Borderline tumors of the ovary: correlation of frozen and permanent histopathologic diagnosis

OBJECTIVE: To evaluate the correlation between the diagnosis of borderline tumor of the ovary by frozen and permanent pathology. METHODS: All pathology reports with diagnoses of borderline tumor of the ovary between 1980 and 1998 at Massachusetts General Hospital were reviewed. Univariate and multivariable logistic regression models were constructed for patient age, tumor size, histology, presence of bilateral or extraovarian disease, and concurrent diagnosis of endometriosis or endosalpingiosis. RESULTS: We reviewed 140 cases. The average age of patients was 52.3 years. Eighty tumors were serous, 47 mucinous, 11 mixed, and two endometrioid. The mean diameter overall was 13.7 cm (range 1-70 cm), 10.2 cm for serous, and 20.1 cm for mucinous. Diagnoses of borderline tumors by frozen and permanent pathology were consistent in 60% of cases. Frozen section interpreted a benign lesion as malignant (overdiagnosed) in 10.7% of cases, and interpreted a malignant lesion as benign (underdiagnosed) in 29.3%. No variable was a significant predicator of overdiagnosis. In univariate analysis, underdiagnosis was more likely for other types of tumors than serous (P <.001), tumors larger than 20 cm (P =.039), and tumors confined to the ovaries (P =. 009). When all variables were included in a multiple regression model, only histology was a significant predictor of underdiagnosis (P =.039). CONCLUSION: Frozen or permanent pathology reports of diagnoses of borderline tumor were consistent 60% of the time, whereas the positive predictive value of borderline by frozen section was 89.3%. Tumors other than serous are more likely to be misinterpreted.