Final Diagnosis -- Pancreatic Islet Cell Tumor


Contributors' Note:

Unlike other gastro-enteropancreatic endocrine tumors, islet cell tumors are nearly always located within the pancreatic parenchyma, and only rarely occur in the duodenum. These are rare tumors, with an incidence of 1.5/100,000 in the general population (1). The endocrine component of the pancreas is comprised of four types of hormone-producing islet cells: glucagon (A cells), insulin (B cells), somatostatin (D cells), and pancreatic polypeptide. Approximately 50% of islet cell tumors are insulinomas. Glucagonomas, somatostatinomas, and pancreatic polypeptide-secreting tumors together comprise an additional 5-10% of islet cell tumors. Interestingly, approximately 40% of pancreatic endocrine tumors (PETs) express hormones that are not native to the human pancreatic, i.e. gastrin and vasoactive intestinal peptide (VIP) (2).

It has been established that islet cell tumors arise from a primitive neuroendocrine stem cell, with the capacity to differentiate into various hormone-secreting cells. Many PETs, when carefully studied, are comprised of more than one hormone-producing cell type. However, one hormone is usually 'dominant' and is associated with hyperfunction. The proportion of these different cell types within the same tumor can actually change with time; and metastatic lesions can secrete hormones different from the parent tumor. While PETs have been well characterized clinically and immunohistochemically, knowledge regarding the molecular etiology of these tumors lags far behind. While the p53 gene has consistently nonmutated in PETs, overexpression and accumulation of the p53 gene is characteristic of the poorly- differentiated and the well-differentiated malignant PETs (3). A role for BK virus infection has been implicated in human insulinomas but has not yet been definitively established (4). A high rate of loss of heterozygosity at the MEN-1 gene locus of chromosome 11q13 has been found in endocrine tumors in patients with multiple endocrine neoplasia type I (MEN-1) (5). However, the role of the MEN-1 gene in sporadic endocrine tumors, such as this case, remains uncertain.

One of the greatest challenges to the pathologist in the histologic evaluation of these tumors is to provide the clinician with information regarding their malignant potential. Overall, insulin-producing tumors are the most common PET, and 94% are benign. Gastrin- and glucagon-producing, as well as nonfunctional, PETs have a greater potential to be malignant. One clinicopathologic classification scheme separates pancreatic endocrine "carcinomas" from well-differentiated pancreatic endocrine "tumors" based on the presence of local invasion and/or metastases, with no deference to the functional status of the tumor (3). According to thise scheme, functioning insulinomas and nonfunctioning tumors confined to the pancreas, which are nonangioinvasive and less than 2 cm, have a near 100% probability of benign behavior. Insulinomas or nonfunctioning tumors greater than 2 cm and/or angioinvasive are considered well-differentiated pancreatic tumors of "uncertain behavior". It is important to note that functioning tumors other than insulinomas (i.e. gastrinomas, VIPomas, somatostatinomas) are considered of "uncertain behavior" regardless of size. Those tumors with angioinvasion, perineural invasion, more than 2 mitoses per 10 high-power fields, and a Ki67 proliferation index of greater than 2% have an increased risk of malignant behavior (6). Poorly-differentiated PETs have highly atypical small, poorly granular cells arranged in ill-defined solid aggregates, often with necrosis. Histologically, they are characterized by 10 or more mitoses per high-power field, distinct angioinvasion, and a high Ki67 proliferation index (3). In addition, immunoreactivity for p53 is usually found (3). The poorly differentiated tumors are highly invasive and generally have liver, and often extra-abdominal, metastases at the time of diagnosis. Radiologic indices that are suggestive of a malignant PET are larger size, cystic and necrotic areas, local/vascular invasion, and calcification, regardless of the functional status of the tumor (7).

In the present case, the tumor was predominantly immunoreactive for insulin, with a smaller percentage of cells staining for glucagon and somatostatin. The patient experienced low fasting glucose levels on several occasions, although he did not complain of hypoglycemic symptoms. It has been suggested that only functional tumors be designated as "insulinomas" or "gastrinomas", excluding clinically silent tumors that are immunoreactive for hormones. There is no definitive data that detectable hormone levels have a prognostic significance (3). Although no obvious evidence of malignancy, such as excessive mitotic activity, necrosis, or definite vascular invasion, were present in this tumor, the large size of this islet cell tumor prompted the designation of "uncertain metastatic potential".


  1. Delcore R and Friesen SR. (1994) Gastrointestinal neuroendocrine tumors. J Am Coll Surg 178:187-211.
  2. Perry RR and Vinik AI. (1995) Diagnosis and management of functioning islet cell tumors. J Clin Endo Metab 80(8):2273-2278.
  3. Rindi G, Capella C, and Solcia E. (1998) Cell biology, clinicopathological profile, and classification of gastro- enteropancreatic endocrine tumors. (1998) J Mol Med 76:413-420.
  4. Corallini A, Pagnani M, Viadana P et al. (1987) Association of BK virus with human brain tumors and tumors of pancreatic islets. Int J Cancer 39:60-67.
  5. Radford DM, Ashley SW, Wells SA et al. (1990) Loss of heterozygosity of markers on chromosome 11 in tumors from patients with multiple endocrine neoplasia type I. Cancer Res 50:6529-6533.
  6. LaRosa s, Sessa F, Capella C et al. (1996) Prognostic criteria in nonfunctioning pancreatic endocrine tumors. Virchows Arch 55:355-361.
  7. Buetow PC, Parrino TV, Buck JL et al (1995) Islet cell tumors of the pancreas: pathologic-imaging correlation among size, necrosis, cysts, calcification, malignant behavior, and functional status. Am J Radiol 165:1175-1179.

Contributed by Karen K Deal, MD, PhD and Uma N. M. Rao, MD

Case  IndexCME Case StudiesFeedbackHome