Diagnostic accuracy of intraoperative frozen section in endometrial cancer: Correlation with final histopathology
Accurate intraoperative assessment of tumor characteristics for endometrial cancer, including histological type, grade, and depth of myometrial invasion (MI), is essential for determining the extent of surgery, particularly lymphadenectomy. This study aims to evaluate the concordance between intra-operative frozen section analysis (IFS) and final histopathology (FH) in endometrial cancer cases. This retrospective analysis included 100 patients who underwent laparoscopic staging surgery for endometrial carcinoma between March 2018 and September 2024. Data on histological type, tumor grade, MI, lymph node involvement, and cervical/adnexal metastases were extracted from medical records. The diagnostic accuracy of IFS was assessed by comparing findings with FH. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and Cohen's kappa (κ) statistics were used to determine agreement levels. IFS demonstrated high concordance with FH for malignancy detection (97%, κ = 0.56). Sensitivity, specificity, PPV, and NPV were 96.9%, 100%, 100%, and 40%, respectively. Tumor grading agreement was 78.2% (κ = 0.67), with the highest accuracy in Grade 3 tumors (sensitivity 85.0%, specificity 98.3%). MI assessment showed strong agreement (κ = 0.851) with 93.7% overall accuracy. Lymph node evaluation by IFS exhibited excellent agreement (κ = 0.942), with 98.3% accuracy. IFS is a reliable tool for intraoperative decision-making in endometrial cancer, particularly for malignancy detection, MI assessment, and lymph node evaluation. However, moderate concordance in tumor grading suggests caution in surgical decision-making based solely on IFS results. Future research should focus on optimizing frozen section protocols to improve diagnostic accuracy and streamline intraoperative management.