Chemoimmunotherapy for Metastatic Melanoma—Sequence is Important

Abstract & Commentary

Synopsis: Sequential administration of cisplatin, vinblastine, and dacarbazine before or after interleukin-2 (IL-2) and interferon-" resulted in a 60% overall response rate in 62 patients. The complete responses were durable.

Source: Legha SS, et al. Ann Oncol 1996;7:827-835.

Therapy for metastatic melanoma is pretty disappointing. Although there have been hints and allegations of activity from combination chemotherapy programs with or without tamoxifen, and biologic agents produce responses in a whopping 10-15% of patients, it is a rare patient who actually appears to benefit dramatically from these treatments. As I recall from undergraduate psychology classes, the best way to get a particular behavior to persevere is to provide intermittent positive reinforcement. That certainly must be the explanation for our persistence in using the currently available toxic and not very effective regimens.

Despite the high response rates to chemotherapy reported in initial phase I studies, the fact is that no combination program has been shown to be superior to single agent chemotherapy and the high response rates tend to fall precipitously when efforts are made to reproduce them. (Is this the oncologic counterpart of the Heisenberg Uncertainty Principle?) Initial efforts to combine chemotherapy and immunotherapy have also followed a similar trajectory: An empirically derived combination program will show an impressive response rate (60-70%), the investigators will go through a hand-waving exercise explaining why a particular regimen was rational and, therefore, why it worked, then others will follow the recipe and get a 15% partial response rate associated with unacceptable toxicity.

This history makes it dangerous to highlight a phase II study in metastatic melanoma. However, a recent study from MD Anderson Cancer Center looks too good to omit. Patients with advanced, inoperable melanoma who had not received prior chemotherapy or biologic therapy were treated with a combination chemotherapy regimen called CVD (cisplatin 20 mg/m2/d for four days, vinblastine 1.6 mg/m2/d for 5 days, and dacarbazine 800 mg/m2 on day 1; cycles repeated every 21 days) integrated in various ways with a combination biotherapy regimen (interferon-a 5 MU/m2/d subcutaneously for five days and IL-2 9 MU/m2/d by continuous intravenous infusion for 5 days). In an initial group of 40 patients, these two regimens were integrated by alternating them in six-week intervals. A subsequent group of 62 patients was randomized to receive either the chemotherapy regimen followed immediately by the biotherapy regimen (CVD/bio) or the biotherapy regimen followed immediately by the chemotherapy (bio/CVD). Results were compared to an historical group who received CVD alone.

The alternating regimen produced two complete and 11 partial responses in 39 evaluable patients for an overall response rate of 33%. The sequential regimens produced 14 complete and 23 partial responses for an overall response rate of 60% (95% CI; 47-72%). The sequential regimen CVD/bio had a higher overall response rate (69%) than bio/CVD (50%) and a truly remarkable complete response rate (11/32, 34% for CVD/bio vs 3/30, 10% for bio/CVD; P = 0.03). The complete responses were also very durable; of the 14 complete responses to sequential therapy, eight remain in their first complete response for periods between 40+ and 52+ months. The median survival on sequential therapy was 13 months, but 24% of patients were alive at a median follow-up of nearly four years. This overall survival was significantly better than the survival of the historical control group treated with CVD alone. In every comparison (survival, response duration, time to progression), the CVD/bio group was significantly better than CVD alone.

Toxicity from the sequential regimens was severe: Nearly all patients had grade IV hematologic toxicity and 61% had grade IV thrombocytopenia. Two patients receiving CVD/bio died of septic shock. Other life-threatening complications included pulmonary edema, congestive heart failure, and gastrointestinal bleeding. Toxicities worsened each day of therapy.


It seems clear that patients who are fortunate enough to get a complete response to therapy have their lives meaningfully prolonged by the aggressive bio-chemotherapy program, CVD/bio. The severe acute toxicities resolve over a short period of time and a prolonged disease-free interval may be obtained. However, for the majority of patients (~2/3) who do not achieve a complete response, the benefit from the therapy may not justify the toxicity. Of course, the difficulty is that we are not able to discern who is likely to obtain a complete response before we subject people to the treatment.

Another serious issue is the lack of availability of dacarbazine. I don’t fully understand this problem, but, in general, I believe that ad hoc modifications to published regimens are to be eschewed. Omitting the most active single agent in the chemotherapy regimen seems particularly bad form. Thus, it will be difficult to give CVD/bio until the supply problem is fixed. Assuming dacarbazine becomes available in abundance, the issue will then be whether the patient wishes to endure myriad side effects, some of life-threatening proportions, in an effort to secure the benefits of complete response. I am reminded of the line from the Dirty Harry movie: "Do you feel lucky?" The question must be addressed on an individual basis.