RESUMO
Neuroendocrine tumors (NETs) with unknown primary (CUP-NET) are associated with a poor prognosis (10-year survival 22%), grade 1 and 2 NETs having a more favorable outcome than grade 3 (also called carcinoma). There is evidence that an effort should be made to localize the primary tumor even in the presence of metastasis because resection of the primary tumor(s) may improve disease-free and overall survival, and because the choice of chemotherapeutic agent depends on the location of the primary tumor. Localization of the tumors remains challenging and often relies on a combination of radiological, endoscopic and functional imaging. The functional imaging protocol for evaluation of these patients has historically relied on somatostatin receptor scintigraphy (SRS). However, the sensitivity and specificity of SRS may be unsatisfactory, especially for NETs of midgut origin. Newer PET radiotracers such as (68)Ga-labeled somatostatin analogs ((68)Ga-DOTA-SSTa) and (18)F-DOPA have shown promise. In direct comparisons between (68)Ga-DOTA-SSTa PET/CT and (99m)Tc-HYNIC-octreotide/(111)In-pentetreotide SPECT(/CT), (68)Ga-DOTA-SSTa performed better than other techniques, giving a compelling reason for switching from SPECT/CT to PET/CT imaging. (18)F-DOPA performs better than SRS and CT in well-differentiated NETs of the small intestine. For detecting pancreatic NETs, the high background uptake of (18)F-DOPA by the normal exocrine pancreas can be somewhat overcome by pretreatment with carbidopa. We have suggested a protocol in which SRS is replaced by one of the two agents (preferably with (68)Ga-DOTA-SSTa, alternatively (18)F-DOPA) as first-line nuclear tracer for detection of CUP-NET in patients with well-differentiated NETs and (18)F-FDG PET/CT may be an additional diagnostic test for poorly differentiated tumors and for prognostication. In the near future, it is expected that patients with CUP-NET will benefit from newly developed PET approaches (radiopharmaceuticals) and intraoperative PET imaging.