A retrospective quality assurance study of intraoperative consultation (frozen section) was carried out to assess the accuracy and to determine the reasons of discordance. Of 14,977 surgical pathology cases accessioned over a 6-month period in Asan Medical Center, frozen sections were done on 1,270 (8.5%) patients and 2,392 frozen sections (1.88 frozen sections/case) were performed. Discordance was noted in 106 cases (4.4%) and diagnosis was deferred in 26 cases (1.1%).
All deferred cases were reviewed with the result of 53.8% justified and 46.2% unjustified. The discordant cases were divided into three categories as to their clinical significances: category A (no affect on patient care) 61.3%, B (minimal affect) 9.4%, and C (major affect) 29.2%. Of 31 category C cases, 7 cases were false positive and 24 cases were false negative. Misinterpretation (70.8%) was the leading cause of discordance, followed by sampling error (15.1%), failure to identify lesion (8.5%), and technical problem (5.7%). More than one-third (35.8%) of all discordances were of central nervous system cases. Total central nervous system cases were 403 (16.8%) with a significantly higher disordance rate (9.8%) and deferral rate (2.5%) in comparison to the other cases with 3.4% discordance rate and 0.8% deferral rate. There were 43 colorectal cancer cases of intraoperative consultation for adequacy of resectional margins. The surgical margins were between 0.4 cm and 28 cm (mean: 6.7 cm) away from the tumor and there was no tumor-positive case. The study indicates surgical pathology should 1) promote interpretative skills in cases involving minute fragments of neurosurgical cases, 2) defer the diagnosis and ask for more tissue on inadequate or inappropriate specimens and 3) give only gross opinions without unnecessary frozen section procedures in the event of simple, clear-cut cases.