Background: Preoperative localisation of insulinomas has been regarded unnecessary, given the significantly higher detection rates of intraoperative ultrasonography and bidigital palpation. These are mandatory before endoscopic surgery.
Methods: 67 patients operated on for organic hyperinsulinism were retrospectively analysed regarding tumour localisation within the pancreas, tumour size, histological findings, sensitivities of preoperative imaging methods, and surgical techniques.
Results: 59 patients (88%) had solitary insulinomas, four patients (6%) multiple insulinomas and four adult patients (6%) nesidioblastosis. Well-differentiated neuroendocrine tumours with benign behaviour (including four patients with nesidioblastosis) were diagnosed in 53 patients (79%), tumours with uncertain behaviour in nine patients (13%) and well-differentiated neuroendocrine carcinomas in five (8%). Tumours were evenly distributed throughout the pancreas. Endoscopic ultrasound localised tumours in 15 out of 21 patients (71%), conventional computed tomography (CT) in 7 out of 21 (33%), single-slice helical CT in 7 out of 12 (58%), multidetector CT in 5 out of 5 (100%), magnetic resonance imaging in 11 out of 13 (85%) and angiography in 15 out of 23 (65%). Various combinations of available methods achieved a sensitivity of 88% (49 patients true positive, 4 true negative, 7 false negative). Of 59 patients, solitary insulinomas were enucleated in 47 (80%), 11 patients underwent conventional open resection and one patient endoscopic distal pancreatic resection. Patients with nesidioblastosis or multiple tumours underwent pancreatic resections alone or in combination with enucleations.
Conclusion: After biochemical diagnosis of organic hyperinsulinism, preoperative localisation is necessary for planning endoscopic pancreatic surgery, because of the possibility of multiple insulinomas, malignancy or nesidioblastosis in adults.