U. Stölzel

BACKGROUND: Endoscopic ultrasonography (EUS) and somatostatin receptor scintigraphy (SRS) can detect a high percentage of gastroenteropancreatic neuroendocrine tumours especially in the upper gastrointestinal tract. The ability of these procedures to localise primary tumour lesions and metastases of gastrinomas and insulinomas was evaluated in comparison with transabdominal ultrasonography (US), computed tomography (CT) and magnetic resonance imaging (MRI). PATIENTS AND METHODS: In a prospective trial, patients with gastrinomas (n = 10) and insulinomas (n = 10) diagnosed by clinical signs and laboratory tests were assessed by EUS, SRS, US, CT and MRI. RESULTS: In 10 patients with gastrinoma and 10 patients with insulinoma, a total of 14 separate primary tumour lesions were histologically confirmed for each of the tumour entities. The mean diameter was 2.1 cm for gastrinomas and 1.5 cm for insulinomas. All insulinomas and nine gastrinoma lesions were located in the pancreas. Three gastrinomas were found in the duodenal wall, one in a periduodenal lymph node, and one in the liver, For gastrinomas, sensitivities were 79% with EUS, 86% with SRS and 29% with CT, US, and MRI. For insulinomas, sensitivities were 93% with EUS, 14% with SRS, 21% with CT and 7% with US and MRI. CONCLUSIONS: EUS is of high value for localising primary lesions of both tumour entities. SRS is a very sensitive procedure for diagnosing of gastrinomas but not insulinomas. CT, US and MRI are primarily useful for visualising metastases.

Well differentiated neuroendocrine tumors (NETs) of the stomach (gastric carcinoid tumors) are observed more often, with a tenfold increase in the US in the last 30 - 35 years, and the prognosis has improved greatly in that time. Nowadays most carcinoids of the stomach are diagnosed at an early stage. Four types of gastric NETs have been proposed and recognition of the type is important for defining the diagnostic approach and treatment. Often gastric NETs (especially type 1) are found incidentally during a gastroscopy performed for other reasons; most of these NETs are smaller than 20 mm in size. Conservative management and endoscopic surveillance is adequate for well differentiated, multifocal gastric carcinoids (type 1 or type 2 gastric NETs) that are less than 10 - 20 mm in diameter, unless they show angioinvasion, infiltrate the muscular wall, or have a proliferation rate above 2 %. Endoscopic ultrasound is the method of choice to determine tumor size and depth of infiltration. It is essential to distinguish between multifocal (types 1 and 2) and unifocal type 3 or type 4 gastric NETs, since surgery is indicated for type 3 gastric NETs larger than 10 mm in diameter and for poorly differentiated (localized) neuroendocrine gastric carcinomas (type 4 gastric NET). For optimal management, the type, biology, and stage of the tumor as well as the individual situation of the patient must be considered. Most patients with well differentiated gastric NETs can be treated conservatively and be followed up with endoscopic surveillance.

Neuroendocrine tumours (NETs) of the upper gastrointestinal tract are mainly located in the pancreas, stomach or duodenum. The aims of preoperative work-up are the localization of primary tumour(s), determination of local tumour invasion, of lymph node metastases and of the hormones secreted by the tumour. Endoscopic ultrasonography (EUS) offers ideal conditions to localize and stage NETs of the foregut. We report our results in localizing and staging NETs of the foregut in 40 patients examined between 1990 and 1997 by EUS, somatostatin receptor scintigraphy (SRS), computed tomography (CT), magnetic resonance imaging (MRI) and transabdominal ultrasound (US). EUS shows the highest sensitivity in localizing insulinomas compared with SRS, US, CT and MRI. US and EUS should be the first-line diagnostics if insulinoma has been proven by a fasting test. Further diagnostic procedures are unnecessary in most cases. Further diagnostics such as CT or MRI to search for distant metastases are necessary in large tumours or local invasive tumours. EUS shows the highest accuracy to detect or exclude pancreatic gastrinomas, but fails to detect extrapancreatic gastrinomas in about 50%. The combination of EUS and SRS gives additional information. First-line diagnostics in gastrinoma patients should be SRS and CT or MRI. If no metastases are detected, EUS should be the next preoperative imaging procedure. In nonfunctional NETs, EUS provides the best information on local tumor invasion and regional lymph node involvement.