Prospective study on the detection of insulinomas by endoscopic ultrasonography.

Abstract

The correct localization of insulinomas using endoscopic ultrasonography (EUS) has been reported to be as high as 80% in multicenter patient cohorts. Over 24 months, we prospectively investigated 14 patients (11 women, three men) with a definite biochemical diagnosis of endogenous hyperinsulinism prior to surgical exploration and removal of an insulinoma. The endoscopic investigator was not aware of any other imaging results if they had been performed in referring hospitals. The overall sensitivity of EUS in the detection of pancreatic insulinomas was 57% (eight of 14 tumors); the sensitivity for insulinomas in the head of the pancreas was 83% (five of six); and 37% (three of eight) for tumors in the tail of the pancreas. The actual median diameter of undetected tumors was 11 x 9.5 mm, the median volume 0.66 ml (range 0.13 - 2.6 ml). The median diameter of correctly detected tumors was 16 x 11 mm, the median volume 1.37 ml (range 0.7 - 6.3 ml), the differences not being significantly different. In two patients, false-positive results were caused by peripancreatic lymph nodes. The sensitivity of EUS in the detection of pancreatic insulinomas depends on the location of the tumor, and possibly on the size of the tumors. Tumors not detected by EUS were likely to be smaller than detected tumors, and were likely to be located in the tail of the pancreas.