Intraoperative frozen section diagnosis in skull base surgery

During January 1990 through June 1992, we performed 39 surgeries for base of skull tumors with extracranial and intracranial involvement on 33 patients. Intraoperative frozen section was requested on 581 of 904 specimens submitted (64%). There was a discrepancy between the frozen section diagnosis and final diagnosis in 15 specimens for an error rate of 3%. There were two false-positive diagnoses of malignancy and 13 false-negative diagnoses. The discrepancies were a result of sampling error in ten cases and of interpretive error in five cases. Four of the five interpretive errors involved intradural tissues. Only two of the discrepancies were clinically significant. One involved the evaluation of adequacy of surgical margins and a second involved the misinterpretation of metastatic renal cell carcinoma for hemangioblastoma. We recommend careful attention to cryostat sectioning and interpretive experience in head and neck pathology and neuropathology. The importance of vigilant communications between surgeon and pathologist before, during, and after surgery cannot be overstated. We are utilizing a video-linked microscopic network that allows the surgeon to view the frozen section histologic sections in the operating room. Future trends may include the use of monoclonal antibodies and morphometry to improve accuracy in frozen section diagnosis.