Immunohistochemical Characterization of the Origins of Metastatic Well-differentiated Neuroendocrine Tumors to the Liver.

Metastatic neoplasms of unknown primary site pose a major challenge to patient management. As targeted therapies are now being tailored to neuroendocrine tumors (NETs) of different primary sites, identifying the origin of metastatic NETs has become increasingly important. Compared with more extensive efforts on metastatic adenocarcinomas of unknown primary, the literature on metastatic NETs (often to the liver) is relatively sparse and most studies are based on primary tumors. We sought to study metastatic well-differentiated NETs to the liver to identify markers that predict the site of origin. Eighty-five metastatic NETs to the liver were retrieved from the pathology archive. The primary sites were determined based on either pathologic review of the primary tumors (in most cases) or radiologic/clinical findings. Immunohistochemical labeling for TTF1, CDX2, ISL1, NKX2.2, and PDX1 was performed on either tissue microarrays or whole sections. The primary sites of the NETs in the study cohort included: pancreas (35%), small intestine (32%), rectum (8%), stomach (2%), bile duct (1%), lung (9%), and unknown primary (12%). We found predominant expression of TTF1 in lung carcinoid (63%), CDX2 in small intestinal (89%) and ISL1 in pancreatic NETs (77%), respectively. NKX2.2 was mainly expressed in NETs of the digestive organs. PDX1 was detected in a small percentage of pancreatic, small intestinal and the single bile duct NET. There was no statistically significant association between tumor grade (World Health Organization G1 vs. G2) and the expression of any of the above markers. The 3-marker panel (TTF1, CDX2, and ISL1) had sensitivities of 81%, 89%, and 63%, specificities of 100%, 94%, and 100%, positive predictive values of 100%, 89%, and 100%, and negative predictive values of 84%, 94%, and 96% in separating metastatic NETs into 3 major primary sites: pancreas/rectum, small intestine, and lung, respectively, with an overall accuracy of 82%. Furthermore, this panel predicted a primary site for 6 of the 10 NETs of unknown primary, which reduced the NETs of unknown primary from 12% to 5%. Thus, through immunohistochemical study of a large series of metastatic NETs to the liver, we have demonstrated the utility of a 3-marker panel for the identification of one or more potential primary sites of most metastatic NETs, which could provide practical guidance in patient management.