What is the accuracy of frozen section in the diagnosis of mucinous ovarian tumours? A 9-year review of performance in a Greek tertiary referral centre.

PURPOSE: This retrospective study aimed to evaluate the diagnostic accuracy of the intra-operative frozen sections (FS) of mucinous ovarian tumours (mOT).

METHODS: Between 2007 and 2015, a total of 105 mucinous ovarian samples were collected during laparotomy. The intra-operative FS accuracy was evaluated and potential factors correlated with increased inaccuracy assessed using both univariate and multivariate analysis.

RESULTS: The overall FS accuracy was 82.6%, while diagnostic discrepancy observed in 18/105 cases, including under-diagnosis in 14 and over-diagnosis in four cases. Amongst six cases diagnosed as benign with FS, five were upgraded to low malignant potential and one to malignant in the final formalin fixed, paraffin embedded section (FFPES). Amongst the 37 low malignant potential (LMP) cases, two were finally diagnosed as benign and eight as malignant. Amongst malignant tumours the diagnostic agreement occurred in 21/23 cases, while solely two cases were over-diagnosed. The false FS interpretation resulted in inadequate surgical management in 8/105 (7.6%) cases. Misdiagnosis had a statistically significant association with tumour size greater than 13 cm. The ratio of tumour size per number of frozen sections (TSFSR) less than 8 found to be an independent predictor of inaccuracy [OR 2.46, 95% confidence interval (CI) 1.74-3.46, P < 0.001].

CONCLUSIONS: The accuracy rate of FS in our study was 82.6%. Frozen section had low accuracy amongst LMP tumours adversely influencing the adequate surgical management. This discordance seems to reflect adverse predictors such as the LMP heterogeneity, maximal tumour diameter and low TSFSR.