.2% BOTs, and 25% malignant tumors. Frozen section overdiagnosed benign tumors as BOTs in 33.8% of cases and underdiagnosed malignant tumors as BOTs in 25%. Mucinous histology (P < 0.05), tumor size > 15 cm (P = .049), and tumor confinement to the ovary (P = .007) were significant predictors of underdiagnosis. Poor agreement was observed between frozen section and final histopathology (Cohen’s Kappa = 0.015, P = 0.756). The sensitivity and positive predictive value of frozen section for BOTs were 86.2% and 40.3%, respectively.

Frozen section diagnosis of BOTs demonstrated limited accuracy, with significant rates of overdiagnosis and underdiagnosis, particularly for mucinous and larger tumors. Given its poor agreement with final histopathology, reliance on frozen section alone for intraoperative surgical decision-making should be approached with caution, emphasizing the necessity of final histopathological confirmation.