Objective: To assess the sensitivity and positive predictive value (PPV) of intraoperative frozen section diagnosis of borderline tumors of the ovary (BTO).
Methods: Retrospective analysis at the Department of Obstetrics and Gynecology, University of Vienna, between 1995 and 2007 and review of the literature. Frozen section analysis and definitive histology reports were compared. Univariate and multivariate regression models were used to assess the influence of patient and tumor characteristics on the likelihood of underdiagnosis and overdiagnosis.
Results: Agreement between frozen section diagnosis and definitive histology was observed in 69/96 (71.9%) patients, yielding an overall sensitivity and a positive predictive value of 75.0% and 94.5%, respectively. Underdiagnosis and overdiagnosis occurred in 27/96 (28%) and 0/96 (0%) patients, respectively. In a univariate and multivariate analysis, tumor diameter, but not patient age, tumor histology, tumor stage, presence of a bilateral tumor, serum CA-125 and concurrent presence of endometriosis was a predictor of underdiagnosis of frozen section analysis. We identified 29 studies investigating the accuracy of frozen section analysis of BTO. Three studies exclusively examined BTO in 140, 48 and 33 cases, respectively. Data of these three studies and the present study were pooled, yielding an overall sensitivity and PPV of 71.1% and 84.3%, respectively. Overdiagnosis and underdiagnosis were identified in 21/317 (6.6%) and in 97/317 (30.6%) cases, respectively.
Conclusion: Intraoperative frozen section diagnosis of BTO has a low sensitivity and PPV and overdiagnosis and underdiagnosis are frequent. Surgical management based on intraoperative frozen section diagnosis should be used with caution.