RATIONALE: Drugs used in chemotherapy work in different ways to stop the growth of tumor cells, either by killing the cells or by stopping them from dividing. Monoclonal antibodies, such as cetuximab, can block tumor growth in different ways. Some block the ability of tumor cells to grow and spread. Others find tumor cells and help kill them or carry tumor-killing substances to them. Giving more than one chemotherapy drug (combination chemotherapy) together with cetuximab may kill more tumor cells.

PURPOSE: This randomized phase II trial is studying three different combination chemotherapy regimens to compare how well they work when given together with cetuximab in treating patients with metastatic esophageal cancer or gastroesophageal junction cancer.

OUTLINE: This is a randomized, multicenter study. Patients are stratified according to histology (squamous cell carcinoma vs adenocarcinoma) and Eastern Cooperative Oncology Group (ECOG) performance status (0 or 1 vs 2). Patients are randomized to 1 of 3 treatment arms. For more information please see the "Arms" section which includes a detailed description of the treatment regimens.

The primary objective of the study is evaluate the tumor response rate (RR) for each of the regimens in this trial and to select the most promising regimen based on RR for further testing in patients with metastatic esophageal or GE junction adenocarcinoma. The secondary objectives are:

1. To evaluate overall survival (OS) for each of the regimens in this trial in patients with metastatic esophageal or GE junction adenocarcinoma.
2. To evaluate progression-free survival (PFS) for each of the regimens in this trial in patients with metastatic esophageal or GE junction adenocarcinoma.
3. To evaluate time to treatment failure (TTF) for each of the regimens in this trial in patients with metastatic esophageal or GE junction adenocarcinoma.
4. To determine the type and severity of toxicities associated with each of these regimens in the multi-institutional phase II setting.
5. Quantitative immunohistochemistry results will be correlated with objective response rate, overall survival and time to progression.
6. To evaluate the cellular damage (apoptosis) as a result of oxaliplatin.
7. To determine if germline EGFR variants correlate with skin rash in patients treated with cetuximab.
8. To evaluate if a correlation exists between germline EGFR variants and tumor EFGR expression as measured by immunohistochemistry.

All subjects must be premedicated with diphenhydramine hydrochloride 50 mg IV (or a similar agent) prior to the first dose of cetuximab in an effort to minimize infusion and hypersensitivity reactions. Premedication is recommended prior to subsequent doses, but the dose of diphenhydramine (or similar agent) may be reduced at the investigator's discretion. More information is detailed in the protocol including a description of the premedication requirements. Patients were closely monitored for treatment-related adverse events. After completion of study treatment, patients are followed periodically for up to 2 years.

1. Metastatic disease of the esophagus or gastroesophageal junction

   1. Histologic, cytologic or radiologic documentation of metastatic squamous cell carcinoma or adenocarcinoma of the esophagus or gastroesophageal junction. Radiologic, endoscopic, histologic or cytologic evidence of locally recurrent or locally residual (post-resection) disease is also permitted.

   2. For the purposes of this study, undifferentiated adenocarcinomas and adenosquamous tumors will be considered as adenocarcinomas. In addition, tumors involving the gastroesophageal junction will be defined by the Siewert classification.

   3. Patients with gastroesophageal junction tumors who are eligible:

      • AEG Type I: Adenocarcinoma of the distal esophagus which usually arises from an area with specialized intestinal metaplasia of the esophagus (eg, Barrett's esophagus, and may infiltrate the esophagogastric junction from above).
      • AEG Type II: True carcinoma of the cardia arising from the cardiac epithelium or short segments with intestinal metaplasia at the esophagogastric junction.

   4. Patients with gastroesophageal junction tumors who are NOT eligible:

      • AEG Type III: Subcardial gastric carcinoma which infiltrates the esophagogastric junction and distal esophagus from below.

2. Patients must have at least one paraffin block available (or at least 15 unstained slides for analysis of tumor EGFR status.

   1. Patients with a history of esophageal and GE junction carcinoma treated by surgical resection who develop radiological or clinical evidence of metastatic cancer do not require separate histological or cytological confirmation of metastatic disease unless an interval of greater than five years has elapsed between the primary surgery and the development of metastatic disease OR the primary cancer was stage I.
   2. Clinicians should consider biopsy of lesions to establish the diagnosis of metastatic esophageal or GE junction carcinoma if there is substantial clinical ambiguity regarding the nature or source of apparent metastases.

3. Patients with Measurable Disease - Lesions that can be accurately measured in at least one dimension (longest diameter to be recorded) as ≥ 20 mm with conventional techniques or as ≥ 10 mm with spiral CT scan.

4. Prior Treatment:

   1. No prior chemotherapy or radiotherapy. No prior therapy which specifically and directly targets the EGF(R) pathway.
   2. No prior allergic reaction to chimerized or murine monoclonal antibody therapy or documented presence of human anti-mouse antibodies (HAMA).
   3. Patients must have completed any major surgery ≥ 4 weeks or any minor surgery ≥ 2 weeks before registration. Patients must have fully recovered from the procedure. Insertion of a vascular access device is not considered major or minor surgery.
   4. No concurrent use of investigational agents is allowed while participating in this study.

5. Patient Characteristics:

   1. ECOG Performance Status of 0-2
   2. ≥ 18 years of age
   3. Patients must be documented to have a stable weight (or less than one pound weight loss) for at least one week prior to registration.
   4. Non-pregnant and not breast-feeding. The effects of cetuximab, cisplatin, epirubicin, fluorouracil, leucovorin, irinotecan, and oxaliplatin on a developing human fetus are not well-known. Because the risk of toxicity in nursing infants secondary to cetuximab, cisplatin, epirubicin, fluorouracil, irinotecan, and oxaliplatin treatment of the mother is unknown but may be harmful, breastfeeding must be discontinued.

6. No myocardial infarction < 6 months prior to registration or New York Heart Association classification III or IV.

7. No ≥ grade 2 diarrhea within 7 days prior to registration.

8. Patients may not concurrently have any of the following conditions:

   1. Known central nervous system metastases or carcinomatous meningitis
   2. Interstitial pneumonia or symptomatic interstitial fibrosis of the lung
   3. Seizure disorder or active neurological disease requiring anti-epileptic medication
   4. ≥ grade 2 peripheral neuropathy

9. No evidence of Gilbert's Syndrome - Patients with Gilbert's Syndrome may have a greater risk of irinotecan toxicity due to the abnormal glucuronidation of SN-38, the active metabolite of irinotecan. Evidence of Gilbert's Syndrome would include documentation of elevation of indirect bilirubin at any time in the patient's medical history.

10. Required Initial Laboratory Data:

    1. Granulocytes ≥ 1500/µl
    2. Platelet count ≥ 100,000/µl
    3. Creatinine ≤ 1.5 mg/dL
    4. AST (SGOT) ≤ 5.0 x Upper limits of normal
    5. Total bilirubin ≤ 1.5 mg/dL
    6. Albumin ≥ 2.5 grams/dL