Chemotherapy for metastatic melanoma

Time for a change?


  • Helen J. Gogas MD,

    Corresponding author
    1. First Department of Internal Medicine, University of Athens, Athens, Greece
    • First Department of Medicine, University of Athens, PO Box 14120, 115 10 Athens, Greece
    Search for more papers by this author
    • Fax: (011) 30 2107781517

  • John M. Kirkwood MD,

    1. Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
    2. Melanoma Program, University of Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania
    Search for more papers by this author
  • Vernon K. Sondak MD

    1. H. Lee Moffitt Cancer Center and Research Institute, University of South Florida, Tampa, Florida
    Search for more papers by this author


Melanoma is a neoplasm with a rising incidence. Early-stage melanoma is curable, but advanced, metastatic melanoma almost uniformly is fatal, and patients with such advanced disease have a short median survival. Systemic therapy remains unsatisfactory, inducing complete durable responses in a small minority of patients. For the current review, the authors focused on the current role of cytotoxic chemotherapy in the treatment of metastatic melanoma and the future prospects for improvements for multiagent chemotherapy and chemotherapy combined with immunomodulatory and/or molecularly targeted agents. They discuss roles of single-agent chemotherapy, combination chemotherapy, combinations of chemotherapy with immunomodulatory or hormone agents, biochemotherapy, and combination chemotherapy with targeted therapies. Cancer 2007;109:455–464. © 2007 American Cancer Society.

Patients with metastatic melanoma have a short median survival, generally in the range of only 6 to 9 months in most reported series. Systemic therapy induces complete durable responses in a very small minority of patients.1 Most discouragingly, the median survival has not improved appreciably despite decades of investigations of cytotoxic, immunologic, and now molecularly targeted agents. In this review, we have focused on the current role of cytotoxic chemotherapy in the treatment of metastatic melanoma and the future prospects for improvements for multiagent chemotherapy and chemotherapy combined with immunomodulatory and/or molecularly targeted agents.


Dacarbazine and Temozolomide

Single-agent chemotherapy produces objective response rates of <20%. However, a few patients—most of them with metastases to the skin, soft tissue, lymph nodes, or lungs and a good performance status as well as normal blood lactate dehydrogenase (LDH) enzyme levels (International Union Against Cancer/American Joint Committee on Cancer M1a and M1b)—achieve long-term disease control with a good quality of life, and some of them may achieve a cure.2 Historically, dacarbazine (DTIC) has produced response rates of from 15% to 25% in single-institution trials with median response durations of 5 to 6 months, but <5% of responses are complete.3 Long-term follow-up of patients treated with DTIC alone shows that <2% can be anticipated to survive for 6 years. However, in recent Phase III trials that used strict response assessment criteria, the response rates with single-agent DTIC did not exceed 12%.4–7 It has been demonstrated that single doses of DTIC (850–1000 mg/m2) are tolerated, appear to deliver clinical improvements similar to those observed with multiple doses that provide the same total dose per cycle, and should be the reference standard for randomized trials comparing new therapies with DTIC.8

DTIC remains the only currently used cytotoxic drug approved by the U.S. Food and Drug Administration (FDA) for the treatment of metastatic melanoma (hydroxyurea also was granted approval for the treatment of this disease but has never been employed widely). Despite its low single-agent activity, DTIC has remained the mainstay of many combination chemotherapy regimens; and evaluations of resistance-reversing agents, and despite >20 years of research, it is remains the standard against which most new chemotherapy agents are compared.

A closely related agent is temozolomide (TMZ), an oral congener of DTIC that has demonstrated efficacy in the treatment of variety of solid tumors, especially in brain malignancies, which is a manifestation of its far greater ability to penetrate the central nervous system (CNS). Taking into account the high rate of CNS recurrence as a site of failure after cytotoxic chemotherapy, TMZ may represent a viable alternative to DTIC, which is ineffective against CNS metastases, even if the single-agent efficacy in non-CNS disease is similar for both drugs.9 The single-agent activity of TMZ in metastatic melanoma has been established in several Phase I/II studies.10, 11 In a randomized trial of 305 patients with advanced melanoma, TMZ showed efficacy at least equivalent to that of DTIC in terms of objective response rate, time to progression, and overall and disease-free survival.5 in that study, the median survival was 7.7 months for patients who received TMZ and 6.4 months for patients who received DTIC. The median progression-free survival essentially was identical in the TMZ-treated group (1.9 months) and the DTIC-treated group (1.5 months). Overall, TMZ was tolerated very well and showed an advantage in terms of improvement in the quality of life. More patients showed improvement or maintenance of physical functioning at Week 12. That trial excluded patients who had brain metastases. Because the trial design was intended to demonstrate the superiority of TMZ over DTIC, rather than equivalence, the FDA did not accept the results of that trial as grounds for approving a melanoma indication for TMZ.

There are suggestions that the administration of TMZ in multiple doses per day or as a prolonged daily administration may overcome some chemotherapy drug-resistance mechanisms. The rationale for those suggestions lies in an evaluation of recent data on the mechanism of action of TMZ. Five-(3-methyltriazen-1-y) imidazole-4-carboximide, which is the active metabolite of DTIC and the end product of spontaneous metabolism of TMZ, methylates guanine residues in DNA at the O6 position.10 Increased levels of O6-alkylguanine-DNA alkyltransferase (ATase) are associated with resistance to agents that produce O6 methylation (DTIC, TMZ, and nitrosoureas). Administration of TMZ results in decreased ATase activity within 4 hours of an oral dose that persists in peripheral blood mononuclear cells for at least 24 hours.12 Daily administration resulted in the progressive depletion of ATase activity over the 5 days of treatment. Thus, on each day of treatment, there is increased sensitivity to the cytotoxic effects of TMZ resulting from the decrease in ATase activity induced by the prior day's chemotherapy. Currently, there is an ongoing trial by the European Organization for Research and Treatment of Cancer (EORTC) comparing DTIC with prolonged daily administration of TMZ. The extended dose of TMZ (75 mg/m2 per day) administered for 6 weeks followed by a 2-week rest is tolerated well and may be used in patients who have a borderline performance status. Other agents that lower ATase are Lomeguatrib (Patrin, PM), which is an orally bioavailable, highly potent O6-MeG analog that is tolerated well as a singe agent13 and poly (ADP-ribose) polymerase inhibitor.14

Other Single Agents

Cisplatin and carboplatin have shown modest activity as single agents in patients with metastatic melanoma. Cisplatin as single-agent therapy induced a 15% response rate with a short median duration of 3 months.1, 15 Evidence that the activity of cisplatin may be dose-dependent has come from single-institution studies. Doses up to 150 mg/m2 in combination with amifostine produced tumor responses in 53% of patients.16 However, all of those responses were partial, and the median response duration was only 4 months. In a randomized Phase II study of cisplatin and WR-2721 versus cisplatin in 94 patients, the response rates were 23.3% and 16.3%, respectively, and the median survival was 7.58 months.17 A response rate of 19% has been reported in 26 chemotherapy-naive patients with metastatic melanoma who received carboplatin. In those patients, there were 5 partial responses, and thrombocytopenia was the dose-limiting toxicity.18 Oxaliplatin has yet to be tested extensively in patients with melanoma. Although in vitro studies have suggested that it may be more active than cisplatin or carboplatin,19 a small Phase II trial in 10 patients who had received and failed prior chemotherapy produced no objective responses.20

The nitrosoureas (carmustine, lomustine [CCNU], and semustine [methyl-CCNU]) induce objective responses in from 13% to 18% of patients. They cross the blood-brain barrier; however, at conventional doses, little or no activity was observed against melanoma brain metastases.1, 15, 21 Furthermore, the nitrosureas induce prolonged myelosuppression. Despite this, they have been included frequently in multiagent chemotherapy regimens, presumably for their ability to penetrate into the CNS. Fotemustine is probably the most active nitrosourea in metastatic melanoma. It has been tested in 5 Phase II trials on 351 patients with response rates of 20% to 25% and complete response rates of 5% to 8%.22–26 Fotemustine was the first drug to show significant efficacy in brain metastases, but it never has been widely available in the United States.

The vinca alkaloids, particularly vindesine and vinblastine, have produced responses in approximately 14% of patients,27 and the taxanes have produced responses in from 16% to 17% of patients.28–30 None of those drugs have been evaluated as single agents in Phase III trials; however, based on the experience with DTIC, it is likely that the Phase III trial objective response rates for all of them would be less than the rates reported from the various small Phase II trials that have been conducted to date. All of these drugs are rarely used currently as single-agent therapy in metastatic melanoma, but they frequently have been incorporated into combination chemotherapy and biochemotherapy regimens. Finally, a potentially active new drug is ABI-007 (abraxane), which is an albumin-bound nanoparticle formulation of paclitaxel with an improved therapeutic index. It has been tested in a Phase II trial in 37 previously treated and chemotherapy-naive patients with metastatic melanoma and produced a >30% overall response rate. ABI-002 will now be explored in a randomized Phase III trial.31


Because single-agent chemotherapy regimens have demonstrated only modest activity in the treatment of metastatic melanoma, in an attempt to improve response rates, many combination regimens have been evaluated in clinical trials. Although it is legitimate to question the advantages that may accrue from the use of multiple agents with minimal single-agent activity and no laboratory or clinical evidence of synergistic interactions, the lack of viable alternatives drove the conduct of these trials more than any biologic rationale. Initially, 2-agent regimens were used in which DTIC was combined with a nitrosourea, vinka alkaloid, or platinum compound. In most of those trials, response rates from 10% to 20% were observed, and there was little evidence to suggest superiority to single-agent DTIC.32–36

After those disappointing results, investigations of more aggressive, multiagent regimens were conducted. Phase II studies of 3- and 4-drug combinations generally produced response rates that ranged from 30% to 50% in single-institution studies. Two of the most widely studied combinations are cisplatin, vinblastine, and DTIC (CVD) and the Dartmouth regimen. The 3-drug combination CVD that was developed by Legha et al at The University of Texas M. D. Anderson Cancer Center produced responses in 40% of 50 patients, a 4% complete response rate, and a median response duration of 9 months.37 However, in a randomized trial comparing CVD with single-agent DTIC that involved approximately 150 patients, the CVD arm produced a 19% response rate compared with 14% for the DTIC arm, and there was no differences in either response duration or survival.38

The 4-drug combination of cisplatin, DTIC, carmustine, and tamoxifen (CBDT), also called the Dartmouth regimen, produced responses in 46% of 141 patients (16 complete responses and 49 partial responses). The median response duration was 7 months.39 The authors suggested that the inclusion of tamoxifen was essential, with response rates of 10% for the 3 cytotoxic drugs in the same doses when tamoxifen was omitted.40 However, a randomized Phase III trial conducted by the National Cancer Institute of Canada that compared CBDT with cisplatin, DTIC, and carmustine (CBD) produced a response rate of 30% for the CDBT arm compared with 21% for the CBD arm (P = .187). In fact, twice as many patients achieved a complete remission in the CBD group compared with the group that received tamoxifen (6% vs 3%), although the difference was not statistically significant (P = .33).

However, in that study, more patients in the tamoxifen group achieved a partial remission (27% vs 14%). This difference was of borderline significance (P = .05). Gender did not seem to be a factor: The response rate and the median survival for women in the tamoxifen group did not differ statistically compared from those for women in the chemotherapy-alone group. Similarly, there was no difference between men in the treatment groups. Progression-free survival and overall survival did not differ significantly between the 2 arms (P = .86 and P = .52, respectively).41 In another randomized Phase III trial, the CDBT combination was compared with single-agent DTIC. That cooperative group trial involved 240 patients, and the response rate was 10.2% for the DTIC regimen compared with 18.5% for the CDBT regimen (P = .09). The median survival was 7 months, with no significant difference between the 2 treatment arms. Toxicity was substantially greater for the combination regimen, with bone marrow suppression, nausea, emesis, and fatigue significantly more frequent with CDBT than with DTIC.42 Taken together, to date, controlled trials have produced no compelling evidence to support the value of combination chemotherapy, with or without tamoxifen, in patients with metastatic melanoma. It is difficult to justify the use of either CVD or CBDT instead of single-agent DTIC or TMZ for the treatment of most patients with metastatic melanoma.


Tamoxifen and Interferon α

Similar to investigations of multiagent cytotoxic chemotherapy, combinations of cytotoxic drugs that have minimal efficacy with immunomodulatory or hormonal agents have been investigated despite the lack of a compelling rationale. The results of adding tamoxifen to CBD chemotherapy were discussed above; however, tamoxifen also has been added to single-agent DTIC in several studies. Results from a small, randomized trial of DTIC without or with tamoxifen indicated that combination therapy may be more effective.43 A response rate of 28% and a median survival of 41 weeks were reported for patients who received DTIC plus tamoxifen compared with a response rate of only 12% and a median survival of 23 weeks for patients who received DTIC alone.

Similar results were reported with the combination of DTIC plus interferon α (IFN-α). In a small, randomized trial that compared DTIC without or with IFN-α2b, the combination therapy produced 12 complete responses and 4 partial responses in 30 patients compared with only 2 complete responses and 4 partial responses among 30 patients who received DTIC.44 The median response and survival were prolonged significantly for the DTIC plus IFN-α2b arm in that trial. To further evaluate the potential benefits of combining DTIC with either tamoxifen, or IFN-α, or both, the Eastern Cooperative Oncology Group (ECOG) conducted a 4-arm, 2 × 2, factorial design, Phase III trial that failed to confirm the initial encouraging observations.4 The overall response rate in that trial (ECOG 3690) was 18% (range, 12–21% for the 4 arms), and the median time to treatment failure was 2.6 months. The median survival was identical (9.1 months) for all 4 arms tested. In that trial, there was no advantage in terms of response or survival attributable to the addition of IFN-α2b, tamoxifen, or both agents to DTIC. In a recent meta-analysis of 6 published, randomized trials that combined a total of 912 patients who were randomized to receive chemotherapy or biochemotherapy with or without tamoxifen, no improvements in the rates of overall response, complete response, or survival were demonstrated.45

Although results from early, limited, Phase II and III trials suggested a benefit of tamoxifen, IFN-α2b, or cisplatin when added to DTIC for patients with metastatic disease, those benefits have not been confirmed in large-scale, multicenter, Phase III trials (Table 1). The main reason for such discrepancies between the findings from single-institution studies and the findings from large, multicenter, cooperative trials probably is selection bias. Differences in performance status, percentages of patients with visceral involvement, and number of metastatic sites easily could account for some of the observed differences; in fact, all of those factors are known to have an impact on both response rate and survival.46

Table 1. Randomized Studies Evaluating Dacarbazine Versus Combination Chemotherapy
ReferenceRegimen(s)No. of patientsRR, %PFS, moSurvival, months
CombinationSingle agentCombinationSingle agent
  1. RR indicates response rate; PFS, progression-free survival; DTIC, dacarbazine; NA, not available; IFN, interferon; CVD, cisplatin, vinblastine and DTIC; CBDT, cisplatin, DTIC, carmustine, and tamoxifen.

Cocconi et al., 199243DTIC and tamoxifen1172812NANA48 vs 29
Falkson et al., 199144DTIC and IFN645320   
Falkson et al., 19983DTIC and IFN;DTIC and tamoxifen; DTIC, tamoxifen, and IFN28021152.62.68.9
Buzaid et al., 199338CVD1501914   
Chapman et al., 199942CBDT24018.510.2NANA7


It has been reported that the combination of a low-dose, daily schedule of TMZ with thalidomide has significant clinical activity in patients with metastatic melanoma.47 The overall objective response rate of 32% was higher than that achieved with TMZ alone, and further study is warranted. However, the legacy of failure left by prior attempts to combine cytotoxic agents with immunodulators and other drugs in the absence of a compelling biologic rationale suggests that the combination of TMZ plus thalidomide should not be accepted as a standard therapy until and unless demonstrates superiority in Phase III trials or at least until the high response rate can be confirmed in larger, multi-institutional Phase II trials using strict response assessment criteria. In addition, it should be noted that, in a Phase II study of TMZ, thalidomide, and whole-brain radiation therapy in patients with brain metastases from melanoma that was conducted by the Cytokine Working Group in 40 patients, no patient exhibited a systemic response.48


In recent years, trials have evaluated the role of combination chemotherapy with other agents, such as IFN-α and interleukin 2 (IL-2). Many investigators have combined IL-2 with both IFN-α and chemotherapy in an attempt to improve both the response rate and the percentage of durable complete remissions. A large body of data now exists demonstrating that such biochemotherapy combinations can produce response rates in the range from 40% to 60% with an approximately 10% complete response rate.49–51 Durable remissions exceeding 5 years were achieved by approximately 5% to 10% of patients. Recurrences beyond the 2-year time point were uncommon, thus suggesting that these patients who exhibited durable responses may have achieved a cure.51, 52 However, many of the most active biochemotherapy regimens are associated with moderate to severe toxicity that has limited their broader use and acceptance. Many trials have been performed in an attempt to identify biochemotherapy regimens that may be administered in an outpatient setting, with less toxicity, and using lower doses of intravenous or subcutaneous IL-2.52, 53 Unfortunately, biochemotherapy regimens that involved low-dose, subcutaneous IL-2 administration appeared to produce lower response rates than were observed generally with regimens that involved intravenous IL-2.53

Two meta-analyses of patients with metastatic melanoma who were treated in studies that used various combinations of chemotherapy, biotherapy, or biochemotherapy demonstrated an improvement in response rates, but not in survival, with the use of biochemotherapy. In a report by Keilholz et al,54 patients with metastatic melanoma who were treated with IL-2/IFN-α/chemotherapy, IL-2/IFN-α without chemotherapy, and IL-2/chemotherapy without IFN-α regimens had response rates of 45%, 21%, and 15%, respectively. However, in that trial, survival did not differ significantly between the groups (10.5 months), with 20% and 10% survival rates at 2 years and 5 years, respectively, that did not differ among the groups. In a second meta-analysis of combined studies that involved 7711 patients with advanced melanoma from 168 published trials, treatment with an IL-2/IFN-α/chemotherapy combination resulted in a significantly improved response rate compared with treatment that used chemotherapy or IL-2-based biotherapy.55

Although the results of those meta-analyses speak for themselves, in the evidence-based treatment of malignant melanoma, the superiority of biochemotherapy regimens to either IL-2-based immunotherapy or cytotoxic chemotherapy should be derived from Phase III clinical trials; thus, 7 prospective, randomized studies in the United States and Europe evaluated aggressive biochemotherapeutic regimens that contained IL-2 and IFN-α (Table 2). Those results were inconsistent and were not compelling. However, it is worthwhile describing several of the randomized trials in more detail.

Table 2. Randomized Studies Evaluating Chemotherapy Versus Biochemotherapy
ReferenceRegimenNo. of patientsRR, %PFS, moSurvival, mo
  • RR indicates response rate; PFS, progression-free survival; Biochemo, biochemotherapy; Chemo/Bio, chemotherapy or immunotherapy; DTIC, dacarbazine; IFN, interferon; ±, with or without; CDDP, cisplatin; IL2, interleukin 2; NR, not reported; BCNU, carmustine; CVD, CDDP, vinblastine and DTIC.

  • *

    Time to progression in days.

  • Median duration of response.

Kielholz et al 199756*IFN and IL2 ± CDDP138331892539
Kielholz et al, 200557DTIC, CDDP, and IFN ± IL236320.822.83.03.99
Rosenberg et al, 199958DTIC, tamoxifen, and CDDP ± IL2 and IFN1024427NRNR10.7 vs 15.8
Ridolfi et al, 200278CDDP, DTIC, and BCNU ± IL2 and IFN17625.320.24311 vs 9.5
Dorval et al, 199979CDDP and IL2 ± IFN10124169.16.610.9 vs 10.4
Eton et al, 200259CVD ± IL2 and IFN18348254.92.411.9 vs 9.2
Atkins et al, 200360CVD ± IL2 and IFN48216.611.953.18.4 vs 9.1

In a prospective European trial, 138 patients with metastatic melanoma were randomized to receive IFN and decreasing doses of IL-2 with or without cisplatin.56 The results of that trial demonstrated a significant increase in the response rate from 18% to 33% in the group that received biochemotherapy compared to the group that received biotherapy and an increase in progression-free survival from 53 days to 92 days without any statistical differences in terms of survival.

A second trial that was conducted by the EORTC evaluated cisplatin, DTIC, and IFN-α with or without IL-2.57 Evaluation of 363 treated patients demonstrated no statistical improvement in response rate with the addition of IL-2 (22.8% vs 20.8%, respectively) and in progression-free survival (median, 3.0 months vs 3.9 months, respectively). The median survival was 9 months in both arms, and the 2-year survival rate was 12.9% and 17.6%, respectively (P = .32). Another study that was conducted by Rosenberg and colleagues58 at the National Cancer Institute-Surgery Branch randomized 102 patients with stage IV melanoma to receive either chemotherapy (DTIC, cisplatin, and tamoxifen) or biochemotherapy (IL-2, IFN-α, DTIC, cisplatin, and tamoxifen). Although the response rate in the biochemotherapy arm (44%) was almost twice that of the chemotherapy arm (27%), the difference was not statistically significant (P = .07). Furthermore, there was a survival advantage in the chemotherapy arm compared with the biochemotherapy arm (median survival, 5.8 months in the biochemotherapy arm vs 10.7 months in the chemotherapy arm; P = .05). There has been speculation that this survival finding may have been caused by the administration of high-dose IL-2 as salvage therapy to a significant proportion of the patients who failed to respond to chemotherapy alone.

The study by Eton et al (CVD vs CVD with intravenous IL-2 and subcutaneous IFN-α) in 183 evaluable patients was the only one that demonstrated a statistically significant advantage of biochemotherapy over chemotherapy alone in terms of the response rate (48% vs 25%, respectively), the complete response rate (7% vs 2%, respectively), and the median time to recurrence (4.9 months vs 2.4 months). A modest but statistically significant increase in median overall survival (11.9 months vs 9.2 months, respectively) was observed.59

A large, Phase III trial that involved 482 patients compared CVD chemotherapy with CVD plus intravenous IL-2 and subcutaneous IFN-α (a modified version of The University of Texas M. D. Anderson Cancer Center's concurrent biochemotherapy regimen; ECOG trial E3695). Important aspects of the E3695 protocol were that it was large enough to distinguish clinically meaningful differences in survival and durable responses and that it involved a population with a relatively large percentage of patients who had prior IFN exposure in the adjuvant setting. There were no statistically significant differences between the chemotherapy and biochemotherapy arms in response rates, progression-free survival, or overall survival. It should be noted that there were many inevaluable patients, many patients were not treated according to the protocol, and more patients were randomized to the biochemotherapy arm.60 Nonetheless, the conclusion from that and all randomized trials of biochemotherapy performed to date was inescapable—that biochemotherapy should not be used routinely outside of a clinical trial. However, biochemotherapy may be useful in a symptomatic patient with rapidly progressive disease to provide symptomatic relief if high-dose IL-2 (FDA approved) is not an option.


The intrinsic resistance of melanoma to conventional chemotherapy has led investigators to evaluate new approaches such, as protein kinase inhibitors (eg, sorafenib), agents that act on cytotoxic T-lymphocyte antigen 4 (CTLA-4) or on apoptotic mechanisms (eg, oblimersen sodium; previously referred to as G3139), and antiangiogenic agents (eg, bevacizumab, SU5416, MEDI-522, PI-88).

Sorafenib plus Chemotherapy

Sorafenib (Nexavar, previously referred to as BAY 43-9006) targets the adenosine triphosphate-binding site of the BRAF kinase and inhibits both wild-type and mutant BRAF in vitro. In addition, the spectrum of kinases inhibited by sorafenib includes CRAF, vascular endothelial growth factor receptor 2, platelet-derived growth factor receptor β, p38, flt-3, and c-kit.61 Preclinical studies demonstrated a significant retardation in the growth of human melanoma tumor xenografts with sorafenib.62 In a Phase I study, the maximum tolerated dose of sorafenib as a single agent was established at 400 mg twice daily, and the most common toxicities were gastrointestinal (mainly diarrhea), dermatologic (skin rash, hand-foot syndrome), and fatigue.63 In a Phase II trial that was conducted in 20 patients with refractory metastatic melanoma, sorafenib showed modest activity with 1 partial response and 3 patients who achieved stable disease.64 In another Phase II, randomized, discontinuation trial, no objective responses were achieved, and 19% of patients achieved stable disease.65 However, in a Phase I/II study that combined carboplatin and paclitaxel with escalating doses of sorafenib in 35 patients, a promising response rate of 31% was observed, and another 54% of patients experienced stable disease that lasted ≥3 months.66 That study recently was updated to include 105 patients, and the current response rate is 27%. It is noteworthy that responses to sorafenib have not been correlated to date with BRAF mutation status.65 Responses were observed in previously treated patients and in at least 1 patient with a noncutaneous primary melanoma. A second Phase II trial of the same regimen conducted at another institution (Vanderbilt University) yielded similarly promising results (unpublished results). On this basis, 2 Phase III trials have been launched to assess the efficacy of carboplatin and paclitaxel plus sorafenib versus placebo in chemotherapy-naive patients (E2603) and in previously treated patients. Relatively little is known about the inherent activity of carboplatin and paclitaxel as systemic therapy for patients with stage IV melanoma, particularly at the doses employed in the Phase II trial and, now, in the Phase III trial. In report on previously treated patients, no responses were observed among 19 patients who received the 2-drug combination,67 whereas a second trial in chemotherapy-naive patients resulted in 3 partial responders and 7 patients with stable disease among 15 evaluable patients.68 Thus, it is likely that the combination of carboplatin and paclitaxel, by itself, has inherent activity in metastatic melanoma. The results of the Phase III trials will define the real impact of adding sorafenib to this regimen in patients with metastatic melanoma. A randomized Phase II trial of 2 schedules of TMZ plus sorafenib also is underway, and the preliminary results have been considered encouraging.69

Anti-CTLA-4 Antibodies and Chemotherapy

Two human anti-CTLA-4 monoclonal antibodies have been tested in clinical trials: MDX-010 and CP-675,206. Responses have been observed with both antibodies administered as single-agent therapy in patients with metastatic melanoma,70, 71 providing a rationale for combinations with chemotherapy. In a Phase II study, the activity of MDX-010 alone or in combination with DTIC was assessed. There were 2 partial responses in the antibody alone arm and 1 complete response and 4 partial responses in the antibody plus DTIC arm, suggesting more activity for the combination.72 In the long term follow-up of that study, 1 additional complete response was observed in the combination arm, and durable clinical responses were noted.73 These results are intriguing, but it remains unclear whether the activity of the combination is simply additive or truly synergistic, and further study probably is warranted in preclinical models.

Anti-BCL2 Antisense Oligonucleotide

Oblimersen sodium (anti-BCL2 antisense oligonucleotide) originally was tested in a Phase I/II trial in combination with DTIC that was followed by a randomized Phase III trial in 771 patients. The primary endpoint of that trial was overall survival, which was not statistically significantly different between the 2 arms (9.1 months for the combination arm vs 7.9 months for DTIC alone arm), although overall and complete response rates were significantly better for the combination arm (overall response, 11.7% vs 6.8%, respectively; P = .019) and progression-free survival was improved with the combination arm (74 days vs 49 days; P = .0003). In an updated analysis, for the subgroup of patients with LDH values ≤2 times the institutional upper limit of normal, there was a statistically significant survival benefit for combination therapy (10.2 months vs 8.7 months; P = .02).7 These data support the idea that oblimersen has at least modest activity when it is combined with DTIC and justify further studies of this compound and similar strategies to overcome drug resistance in melanoma, as discussed by Eggermont.74


MEDI-522 is a humanized monoclonal antibody directed against the αVβ3 integrin. αVβ3 is essential for endothelial cell proliferation, maturation, and survival; and, when it is blocked, proliferating endothelial cells undergo apoptosis and regress. In addition, αVβ3 is highly expressed in melanomas and is associated with tumor growth.76 In preclinical studies using αVβ3 antagonists, inhibition of melanoma tumor growth independent of its antiangiogenic effects was reported.76 MEDI-522 has been investigated in 3 Phase I, dose-escalation studies in patients with refractory cancer. In the Phase II trial, 57 patients received MEDI-522 alone, and 55 patients received MEDI-522 plus DTIC. MEDI-522 with or without DTIC generally was well tolerated and was active in patients with metastatic melanoma. The median survival was 12.6 months for the group that received MEDI-522 with DTIC and 9.4 months for the group that received MEDI-522 without DTIC.77 Those results provided the impetus to evaluate single-agent MEDI-522 in a confirmatory Phase III trial in patients with melanoma.


A discouragingly long list of targeted agents already have been tested in patients with melanoma and have failed. This failure may attributed to the empiricism of trials that generally combined agents because they were there (CVD, IL-2, and IFN) and a general lack of studies that evaluated the posited mechanisms of action for the agents used in combination, such as antisense to BCL2 or combinations of antiestrogens with chemotherapy, to name a few—in which the pursuit of tissue studies to validate the hypothetical intervention likely would have been helpful critically in reformulating the approaches before huge trials were undertaken. We are entering a new era for the systemic therapy of melanoma in which the molecular events that drive tumor progression, on the one hand, and the suppression of the host immune response to tumor, on the other, increasingly are understood. With this understanding, the pursuit of trials that target the events critical to progression and the lesions with immunity, which may be associated with progression, increasingly are feasible. It is incumbent upon investigators to design clinical trials that can inform us why we succeed when we do and, perhaps equally important, why we fail when we fail at the levels of the tumor and the immune response. We have a pressing need for more effective single agents, which, either alone or in combination with agents that may abrogate resistance and agents that may impact collateral pathways in a rational manner, will achieve more a significant impact on this disease.