There are six major steps in the management of patients with neuroendocrine tumors (NETs) (carcinoids and pancreatic endocrine tumors). One of the steps that is increasing in its importance is the need to assess primary tumor location and tumor extent in these patients. Without such information, it is not possible to adequately manage these patients. Conventional imaging studies (CT scan, MRI, ultrasound, angiography), functional localization studies measuring hormonal gradients, endoscopic ultrasound, and most recently, somatostatin receptor scintigraphy (SRS) with [125I-DTPA-DPhe1]-octreotide have all been advocated to localize NETs in different studies. Whereas it is now established that for all NETs, except insulinomas, SRS has the greatest sensitivity, it remains unclear whether this increased sensitivity translates into increased clinical usefulness. It, therefore, remains unclear based on fiscal and clinical considerations what should be the recommended algorithm for the use of the different localization methods. To address this issue, we have recently performed two prospective studies on patients with gastrinomas. In this paper, the methods and results of each are summarized and based on these results, an algorithm for localization studies in NETs is proposed. One study assessed the role of SRS in management in 122 patients and shows that the use of SRS changed management in 47 percent of patients according to six different criteria when the patients were stratified according to their principal management problem. Determining whether liver metastases were present is one of the major goals of tumor localization studies and is frequently a source of confusion because of the difficulty in distinguishing small NETs liver metastases from hemangiomas. In the second study, the ability of SRS and other tumor localization methods to distinguish these two possibilities was assessed in 15 patients with small hemangiomas and 15 patients with small hepatic metastases (mean size 1.3 cm). SRS correctly identified 93 percent of the patients with liver metastases and was not positive in any patient with a hemangioma, suggesting it was not a liver metastases. SRS had greater negative and positive predictive value than conventional studies. Based on these two studies, and SRS's greater sensitivity and fiscal considerations, it is proposed that SRS should be the initial tumor imaging study in all NETs except insulinomas, and algorithms for the use of other localization studies in both NETs and insulinomas are proposed.