Since Ödman (1) first described the potential of selective angiography in the diagnosis of pancreatic diseases, many reports dealing particularly with pancreatic tumors have appeared (2–7). The vascular changes caused by pancreatic tumors have been well described, but only a few reports have given any statistics regarding preoperative accuracy (8–10). Although these reports have indicated a high incidence of false negative diagnoses, accurate data regarding false positive diagnoses and comparison of preoperative and retrospective diagnoses is lacking. Furthermore, no attempt has been made to define the frequency of findings simulating tumor in a large control series of angiograms in patients without carcinoma of the pancreas. The purpose of this paper is to report the experience with pancreatic angiography at the Massachusetts General Hospital, emphasizing preoperative accuracy and retrospective analysis, and to determine the incidence of angiographic findings mimicking tumor which lead to false positive diagnoses. These data are drawn from a series of 300 selective celiac and superior mesenteric angiograms. Technic The Seldinger technic (11) of catheter introduction usually from a femoral, but occasionally an axillary, artery was utilized, and a preformed red Kifa catheter was inserted. During the earliest part of the series sequential injection into the celiac axis and superior mesenteric artery was employed. During the second quarter of the series, simultaneous injection into the celiac and superior mesenteric arteries, using 2 catheters (one from each femoral artery), was performed. During the last half of the series sequential selective splenic and selective hepatic arterial catheterizations and occasionally selective pancreatic (dorsal pancreatic artery and pancreaticoduodenal artery) injections were made in addition to superior mesenteric angiography. Renografin-762 was used in varying volumes, depending upon the vessel catheterized. Average volumes were: selective splenic and selective superior mesenteric arteries 50–70 cc each, selective hepatic artery, 35–50 cc; selective gastroduodenal artery, 10 cc; selective pancreatic artery 3–5 cc. The object of the angiogram was not only to show the arterial and capillary phases, but also to demonstrate the portal venous phase and thus show possible compression or obstruction of the splenic or superior mesenteric vein. Therefore, serial films were always used with 2 films per second for five seconds and one film every third second for twenty additional seconds. Anteroposterior and right posterior oblique projections were employed in almost all cases where tumor was suspected in the head of the pancreas. Anteroposterior and occasionally left posterior oblique projections were used when the lesion was in the tail.