RATIONALE: Drugs used in chemotherapy, such as gemcitabine, work in different ways to stop the growth of tumor cells, either by killing the cells or by stopping them from dividing. Radiation therapy uses high-energy x-rays to kill tumor cells. Giving chemotherapy together with radiation therapy may kill more tumor cells.

PURPOSE: This phase I trial is studying the side effects and best dose of gemcitabine when given together with radiation therapy in treating patients with locally advanced upper gastrointestinal cancer.

OBJECTIVES:

• To determine the feasibility of combining preoperative or intraoperative gemcitabine hydrochloride with intraoperative radiotherapy.
• To determine the tolerance of gemcitabine hydrochloride given concurrently with external-beam radiotherapy.
• To measure biochemical parameters in tumors that may correlate with the effectiveness of therapy.

OUTLINE: Patients receive gemcitabine hydrochloride IV 12-18 hours prior to planned surgery. All patients then undergo an exploratory laparotomy that may include tumor debulking, Whipple-type resection (pancreaticoduodenectomy), total pancreatectomy, gastrojejunostomy, total or partial gastrectomy, or cholecystectomy and en bloc resection depending on the extent of the disease. Patients with no metastatic disease beyond regional lymph nodes also undergo intraoperative radiotherapy.

Beginning 2-6 weeks after surgery, patients undergo external-beam radiotherapy (EBRT) once a day 5 days a week for up to 7 weeks. Patients also receive escalating doses of gemcitabine hydrochloride IV at the beginning of each week of EBRT.

Patients undergo tissue sample collection periodically for correlative studies. Samples are analyzed for thymidylate synthase (TS), ribonucleotide reductase (RR), excision-repair-cross-complementing (ERCC)-1 protein, deoxycytidine kinase mRNA. Biopsy tissues are also analyzed for gemcitabine triphosphate, dATP, and dCTP content. p53 status is assessed via immunohistochemistry and mRNA levels via quantitative polymerase chain reaction (PCR).

After completion of study treatment, patients are followed every 3 months for 2 years, every 6 months for 2 years, and then once a year thereafter.

DISEASE CHARACTERISTICS:

• Diagnosis of any of the following upper gastrointestinal malignancies:

  • Localized pancreatic adenocarcinoma

    • Stage I, II, or III disease
    • Parapancreatic node involvement and locally recurrent disease allowed

  • Locally advanced biliary, gallbladder, or ampullary adenocarcinoma

    • Stage II, III, or locally recurrent disease

  • Histologically confirmed locally advanced gastric adenocarcinoma

    • T3, T4, or node positive OR locally recurrent disease

  • Histologically confirmed locally advanced duodenal cancer

    • Stage II or III disease

• Locally advanced, but unresectable cancers may be included on protocol if appropriate for intraoperative radiotherapy (IORT)

• Other histologies may be considered for this protocol except for lymphoma, sarcoma, or neuroendocrine tumors

• Patients with evidence of metastatic disease are eligible if there is significant local disease warranting surgery and IORT

PATIENT CHARACTERISTICS:

• Karnofsky performance status > 60%
• Life expectancy > 4 months
• Absolute neutrophil count > 1,500/mm^3
• Platelet count > 100,000/mm^3
• Serum creatinine < 2.0 mg/dL
• ALT < 3 x normal
• Bilirubin < 2 x normal
• Must be able to give voluntary informed consent
• No severe intercurrent illness that would make the patient inappropriate for laparotomy or otherwise inappropriate for treatment on protocol
• Prior history of malignancy allowed

PRIOR CONCURRENT THERAPY:

• More than 4 weeks since prior chemotherapy (6 weeks for mitomycin C)
• Prior gemcitabine hydrochloride allowed