Pancreatic neuroendocrine tumors (PNETs) account for <3% of all pancreatic tumors and occur with an incidence of 1 to 2 per 100,000 people.^1–3 Most are sporadic; however, they may also be associated with multiple endocrine neoplasia 1 (MEN1—parathyroid, pituitary, and PNETs), Von-Hippel-Lindau syndrome (eye, CNS, renal cell carcinoma, pheochromocytoma, and PNETs), or tuberous sclerosis. PNETs are often metastatic at diagnosis, most frequently to the liver. Regional node metastases are particularly common and a regional lymph node dissection is an important part of most surgical procedures. PNETs have a wide spectrum of biologic behavior which is reflected in patients’ 5-year survival rates (97% for insulinomas, 30% for nonfunctional PNETs with metastases). Lesions with higher mitotic rates and Ki-67 index demonstrate a more aggressive biologic behavior.

Clinically, PNETs are divided into two classes based on hormone production. Functional tumors (10% to 30%) refer to those PNETs that overproduce hormones and cause clinical syndromes. The most common examples are insulinomas causing hypoglycemia, gastrinomas causing peptic ulcers, VIPomas resulting in watery diarrhea, and glucagonomas stimulating necrolytic migratory erythema. However, nonfunctional PNETs are actually more common (70% to 90%) and cause symptoms related to local mass effect (abdominal pain, jaundice, intestinal bleeding, and obstruction) and/or distant organ metastases.⁴ Although classified as nonfunctional from a clinical perspective, these tumors may actually produce multiple hormones and peptides including neurotensin, pancreatic polypeptide, chromogranin A, and neuron-specific enolase. Plasma chromogranin A levels are elevated in 60% to 100% of patients with nonfunctional PNETs and may be used to identify disease progression, recurrence following surgical resection, and response to therapy in patients with metastatic disease.⁵

Ki-67

In addition to categorization based on functionality, PNETs have also been grouped by proliferation index termed Ki-67. Ki-67 is a proliferation antigen that is present in all phases of the cell cycle excluding G0. Because it is not involved in the DNA repair process, it is an excellent marker to determine the growth fraction of a cell population. Ki-67 is determined by scanning a slide for maximal staining intensity and then calculating the number of positive cells among ≥2000 cells in ≥10 representative high-powered fields. The index is also reported as MIB-1 which is a monoclonal antibody against Ki-67.⁶ The World Health Organization (WHO) utilizes the Ki-67/MIB-1 index values of <2, 2 to 20, and >20 as indicators of benign/uncertain, low, and high malignant potential, respectively.⁷ Other authors have demonstrated a significant decline in survival with a Ki-67 index >2%.⁸ Highly proliferative tumors are usually >2 cm in size, demonstrate extensive angioinvasion, a Ki-67 of >10%, and, as a result, have a high risk for harboring or developing metastatic disease. Ki-67 has been shown to be an independent prognostic factor for local-regional and distant organ disease.⁶

Chromogranin A

Chromogranin A (CgA) is believed to be the most important tumor marker for PNETs. CgA is a secretory glycoprotein found in neuroendocrine cell vesicles that is co-released with biologically active products and has value both in diagnosis and treatment response. Massironi et al⁹ published the prognostic significance of a CgA response to octreotide in gastroenteropancreatic neuroendocrine tumors. Patients included in the trial had to have either a positive octreoscan or a >50% decline in CgA with administration of octreotide. Two hundred micrograms of octreotide was administered and CgA levels were obtained at 0, 3, and 6 hours. The CgA level correlated significantly with WHO class, TNM stage, and Ki-67 index. A decline in plasma CgA of >30% was indicative of an improved response to somatostatin analogue therapy and an overall better prognosis.⁹

Patients with PNETs are usually long-lived. Over the past decade, readily available, high-quality cross-sectional imaging has resulted in the more frequent diagnosis of PNET at younger and younger ages. For these reasons, quality of life is a very important component in determining optimal treatment as the treatment should never be worse than the disease. Ideally, it would be best to select patients for therapies based on both biological and clinical information. Baseline data should include extent of disease, degree of tumor differentiation, proliferation index, and somatostatin receptor imaging (with or without 18-FDG PET imaging) to assess tumor biology and cell proliferation. What remains unavailable, but potentially relevant, is differential somatostatin receptor expression, the degree of VEGF and mTOR activation, and the mutational status of NET-related genes involved in apoptosis/repair machinery (P53, BCL2, MGMT, MMR, and NER).¹⁰ In addition, it is important to remember that the primary tumor and the metastatic lesion may be disparate and respond differently to the same therapy.

Chemotherapy

Streptozotocin (STZ), 5-fluorouracil (5-FU), and doxorubicin were traditionally the chemotherapeutic drugs of choice for PNET. The Eastern Cooperative Oncology Group (ECOG) studies of the 1970s and 1980s showed response rates for STZ combined with 5-FU or doxorubicin in the 63% to 69% range; however, later studies refuted these results showing response rates of only 39% with a median duration of only 9.3 months.^11–13 Significant toxicity profiles have made their use difficult in patients with large locally advanced tumors, widely metastatic disease, or those with a poor performance status.

O⁶-methylguanine DNA methyltransferase, otherwise known as MGMT, is a DNA repair enzyme that induces cancer cell resistance to O⁶ alkylating agents. The lower the level of available MGMT, the greater the tumor’s sensitivity is to therapy. Temozolomide is a cytotoxic oral alkylator that induces DNA methylation at the O⁶ position of guanine. The result is DNA mismatch, apoptosis, and cell death. Use of capecitabine, which depletes MGMT, together in combination with temozolomide has resulted in a relative response rate as high as 60% to 70%.¹⁴ Strosberg et al¹⁴ reported a series of 30 patients treated with this synergistic combination as having a median progression-free survival of 18 months and a 2-year survival of 92%. They concluded that the combination of capecitabine/temozolomide had a high and durable response rate in patients with PNET. In comparison to streptozotocin-based regimens, response rates, survival, and toxicity were superior with this two-drug combination.¹⁴

Somatostatin Analogues

Somatostatin analogues (SSAs) were initially developed as antisecretory agents to help control symptoms of hormonal syndromes associated with neuroendocrine tumors. More recently, data has shown that these agents also have an antiproliferative effect resulting in stabilization of tumor growth. The PROMID trial was a phase III placebo-controlled prospective randomized controlled trial of well-differentiated inoperable or metastatic, neuroendocrine midgut tumors. Participants were given octreotide LAR 30 mg IM monthly or placebo. Those receiving octreotide experienced a median time to progression of 14.3 months compared to 6 months for placebo (p = 0.000072). At 6 months, tumor progression was 24% for patients receiving octreotide and 66% for those receiving placebo (p = 0.0079). On multivariate analysis, the highest rates of disease stabilization were observed with low hepatic tumor load (<10%) and a resected primary tumor. However, even those with a higher hepatic tumor burden (>10%) experienced a near doubling of the time to progression while on octreotide LAR. Current recommendations support the early use of octreotide LAR (30 mg IM monthly) in patients with metastatic disease—this is obviously an extrapolation of the experience with midgut carcinoid tumors to patients with PNET.¹⁵

Most well-differentiated gastroenteropancreatic neuroendocrine tumors (GEP-NETS) are well vascularized with a high expression of proangiogenic molecules such as vascular endothelial growth factor, along with overexpression of tyrosine kinase inhibitors (epidermal growth factor receptor, insulin growth factor receptor, and downstream signaling pathways IPI3K-AKT-mTOR).

Mammalian Target of Rapamycin

Mammalian target of rapamycin (mTOR) is an intracellular serine/threonine kinase that is a central regulator of multiple signaling pathways (i.e., IGF-1, EGF, VEGF). It regulates apoptosis, cell proliferation, and cell growth by modulating cell cycle progression. Activation of mTOR is linked to increased rates of proliferation and cancer progression; such patients have been found to have shorter progression-free survival and overall survival.¹⁶ Recent whole-exome genomic analysis of PNETs demonstrates that approximately 15% of tumors are associated with somatic mutations in genes associated with the mTOR pathway such as PTEN, PI3K, and TSC2.¹⁷ Everolimus is an oral mTOR inhibitor. The RADIANT 3 trial, published in the New England Journal of Medicine in 2011, was a multicenter, double-blinded, phase III trial of everolimus versus placebo. Those receiving everolimus experienced a prolongation in median progression-free survival of 6.4 months compared to placebo and had a 65% reduction in relative risk of disease progression (p < 0.001). Since crossover was allowed in this trial, no difference in overall survival was detected.¹⁸

Vascular Endothelial Growth Factor

Vascular endothelial growth factor (VEGF) is known to be a key driver of angiogenesis in PNET. PNETs show widespread expression of VEGF receptors along with platelet-derived growth factor receptors (PDGFRs) and c-kit. Sunitinib is a multitarget tyrosine kinase inhibitor that targets VEGFR 1–3, PDGFR alpha and beta, c-kit, flt-3, and the RET proto-oncogene. A recent randomized, double-blinded, placebo-controlled phase III trial of advanced well-differentiated PNET compared sunitinib to placebo. In this study, patients who received sunitinib experienced a median progression-free survival of 11.4 months as opposed to 5.5 months (p < 0.001) for those receiving placebo. The objective response rate was 9.3% versus 0% (p = 0.0066). Median overall survival was never reached; however, hazard ratios favored those treated with sunitinib. The authors concluded that sunitinib improved progression-free survival, overall survival, and overall response rates in patients with PNETs.¹⁹

Building on the synergistic effects of mTOR and VEGFR inhibition, a phase II trial of temsirolimus (mTOR inhibitor) combined with bevacizumab demonstrated that 80% of patients (44 of 50) were free of disease progression at 6 months and the objective response rate was 37%.²⁰ Everolimus has also recently been combined with depot octreotide in a phase II Italian trial of gastroenteropancreatic and lung NETs with 92% of treated patients demonstrated to have clinical benefit and 72% found to maintain stable disease for >6 months.²⁰

Transarterial Embolization and Chemoembolization