Prognostic factors in metastatic melanoma

BACKGROUND: Each year, 6000 people die in the United States from metastatic melanoma. Further study of factors affecting the prognosis of patients with this disease is needed. METHODS: The authors analyzed response and survival data from 635 patients who had entered three Eastern Cooperative Oncology Group trials for metastatic melanoma. RESULTS: Factors associated with poorer survival after study entry included poor performance status and the presence of symptoms, such as reduced appetite, fever, or nausea/vomiting. Male patients had poor survival, as did patients entering the study less than 1 year after a documented recurrence to study entry. As expected, characteristics of the initial primary disease (treatment and symptoms) had little association with survival after entering the advanced disease protocol. Two summary measures of the extent of metastatic involvement had a strong influence on survival. These were the number of nonbone metastases and the clinician's assessment as to the most significant metastatic site. Patients with the liver as their clinically most significant metastatic site had a poorer prognosis than those otherwise classified, including those with central nervous system metastases. The prognosis also worsened with an increasing number of sites of nonbone metastases, including skin and soft tissue. Tumor response occurred in only 11% of the patients. Patients with poor performance status and those with lung involvement had a significantly lower response rate than did others. Although the frequency of response was low, patients with objective responses survived significantly longer than did the nonresponders (based on an analysis appropriately adjusted for the time of response using a time-dependent proportional-hazards model). CONCLUSIONS: These results provide useful guidelines for the design and analysis of clinical trials in metastatic melanoma.