Objective The aim of this study was to perform a meta-analysis comparing diagnostic performance of intraoperative gross evaluation (IGE) and intraoperative frozen section (IFS) for the assessment of myometrial invasion in patients with endometrial cancer.

Methods An extensive search was performed in several databases from January 1989 to May 2015. Eligibility criteria were studies using intraoperative gross evaluation or intraoperative frozen section to determine deep myometrial invasion in patients with endometrial cancer using the final histopathology report with reference standard. Quality was assessed using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines and the Quality Assessment of Diagnostic Accuracy Studies 2 tool.

Results Forty-seven articles were identified. Of these, 35 studies were selected and included in the meta-analysis. A total of 6387 women were evaluated intraoperatively with any of the 2 methods mentioned. Pooled sensitivity and specificity for IGE were 71% (95% confidence interval [CI], 63%–77%) and 91% (95% CI, 89%–93%), respectively. Heterogeneity was found high for sensitivity (I2: 83.6%; Cochran Q: 79.4; P < 0.001) and moderate for specificity (I ², 51.4%; Cochran Q, 29.8; P =0.01). Pooled sensitivity and specificity for IFS were 85% (95% CI, 81%–88%) and 97% (95% CI, 96%–98%), respectively. Heterogeneity was found moderate for sensitivity (I ², 56.4%; Cochran Q, 45.9; P < 0.001) and high for specificity (I ², 83.2%; Cochran Q, 118.9; P < 0.001). Both sensitivity (P = 0.0008) and specificity (P = 0.0021) were significantly higher for IFS compared to IGE.

Conclusion Intraoperative frozen section has better diagnostic performance than intraoperative gross evaluation for the intraoperative diagnosis of deep myometrial invasion in patients with endometrial cancer.