Research in Medicine

Abstract Background : Insulinoma tumor is the most common endocrine tumor of pancreas and is diagnosed by clinical symptoms(whipple’s triad) and paraclinical laboratory test results .surgery is main treatment .Optimum utilization of tumor localization techniques can minimize the loss surgery Complications and outcome of surgery in these patients. Aim of this study was to evaluate the diagnostic power of tumor localization by radiologic methods and surgeon palpation and compare this methods to propose useful recommendations for treatment of patients. Finally we want to propose a diagnostic algorithmic approach according to available radiologic tools in our country. Method and materials:First, referral suspicious insulinoma patients from endocrinology ward of Taleghani hospital studied with CT scan and endosonography to localize tumor and then they are transferred to OR for surgery. After pancreas exploration during the surgery , pancreas is palpated exactly by experienced surgeon to localize tumor and findings are registered, then intraoperative sonography is done and again findings are registered. Then appropriate surgery according to our study (Pancreatectomy or Enucleation) is done. After surgery we followed patients checking BS and pathology report. Finally we calculated the power of the mentioned methods in localization of insulinoma tumor in comparison with pathology report (as gold standard). Results:...positive predictive value of CT scan,endosonography,intraoperative sonography & surgeon palpation was ,in order to,83.3%,92.3%,85.7%&83.3% and the negative predictive value of CT scan,endosonography,intraoperative sonography & surgeon palpation was ,in order to, 27.2%,75%,66.6%&40%. Conclusion: CT scan did not help to diagnosis and treatment of patients in presence of other localization methods,.Endosonography in this study had the highest accuracy. Finally, recommended diagnostic algorithmic approach seems to be useful in treatment of insulinoma patients. the diagnostic power of this algorithmic approach needs more evaluation.