Gastrinoma: the predictive value of preoperative localization.

In the management of gastrinoma, a variety of preoperative tests for tumor localization are recommended. Their value in predicting surgical outcome is not well defined. We reviewed the preoperative investigation and operative outcome in 23 patients to assess the predictive value of localization before surgery. With results of laparotomy used as reference, the efficacy of computed tomography, angiography, ultrasonography, and magnetic resonance imaging, individually and combined for primary and metastatic disease, was determined. All investigative modalities were poor for the detection of primary gastrinomas, with individual sensitivities in the range of 20% to 29%. Their combined sensitivity was 53%, with a negative predictive value of only 38%. The detection of metastatic disease was better, with a combined accuracy of 83%. At laparotomy, tumor was identified in 19 of 23 (83%) patients. In 14 patients with negative preoperative workup, primary gastrinoma was found in 10 of 14 (71%) and resected for long-term cure in three of 14 (21%) patients. In all nine patients with positive preoperative test results, tumor was found and was resected for cure in four (44%) patients. In conclusion, preoperative localization of primary gastrinoma is unreliable, and a negative workup does not preclude cure. Therefore, despite the availability of effective palliative antiulcer therapy, all patients exclusive of those with multiple endocrine neoplasia, type I, or metastasis should undergo laparatomy with curative resection in mind.