Localization of occult insulinoma by super-selective pancreatic venous sampling for insulin assay through percutaneous transhepatic catheterization

Preoperative localization of insulinomas by arteriography is successful only 66% of the time. With small tumors, intraoperative localization is usually unsuccessful also. Because surgicl morbidity and mortality are increased greatly when major, blind pancreatic resections or reoperations are performed, additional preoperative localization procedures are needed. We now report the successful localization of an occult insulinoma by means of blood sampling for insulin radioimmunoassay, obtained by selective pancreatic vein sampling through percutaneous transhepatic catheterization. Our patient had symptoms of fasting hypoglycemia for 11 months. Routine studies were normal and two fasting plasma glucose concentrations were 34 and 16 mg/dl to correspond with insulin values of 33 and 75 μU/ml. Celiac arteriography was normal. Percutaneous transhepatic portal and pancreatic venous catheterization revealed insulin concentrations (μU/ml) of between 23 and 36 in portal and splenic veins, 17 in the short, gastric vein, 17 in the dorsal pancreatic vein, 19 in the superior, mesenteric vein, and 61 and 288 in two pancreatic magna veins draining the tail. An insulinoma of 2 x 3 cm was resected from the tail. On the basis of our experience with this patient, it is clear that selective venous sampling may facilitate the localization of occult insulinomas at surgery and thereby avoid extensive, blind, pancreatic resections and the need for reoperation.