This phase II study is designed to assess whether treatment with capecitabine/temozolomide (CAP/TEM) is safe and effective in treating subjects with progressive, differentiated, metastatic neuroendocrine tumors (NET). The primary objective of the study is to determine the radiologic response rate to this regimen in progressive, metastatic, differentiated neuroendocrine cancers.

Neuroendocrine tumor (NET) is a classification that has evolved over time to include a group of related tumors which all originate from neuroendocrine cells. This group includes carcinoid tumors, pancreatic endocrine tumors (PETs), catecholamine-secreting tumors (e.g. pheochromocytomas), medullary carcinoma of the thyroid and small cell lung cancer. Carcinoid tumors are mostly derived from serotonin-producing enterochromaffin cells, occuring most frequently in the gastrointestinal tract (67.5%) and the bronchopulmonary system (25.3%). Pancreatic endocrine tumors (PETs) arise from the several types of pancreatic islet cells, which manifest as insulinomas, somatostatinomas or glucagonomas. NETs are broadly classified as functional or nonfunctional, as determined by whether plasma hormone elevation and endocrine symptoms occur. NETs are classified into two groups: 1) rapidly growing anaplastic small cell cancers such as small cell lung cancer and small cell carcinomas of the GI tract and 2) slow growing, more differentiated NETs such as carcinoid and PET.

In total, an estimated 12,000 - 15,000 cases of NETs (not counting small cell carcinomas) are diagnosed in the United States annually. The incidence of carcinoid tumors alone is estimated to be 2 per 100,000 in the United States (5,400 cases/yr/U.S.). PETs are less common, with about 1,000 new cases per year in the United States. Carcinoids and PETs are potentially curable by surgical resection; the 5-year survival rates in patients with localized carcinoid is 78.2%. However, these tumors are frequently indolent in their growth and patients often present with unresectable or metastatic disease (80% of all cases). The hormonal symptoms that may accompany their disease, as exemplified by carcinoid syndrome, complicate the management of these patients. Hormonal therapy, namely octreotide, is used to relieve symptoms and has been reported to have a response rate of 1-5% by itself. Metastatic disease is associated with significantly worse prognosis; carcinoid patients with visceral metastases have a 5-year survival rate of 38.5%. Based on the efficacy of the combination of cisplatin and etoposide in treating small cell lung cancer, these agents have been explored in the treatment of pancreatic islet cell tumors and carcinoids. In general, etoposide-cisplatin regimens have poor response rates for the slow growing, differentiated NET group with an average response rate of 7-10%. Furthermore, these cisplatin-etoposide regimens have been associated with significant toxicities, including frequent severe neutropenia, ototoxicity, neurotoxicity and nephrotoxicity.

An Eastern Cooperative Oncology Group (ECOG) trial of patients with metastatic or unresectable progressive pancreatic islet cell tumors, including poorly and well-differentiated PETs, showed that a regimen of streptozocin and doxorubicin had a significantly superior objective response rate compared to a combination of streptozocin and fluorouracil (69 versus 45%, p=0.05). The study used a definition for objective response that included regression of the tumor mass, regression of malignant tumor causing hepatomegaly or a reduction in excessive hormone production. Streptozocin-doxorubicin was also significantly superior to streptozocin-fluorouracil in terms of median time to tumor progression and of median overall survival (2.2 versus 1.4 years, p=0.004). However, significant toxicity was associated with either streptozocin-based regimen with roughly 80% in either arm experiencing vomiting that lasted throughout the 5-day course of streptozocin per cycle. Additionally, Grade 3/4 leukopenia occurred in 25% of patients who received streptozocin-fluorouracil, with one treatment-related death secondary to leukopenia complicated by sepsis. Notably, streptozocin has significant renal toxicity causing significant proteinuria. Thus, the doubtful efficacy of streptozocin-based combinations and the significant associated toxicity has limited the role of cytotoxic chemotherapy in the treatment of differentiated NETs.

In the lab, the investigators have found that capecitabine (5-DFUR), an oral pro-drug for 5-fluorouracil (5-FU), and temozolomide were synergistic for induction of apoptosis in 2 human NET cell lines. The mechanism and pathways involved are under investigation, but it was found to be important for the synergism that temozolomide be exposed to the NET cell lines during the end of the capecitabine exposure. The team believes that the combination of temozolomide and capecitabine will prove to be an effective regimen. Our hypothesis is that the DNA damage induced by capecitabine by incorporation of 5-FdUTP into DNA and reducing thymidine pools by inhibition of thymidylate synthase via 5-FdUMP will synergistically potentiate the effect of temozolomide as an alkylator by reducing the repair activity of O6-alkylguanyl-alkyl-transferase (O6-AGAT). O6-AGAT is a DNA repair enzyme which removes temozolomide-alkylated groups from guanine. A 5-day regimen of temozolomide is vital to decreasing O6-AGAT levels by direct binding which leads to a suicide inactivation of O6-AGAT-mediated DNA repair. This saturates O6-AGAT after 23 days of temozolomide, thus allowing the last 23 days of dosing to induce alkylation of DNA and thereby induce apoptosis. The investigators found that cells with prior 5-FU exposure were more sensitive to the induction of apoptosis by temozolomide.

Another fundamental rationale and hypothesis which the investigators have developed into the synthesis of a novel regimen for NET is based upon cytokinetics and p53. NET are characteristically very slow growing, yet fatal, cancers with the great majority of them having wild type p53. Therefore, their drug resistance is probably not based upon mutational p53 causes because of the wild type p53 status but rather upon their slow cytokinetics. The best way to kill slow growing tumors with a long interval in G0 phase is with lipophilic alkylators (i.e. Temodar) and utilizing continuous exposure to antimetabolites such as Xeloda or continuous infusion 5-FU. Xeloda's half-life is 11 hrs so q 12 hr dosing is roughly equivalent to continuous infusion. The investigators believe that the hypothesis is correct and well grounded in pharmacologic and cell cycle principles.

The investigators have to date pilot experience of ten patients who received capecitabine, total of 1500 mg/m2/day/PO, for fourteen days, with temozolomide 150-200 mg/m2 given on the last five days of their course of capecitabine. All of our initial 10 patients with progressive, differentiated NET have had dramatic symptomatic pain relief and at least 75% reduction in their tumor markers. Five patients had metastatic carcinoid and 5 had metastatic pancreatic NET. All patients had progressive liver metastases, all 10 patients had failed octreotide therapy with long acting somatostatin, and 7/10 had failed prior chemotherapy regimens. One carcinoid patient had a complete response (CR) proven by surgery and is now without any tumor recurrence 22 months out from surgery and chemotherapy. Three patients had a partial response (PR) and one patient had a minor response (MR) in their liver metastases. Two other patients experienced stable disease (SD) for 6 and 8 months while on therapy. The overall response rate proven by CT or MRI scans (CR, PR and MR) is 50% to date. Overall, clinical benefit of this lab based regimen occurred in 7/10 patients (CR, PR, MR and SD). Toxicities have all been minor with none over grade 2 myelosuppression. There were no hospitalizations or complications or side effects except grade 12 nausea during temozolomide therapy. Therefore, the study seeks to evaluate the role of these two drugs in this disease.

Inclusion Criteria:

• Patients must have a tissue diagnosis of any of the following metastatic, well or moderately differentiated, slow growing neuroendocrine tumor and must demonstrate progressive metastatic disease by prior serial computerized tomography (CT) or magnetic resonance imaging (MRI) scans, or have increased symptoms from their tumors while on sandostatin LAR or octreotide.
• Carcinoid tumors originating anywhere in the body including the gastrointestinal (GI) tract or bronchial tree
• Pancreatic neuroendocrine tumors (including functional and non-functional islet cell, insulinomas and glucagonomas)
• Pheochromocytomas, gastrinomas (Zollinger-Ellison Syndrome), multiple endocrine neoplasia (MEN) Type I/II, paragangliomas, adrenal carcinomas with NET markers by immunohistochemistry (IHC) or serum.
• Somatostatinoma, VIPoma, Merkel Cell tumors, medullary thyroid carcinoma
• Neuroendocrine tumors of unknown primary site
• Any other tumors with differentiated neuroendocrine features may be included such as aggressive pituitary adenomas/carcinomas, which are neuroendocrine in origin
• Patients must have progressed on octreotide therapy (up to and including Sandostatin LAR-60 mg/month) and/or radioactive isotopes linked to octreotide or its congeners if they has a positive octreotide scan. Patients who have negative or mildly positive octreotide scans are exempt from this requirement. Exceptions to this requirement are patients who have NETs in the pituitary gland. Sandostatin does not cross into the pituitary blood supply well.
• Measurable disease: Any primary and/or metastatic mass reproducibly measurable in one or two diameters by Response Evaluation Criteria In Solid Tumors (RECIST) parameters by CT scan or MRI scan.
• Ineligible for other high priority national or institutional studies
• Prior radiation and surgery allowed: ≥3 weeks since surgery or chemotherapy or hepatic embolization/chemoembolization or radioactive isotopes (i.e. Yttrium 90) ≥4 weeks since radiation therapy (RT)
• Non pregnant females, not in menopause, who are not breast feeding with a negative serum β-HCG (human chorionic gonadotropin) test within 1 week of starting the study. Men and women of childbearing potential must consent to using effective barrier contraception while on treatment and for 2 months thereafter.

Exclusion Criteria:

• Prior chemotherapy with capecitabine or temozolomide. Patients previously treated with continuous infusion 5-FU or any schedule of DTIC (dacarbazine), which are similar to capecitabine and temozolomide, respectively, will be excluded. Patients can have had prior therapies up to 3 prior chemotherapy regimens such as bolus 5-FU, streptozocin, anthracyclines, Camptothecin-11 (CPT-11), etoposide, or a platinum agent
• Hypersensitivity: Patients with a history of severe hypersensitivity reaction to capecitabine, 5-FU, temozolomide or DTIC will be excluded (i.e. anaphylaxis or anaphylactoid reactions)
• Serious medical or psychiatric illness preventing informed consent or intensive treatment (e.g, serious infection)
• Patients with tumor which has spread to the central brain (cerebral/cerebellum) and spinal cord.
• Patients with compromised immune systems are at increased risk of toxicity and lethal infections when treated with marrow-suppressive therapy. Therefore, HIV-positive patients are excluded from the study
• Prior malignancies in the last 5 years other than; curatively treated carcinoma in-situ previously treated with curative intent (cancer free for the past year)

Number at risk reflects the participants who passed the screening.