Gemcitabine and S-1 for Locally Advanced Unresectable or Metastatic Pancreatic Cancer

OBJECTIVES: Primary * Correlate intratumoral expression level of ribonucleotide reductase subunit 1 (RRM1) with response to gemcitabine hydrochloride therapy in patients with locally advanced unresectable or metastatic adenocarcinoma of the pancreas. Secondary * Correlate intratumoral expression levels of other genes (e.g., deoxycytidine kinase \[dCK\], equilibrative nucleoside transporter 1 \[ENT1\], and concentrative nucleoside transporters 1 and 3 \[CNT1 and CNT3\]) with response in these patients. * Determine, preliminarily, the median survival of these patients, using a therapeutic strategy entailing sequential addition of agents and decision making based on early CA 19-9 biomarker response. * Determine the safety of this approach. * Determine the percentage of patients classified as potential biomarker responders. * Determine the time to progression with each successive line of treatment. * Determine the proportion of patients with ≥ 25% decline in CA 19-9 biomarker (i.e., biomarker response) with each successive line of treatment. Tertiary * Identify other genes that may mediate sensitivity to gemcitabine hydrochloride, S-1, and other agents with activity in pancreatic cancer. * Determine the frequency of host genetic polymorphisms in various nucleoside transporters. OUTLINE: This is a multicenter. * Initial treatment (gemcitabine hydrochloride alone): Patients receive gemcitabine hydrochloride IV over 100 minutes on days 1, 8, and 15. CA 19-9 levels are assessed in weeks 1 and 3 of each course. Patients who are biomarker responders continue to receive treatment with gemcitabine hydrochloride alone. Courses repeat every 28 days in the absence of disease progression or unacceptable toxicity. Patients who are no longer biomarker responders or show other evidence of disease progression proceed to therapy comprised of gemcitabine hydrochloride and S1. * Gemcitabine hydrochloride and S-1 treatment: Patients receive gemcitabine hydrochloride IV over 100 minutes on days 1 and 15 and oral S-1 twice daily on days 1-7 and 15-21. Courses repeat every 28 days in the absence of disease progression or unacceptable toxicity. Patients undergo core needle tumor biopsy and fine-needle aspiration at baseline. Tissue samples are analyzed for correlation between transcript and protein expression by immunohistochemistry and for expression of genes and gene products that may mediate sensitivity to gemcitabine hydrochloride (RRM1, ENT1, CNT1 and 3, dCK); S-1, thymidine phosphorylase \[TP\], TS, DPD, and ORPT; and other anticancer treatments (ERCC-1, epidermal growth factor receptor, GSK-3β) by reverse-transcriptase polymerase chain reaction. Tissue samples are also analyzed by microarray and comparative genomic hybridization to identify new genes that may predict chemotherapeutic response or mediate sensitivity to anticancer therapy. Mutational status of KRAS and p53 gene are also assessed. Blood samples are collected at baseline and are analyzed by genotyping assays to identify polymorphic variants of select genes. After completion of study treatment, patients are followed monthly. PROJECTED ACCRUAL: A total of 100 patients will be accrued for this study.