BACKGROUND: Patients with potentially resectable pancreatic masses, without evidence of metastatic disease, require surgical exploration. We assessed the reliability of frozen section diagnosis of pancreatic malignancy. METHODS: We analysed data from 120 patients who underwent pancreatic exploration, including frozen section assessment, over a period of 41 months. RESULTS: A total of 310 pancreatic biopsies were sent for frozen section analysis. The frozen section result concurred with the final histological diagnosis in 98.1% of biopsies. The false-negative rate was 1.9%, with the correct diagnosis being established on subsequent frozen section in half of these cases. Only 3 patients were not correctly diagnosed intra-operatively by frozen section. The majority of tumours identified were pancreatic ductal adenocarcinomas, but endocrine tumours and lymphoma were also detected. 15 (12.5%) patients required more than one set of biopsies in order to establish a diagnosis of malignancy. 19 (15.8%) patients had malignant disease outside the pancreas, making their tumours unsuitable for curative resection. CONCLUSIONS: Frozen section diagnosis is very reliable in the assessment of pancreatic masses and accurately differentiates between benign and malignant disease. More than one biopsy is needed to make a diagnosis, and if clinically suspicious, successive sets of biopsies may be required.
BACKGROUND:Patients with potentially resectable pancreatic masses, without evidence of metastatic disease, require surgical exploration. We assessed the reliability of frozen section diagnosis of pancreatic malignancy. METHODS: We analysed data from 120 patients who underwent pancreatic exploration, including frozen section assessment, over a period of 41 months. RESULTS: A total of 310 pancreatic biopsies were sent for frozen section analysis. The frozen section result concurred with the final histological diagnosis in 98.1% of biopsies. The false-negative rate was 1.9%, with the correct diagnosis being established on subsequent frozen section in half of these cases. Only 3 patients were not correctly diagnosed intra-operatively by frozen section. The majority of tumours identified were pancreatic ductal adenocarcinomas, but endocrine tumours and lymphoma were also detected. 15 (12.5%) patients required more than one set of biopsies in order to establish a diagnosis of malignancy. 19 (15.8%) patients had malignant disease outside the pancreas, making their tumours unsuitable for curative resection. CONCLUSIONS: Frozen section diagnosis is very reliable in the assessment of pancreatic masses and accurately differentiates between benign and malignant disease. More than one biopsy is needed to make a diagnosis, and if clinically suspicious, successive sets of biopsies may be required.