Affordable Access

[Identification of pancreatic insulinomas. The role of magnetic resonance].

Authors
  • Pavone, P
  • Catalano, C
  • Cardone, G
  • Di Girolamo, M
  • Laghi, A
  • Messina, A
  • Scipioni, A
  • Sarrantonio, A
  • Passariello, R
Type
Published Article
Journal
La Radiologia medica
Publication Date
Dec 01, 1995
Volume
90
Issue
6
Pages
734–739
Identifiers
PMID: 8685457
Source
Medline
License
Unknown

Abstract

Insulinomas are the most common endocrine tumors of the pancreas. Although their clinical and biochemical diagnosis is extremely accurate, these lesions must be correctly located preoperatively to plan the best possible treatment. Many different noninvasive diagnostic techniques have been proposed, with only partially satisfying results. For many years such non-invasive modalities as arteriography have been considered the most reliable and accurate diagnostic tools. Recently, however, MRI has been proposed for best pancreas imaging, especially for tumor detection, and its results have been excellent especially thanks to its optimal contrast resolution. Fat-suppressed and turbo-spin-echo (TSE) sequences now yield even better results. In our study we examined 21 patients, whose symptoms were typical of insulinoma, using two different 0.5T magnets, one of which with 15 mT/m gradients. We used T1-weighted spin-echo (SE) sequences with and without respiratory compensation, SE ad TSE T2-weighted and fat-suppressed T1-weighted SE sequences. In 16 patients the lesions were correctly diagnosed and located, as confirmed at surgery. In the other 5 patients, MRI failed to locate the insulinomas. Three of these 5 patients were submitted to surgery, during which palpation and intraoperative US findings were also negative for a solid mass. The remaining 2 patients underwent clinical follow-up. Our study demonstrated that MRI, especially with fat-suppressed and T2-weighted TSE sequences, is a very accurate modality to detect pancreatic insulinomas and can therefore be proposed as the only preoperative technique.

Report this publication

Statistics

Seen <100 times