The diagnostic value of frozen section for borderline ovarian tumours

Abstract

Borderline ovarian tumours (BOTs) are characterised histologically by a low degree of cellular proliferation and nuclear atypia in the absence of infiltrative growth or stromal invasion. Surgical treatment has been a crucial component of BOT therapy. Surgical decisions are established intraoperatively via the frozen section. We evaluated the accuracy of frozen section diagnosis. The rate of correct diagnosis, underdiagnosis and overdiagnosis of BOTs with frozen sections was 78%, 17% and 5%, respectively. The sensitivity and positive predictive values for the diagnosis of BOTs with frozen sections were 82.3% and 93.3%, respectively. The positive likelihood ratio was 0.82 (95% CI: 0.85-0.96). The histological classification of BOTs had a significant effect on the accuracy of diagnosis (p=0.001). Frozen section diagnosis is not suitable to be considered as the gold standard for a definitive diagnosis. Clinicians should be aware that using frozen sections is insufficient for the accurate staging of BOTs.