Many patients with periampullary mass lesions lack a tissue diagnosis at referral despite advances in body imaging and aggressive biopsy techniques. This review evaluates a consecutive cohort of patients who underwent pancreatoduodenectomy (PD) with and without a diagnosis of malignancy. From 1990 to 2001, 121 patients underwent PD on a gastrointestinal surgical service by a single surgeon with a bias toward "blind" Whipple resections (BWR). Sixty-three per cent of the patients had obstructive jaundice with a mass on CT in 51 per cent. Fifty-three patients (44%) had a preoperative diagnosis of malignancy. Sixty-eight patients (56%) underwent a blind PD based on computed tomography (CT), ERCP, and clinical findings. After PD, 113 patients (94%) had a malignancy (46 pancreatic, 30 ampullary, 13 cholangiocarcinoma, 9 neuroendocrine, 4 duodenal, 10 other). Of the 68 patients (56%) who underwent a blind PD, 61 patients (90%) had a malignancy. Ten per cent of the BWR patients had a pathologic diagnosis of chronic inflammation/pancreatitis. Overall mortality was 3.3% (4 patients), with no deaths in the BWR group. In this review, clinical judgment was correct in 90 per cent of patients undergoing a "blind" PD without a prior diagnosis of malignancy. In patients with "potentially resectable" lesions (based on CT exam), biopsy information does not affect the choice of therapy since a negative biopsy still commits the patients to surgery. Combined CT and/or ERCP data with clinical findings leads most often to a correct diagnosis and procedure. These data question the practice of numerous biopsy attempts in patients with periampullary lesions.