Drs. Ciaccia and Gress raise several points; we agree with some and disagree with others. First, our initial studies with SRS were done to determine the overall role of various imaging studies in the management of patients with gastrinomas. Endoscopic ultrasonography is primarily used in patients before surgery and has either a minimal role or no role in the 40% to 50% of patients who do not undergo surgery or have liver metastases [1]. Conventional studies and SRS have a potential role in assessing all aspects of disease (preoperative assessment, extent of metastases, response to chemotherapy). Therefore, we first did this study comparing the sensitivity of methods used in all aspects of the disease [2] and then did a study [3] to determine the ability of SRS to alter patient management. These studies show that SRS is the most sensitive method and that it frequently alters management. Therefore, in the overall management of patients with gastrinomas, SRS should be the initial imaging study.

Second, we disagree that EUS is established as the foremost method for localization of extrahepatic gastrinomas. In one study [4], it localized only 50% of duodenal gastrinomas. In another study [5], EUS localized gastrinomas in the duodenopancreatic area in 40% of patients, SRS localized them in 56%, and both methods together localized them in 69%. This study suggests that these two procedures may be complementary, but, at present, too few patients have had both studies and surgery to conclude this with confidence. We therefore agree that additional studies are needed on this point, and, with Drs. Lok Tio and Stanley Benjamin (Georgetown University Medical School, Washington, D.C.), we have begun a prospective study.

Third, we question the statement that EUS is the foremost diagnostic method for the localization of other functional islet cell tumors. Except in the case of insulinomas, SRS and EUS have been compared in too few patients with islet cell tumors to determine their relative value. Furthermore, liver metastases in these patients are usually found at presentation; thus, EUS may not be generally needed. We agree that in patients with insulinomas-which are benign in 90% of cases, are intrapancreatic, and have a lower frequency of positive results on SRS-EUS may be particularly useful. However, these characteristics apply only to insulinomas.

Fourth, because functional islet cell tumors are diagnosed by laboratory studies, it is currently unclear what the role of cytologic confirmation by EUS will be. Most patients without metastatic disease will have surgery regardless of the cytologic results.

It is now possible to perform studies that address each of the above issues. The resulting information will allow physicians to decide which of these relatively expensive localization methods is appropriate in which circumstances.