The results of intraoperative consultations in 181 ductal carcinomas in situ of the breast

BACKGROUND The utility of frozen section (FS) examination in the intraoperative management of breast lesions is well established. The accuracy of FS in the diagnosis of borderline noninvasive or preinvasive breast lesions is uncertain. METHODS The authors retrospectively reviewed the results of intraoperative consultations/frozen section examinations of 181 ductal carcinomas in situ (DCIS) of the breast. Various clinical and pathologic factors were analyzed and correlated with FS diagnosis. RESULTS FS examination was performed on 153 cases (85%) and only macroscopic examination on 28 cases (15%). FS diagnoses were as follows: DCIS in 76 cases (50%), atypical ductal hyperplasia/suspicious for DCIS in 8 cases (5%), benign in 55 cases (36%), deferred in 13 cases (8%), and invasive carcinoma in 1 case. FS accuracy, false-negative rate, and false-positive rate were 55%, 36%, and 0.6%, respectively. Sampling error was the main reason for the low detection rate, and technical inadequacy was a major factor contributing to interpretive problems. In multivariate regression analysis, FS accuracy was significantly associated with the clinical presentation of a palpable mass (odds ratio [OR] = 4.16, 95% confidence interval [CI]: 2.04-8.45), the macroscopic finding of a mass (OR = 3.03, 95% CI: 1.45-6.67), and necrosis (OR = 3.13, 95% CI: 1.4-6.67). CONCLUSIONS The authors concluded that the accuracy of FS diagnosis of DCIS was low, mainly due to sampling error. In general, FS examination should not be performed when no lesion/mass is identified by macroscopic examination. Cancer 1997; 80:75-9. © 1997 American Cancer Society