The risk and accuracy of pancreatic biopsy.

Pancreatic biopsy in 171 patients produced complications in 4.7 per cent and death in 1.7 per cent. The biopsy diagnosis was confirmed in 86 per cent but was falsely negative in 14 per cent of pancreatic cancers. Frozen sections were interpreted correctly in all cases. Wedge biopsy resulted in fewer complications and fewer false-negative results than needle biopsy and is preferable except for periampullary lesions in which transduodenal needle biopsy is best. When the biopsy incision transected a pancreatic duct, suture ligation and drainage avoided complications. Open transduodenal biopsy of periampullary lesions was reliable, but the duodenotomy closure leaked in 10 per cent of the cases. Pancreatic biopsy with confirmation of malignancy should precede pancreaticoduodenectomy in most cases of suspected pancreatic cancer.