Rectal neuroendocrine tumors (NETs) can be classified by histologic pattern and secretory products. Recently, rectal NETs have been noted to exhibit immunohistochemical (IHC) positivity for Islet 1 and PAX8, which are generally considered markers for NETs of pancreatic origin. In this study, we sought to characterize the IHC staining profile of rectal NETs and determine whether there was any correlation between the histologic pattern of rectal NETs and their IHC profile. Fifty-six primary rectal NETs were histologically reviewed and stained with antibodies against Islet 1, PAX8, CDX2, chromogranin A, and synaptophysin. In a subset of 31 cases, immunoreactivity for serotonin, pancreatic polypeptide (PP), and prostatic acid phosphatase (PAP) was also studied. By morphology, the tumors studied included 55 % trabecular, 27 % solid nested, 4 % acinar, and 14 % mixed patterns. Islet 1 was positive in 89 % and PAX8 in 79 % of cases. CDX2 was negative in all 56 cases. Cytoplasmic staining was observed for chromogranin A in 30 % of cases and for synaptophysin in all 56 cases. Cytoplasmic staining for serotonin, PP, and PAP was present in 16, 61, and 97 % of cases, respectively. There was no correlation between histologic pattern and IHC staining pattern with any of the antibodies studied. We have demonstrated that Islet 1 and PAX8 are not entirely specific for NETs of pancreatic origin, as they are expressed in a majority of rectal NETs. Since rectal NETs may show an IHC staining profile which mirrors that of pancreatic NETs (Islet 1 and PAX8-positive, CDX2-negative), a metastatic rectal NET should be considered in the differential diagnosis and ruled out clinically in the work-up of a metastatic NET of unknown primary origin which exhibits this staining profile.

Neuroendocrine tumors can present as liver metastases before discovery of the primary tumor. Islet 1 and PAX8 have recently been proposed as markers for neuroendocrine tumors of pancreatic origin. In this study, we compared the utility of Islet 1 and PAX8 in distinguishing pancreatic neuroendocrine tumors from neuroendocrine tumors of other sites and determined the usefulness of an immunohistochemical panel, including TTF1, CDX2, Islet 1 and/or PAX8, in identifying metastatic pancreatic neuroendocrine tumors. A total of 110 primary neuroendocrine tumors (33 pancreatic, 31 pulmonary, 23 ileal, 14 rectal, and 9 gastric) and 73 metastatic neuroendocrine tumors (28 pancreatic, 5 pulmonary, 37 ileal, 1 rectal, 1 colonic, and 1 duodenal) were studied. Islet 1 and PAX8 were positive in 27/33 (82%) and 29/33 (88%), respectively, of primary pancreatic neuroendocrine tumors, and in 19/28 (68%) and 15/28 (54%), respectively, of metastatic pancreatic neuroendocrine tumors. No cases of primary (0/23) or metastatic (0/37) ileal neuroendocrine tumors were positive with either Islet 1 or PAX8. There was Islet 1 positivity in 2/31 (6%) primary pulmonary, 12/14 (86%) primary rectal, and 1/1 metastatic rectal neuroendocrine tumors, and PAX8 positivity in 7/31 (23%) primary pulmonary, 11/14 (79%) primary rectal, and 2/9 (22%) primary gastric neuroendocrine tumors. ROC curve analysis incorporating sensitivity and specificity data of immunohistochemical panels for metastatic pancreatic neuroendocrine tumors showed that a four-stain panel, including Islet 1, PAX8, TTF1, and CDX2 significantly outperformed a three-stain panel composed of PAX8, TTF1, and CDX2 (P=0.019), and also showed a trend for better performance compared with a three-stain panel composed of Islet 1, TTF1, and CDX2 (P=0.072). Both Islet 1 and PAX8 are reliable immunohistochemical markers for pancreatic neuroendocrine tumors and would be useful adjuncts to other markers (TTF1, CDX2) currently used to work up a metastatic neuroendocrine tumor of unknown primary.