RATIONALE: Drugs used in chemotherapy, such as melphalan, work in different ways to stop the growth of tumor cells, either by killing the cells or by stopping them from dividing. Giving melphalan directly into the arteries around the tumor may kill more tumor cells. It is not yet known whether hepatic arterial infusion with melphalan is more effective than standard therapy in treating liver metastases due to melanoma.

PURPOSE: This randomized phase III trial is studying hepatic arterial infusion with melphalan to see how well it works compared to standard therapy in treating patients with unresectable liver metastases due to melanoma.

OBJECTIVES:

Primary

• Compare the hepatic progression-free survival of patients with unresectable liver metastases secondary to ocular or cutaneous melanoma treated with percutaneous isolated hepatic arterial perfusion (PHP) with melphalan with subsequent venous hemofiltration vs the best alternative standard treatment.

Secondary

• Determine the response rate and duration of response in patients treated with melphalan PHP.
• Determine the patterns of recurrence in patients treated with melphalan PHP.
• Compare the overall survival of patients treated with these regimens.
• Compare the safety and tolerability of these regimens in these patients.
• Determine the pharmacokinetics of melphalan after PHP.

OUTLINE: This is a multicenter study. Patients are stratified according to site of disease (ocular vs cutaneous). Patients are randomized to 1 of 2 treatment arms.

• Arm I: Patients undergo an isolated hepatic arterial infusion of melphalan over 30 minutes on day 1. Treatment repeats every 4 weeks for 4 courses in the absence of disease progression or unacceptable toxicity. Patients with complete or partial response undergo 2 additional courses in the absence of ongoing or increasing toxicity.
• Arm II: Patients receive the best alternative therapy comprising supportive care, systemic or regional chemotherapy, hepatic artery (chemo)-embolization, or any other appropriate therapy at the National Cancer Institute or therapy at the discretion of their physician. Patients may cross over to arm I if they have evidence of disease progression.

Blood samples are collected periodically for pharmacokinetic analysis of melphalan.

After completion of study treatment, patients are followed periodically for 4 years and then annually for survival.

PROJECTED ACCRUAL: A total of 92 patients will be accrued for this study.

DISEASE CHARACTERISTICS:

• Histologically or cytologically confirmed liver metastases secondary to cutaneous or ocular melanoma

  • Unresectable disease
  • Predominantly in the parenchyma of the liver

• Measurable disease by CT scan and/or MRI

• Limited unresectable extrahepatic disease allowed provided the life-limiting component of progressive disease is in the liver, including, but not limited to, any of the following:

  • Up to 4 pulmonary nodules, each < 1 cm in diameter
  • Retroperitoneal lymph nodes < 3 cm in diameter
  • Less than 10 skin or subcutaneous metastases < 1 cm in diameter
  • Asymptomatic bone metastases that are eligible for or have undergone palliative external-beam radiotherapy
  • Solitary metastasis to any site that can be resected

PATIENT CHARACTERISTICS:

• Life expectancy ≥ 3 months
• ECOG performance status 0-2
• Bilirubin < 3.0 mg/dL
• PT within 2 seconds of upper limit of normal (ULN)
• AST/ALT ≤ 10 times ULN
• Platelet count > 75,000/mm^3
• Hematocrit > 27% (may be achieved with a transfusion)
• Absolute neutrophil count ≥ 1,300/mm^3
• Creatinine ≤ 1.5 mg/dL OR creatinine clearance > 60 mL/min
• Fertile patients must use effective contraception
• Not pregnant or nursing
• Negative pregnancy test
• No history of congestive heart failure
• LVEF ≥ 40%
• No significant chronic obstructive pulmonary disease (COPD) or other chronic pulmonary restrictive disease
• FEV_1 ≥ 30%
• DLCO ≥ 40% of predicted
• Weight ≥ 35 kg
• No untreated active bacterial infection with systemic manifestations (e.g., malaise, fever, and leucocytosis)
• No severe allergic reactions to iodine contrast unless reaction can be controlled by antihistamines and/or steroids
• No known hypersensitivity to melphalan
• No positive serology for HIV, hepatitis B surface antigen, or hepatitis C antibody (pharmacokinetics portion of the study only)
• No known latex allergy
• No Childs B or C cirrhosis
• No evidence of portal hypertension by history, endoscopy, or radiological study
• No prior history of gastrinoma

PRIOR CONCURRENT THERAPY:

• See Disease Characteristics
• At least 1 month since prior chemotherapy, radiotherapy, or biologic therapy for this cancer and recovered
• No prior regionally delivered melphalan
• No prior Whipple procedure
• No concurrent immunosuppressive therapy
• No concurrent chronic anticoagulation therapy