Predicting delayed diagnosis rate of intraoperative rapid frozen section pathological examination for early-stage breast cancer: a real-world retrospective study

Background and purpose: For few of early-stage breast cancers who are undiagnosed by core needle biopsy, the delayed diagnosis rate of intraoperative rapid frozen section pathological examination is unclear. The purpose of this retrospective cohort study was to investigate the clinical application value of frozen sections in this situation. Methods: This study reviewed data of 876 breast cancer patients that were undiagnosed by core needle biopsy in Fudan University Shanghai Cancer Center from May 1, 2006 to December 31, 2019. Clinical characteristics and image data and pathological data of patients were collected. The correlation between clinical features and delayed diagnosis rate (DDR) of frozen section was explored using logistic regression analysis, then a nomogram was constructed to predict the probability of delayed diagnosis. This study was approved by Ethics Committee of Fudan University Shanghai Cancer Center (No.: 050432-4-2108*). Results: A total of 876 patients met the inclusion standards. The intraoperative diagnosis rate of frozen section for breast cancers that were undiagnosed by core needle biopsy was 67.7%, and the DDR was 32.3%. In multivariate analysis, papillary lesion [odds ratio (OR)=4.251, 95% CI: 2.804-6.492; P<0.001) and sclerosing adenosis (OR=3.727; 95% CI: 1.897-7.376; P<0.001) accompanied by atypical epithelial hyperplasia on core needle biopsy (CNB) were positive correlation factors of delayed diagnosis, while clustered microcalcifications on mammography (OR=0.345; 95% CI: 0.216-0.543; P<0.001) and ultrasonic category 4C-5 according to Breast Imaging Reporting and Data System (BI-RADS) (OR=0.250; 95% CI: 0.081-0.777; P=0.016) were positive correlation factors of intraoperative diagnosis. The nomogram constructed by these factors could better predict the delayed diagnosis rate of frozen section and screen out low delayed diagnosis population. Conclusion: Frozen section has a certain delayed diagnosis rate for breast cancer that is not clearly diagnosed by core needle biopsy. The method of model prediction can effectively eliminate the delayed diagnosis cases and avoid some unnecessary frozen sections.