BACKGROUND To determine the accuracy and safety of a modified intra-arterial calcium stimulation with the venous sampling test (ASVS) for preoperative localization of insulinomas. Modification included stimulation with a fixed low dose of calcium gluconate, additional stimulation in the distal splenic artery, and no insulin sampling in the left hepatic vein. METHODS In 10 patients showing biochemical evidence of organic hyperinsulinemia, 0.45 mmol of Ca2+ was injected into the gastroduodenal, superior mesenteric, proper hepatic, proximal, and distal splenic arteries during angiography. Insulin levels were measured in samples taken from the right hepatic vein before and 30, 60, 90, 120, 180, and 300 s after Ca2+ injection. RESULTS Insulin gradients with an increase of more than fourfold indicated direct tumor supply, two- to fourfold correlated with collateral supply, and less than twofold correlated with normal tissue vascularization. ASVS localized all the adenomas of the pancreatic head (n = 3) and body (n = 2) and two of four adenomas of the tail correctly, as confirmed by surgery. Two adenomas of the proximal pancreatic tail were erroneously localized to the body segment, but the fault was rectified by angiography. In one patient with a negative ASVS and without exploration, the diagnosis of an insulinoma was revised. CONCLUSIONS ASVS with a fixed low dose of calcium gluconate is a highly accurate and safe method for preoperative localization of insulinomas. Sampling in the left hepatic vein can be routinely omitted. Additional stimulation in the distal splenic artery seems helpful in surgical decision making, but additional experience is needed.