Treatment Options for Patients with Stage III & IV Melanoma: Part II: Experimental treatments for stage III/IV or refractory/recurrent melanoma

Contemporary Oncology®, Spring 2010, Volume 2, Issue 1

Various immunotherapeutic strategies to treat advanced melanoma have been under investigation for more than a decade


Various immunotherapeutic strategies to treat advanced melanoma have been under investigation for more than a decade—primarily vaccines and adoptive cell transfer (ACT) therapy. Most of these novel approaches appear relatively safe, but with differing levels of efficacy. Researchers continue to believe that some form of immunotherapy, either alone or combined with other therapeutics, may prove to be effective in late-stage melanoma. Data are accumulating in support of novel therapeutic immunotherapy regimens and schedules; these data are often incongruent with the dogma of the past few decades on how the host immune response becomes activated to cancer.

Tumor Cell—Based Vaccines

Several early-phase clinical trials have shown that autologous and allogeneic tumor cell—based vaccines can be given safely with few adverse effects. One of the first such vaccines tested extensively in clinical trials was onamelatucel-L, a polyvalent whole–tumor cell vaccine derived from 3 human melanoma cell lines. Although early trials failed to show significant clinical benefit, a few complete and durable responses provided the impetus for two multicenter, phase III randomized trials in 1998.1 As part of the design of the onamelatucel-L trials, patients with stage IV melanoma were required to undergo definitive surgical removal of all metastatic disease prior to enrollment. Both trials closed prematurely, after an interim analysis revealed no probable efficacy over placebo.1 Evidence supports a hypothesis of first removing the bulk of disease, and long-term responses have indeed been observed in patients who had complete surgical resection of regional nodal disease without any adjuvant immunotherapy.2,3

Over the years, several other tumor cell—based trials4-16 have yielded interesting clinical outcomes. Overall, however, the application of strict response criteria finds that response rates remain exceedingly low. The next generation of whole—tumor cell vaccines has incorporated advances in gene transfer technology with the immunobiology of the cancer cell and host immune system. Initial preclinical studies have shown that a melanoma whole–tumor cell vaccine transduced with the granulocyte-macrophage colony-stimulating factor (GM-CSF) gene confers a potent and long-lasting anti-tumor immune response.17 The secretion of GM-CSF by the transduced melanoma cells attract immune cells, such as antigen-presenting cells and T-lymphocytes, to the vaccine site.17,18 Several phase I/II clinical trials using this approach have been completed, with a few objective clinical responses noted.19-21 These early studies, however, involve small numbers of patients. Further clinical trials that enroll more patients are needed before strong conclusions can be made about the efficacy of this vaccine approach.

Peptide- and Dendritic Cell-Based Vaccines

Rosenberg and associates performed one of the first clinical trials to use peptides derived from melanoma cells.22 The researchers used a modified immune-dominant peptide of the gp100 antigen, g209-2M, to vaccinate 11 patients with stage IV melanoma. They found that 10 (91%) of the patients showed a consistently high level of immunization against the native g209-217 peptide, but not against g280-288, the control peptide. Although many patients developed an immunologic response to the peptide vaccine, all patients went on to develop progressive disease. This study, however, provided a central proof-of-principle that patients with advanced melanoma could mount an immune response against their own self-antigens, an exciting concept not appreciated previously. The trial also highlighted that an immunologic response does not necessarily equate with a clinical response.

Although the vaccines are seemingly effective at enhancing a tumor-antigen—specific immune response, clinical outcomes have been uniformly disappointing. A recent analysis of 28 different peptide-based vaccines in stage IV patients revealed an objective response rate (ORR) of just 2.9%.23 Other studies using a peptide-based vaccination strategy have produced only limited responses.24,25

Another novel approach to melanoma immunotherapy involves using autologous dendritic cells (DCs) as potent antigen-presenting cells that interact with other immune cells and activate an antigen-specific immune response. The first published clinical trial of DC vaccination was in 1995; this has been followed by 98 additional clinical trials in 15 different countries involving >1000 patients.26 Of these 98 trials, 28 focused on patients with advanced stages of melanoma. The trials have shown that DC-based vaccines are safe, producing few severe adverse effects. Yet, their record of effectiveness has been disappointing, with an ORR of <5% in most trials. A recent DC-based vaccination trial by Schadendorf and colleagues from the German Dendritic Cell Study Group produced similar results.27

Plasmid and Recombinant DNA Vaccines

One approach to immunotherapy is to use viral vectors as carriers of cDNA encoding for the gene of interest. The advantage of using recombinant viruses to carry gene encoding for an antigen is that viruses efficiently infect cells, particularly DC.28 Although the use of viral vectors is appealing for multiple reasons, several induce a tremendous antiviral-neutralizing antibody response to first and subsequent vaccinations, severely limiting the effectiveness of this approach. One exception may be the use of fowlpox viral vectors, which appear not to produce neutralizing antibodies. Other studies have used adenovirus, vaccinia, canarypox, or fowlpox viruses, encoding various melanoma antigens, such as gp100, MART-1, tyrosine, or MAGE. Only occasional immunological or clinical responses have been observed.29-31

In a recent study, recombinant vaccinia and fowlpox virus encoding the antigen NY-ESO-1, used in a prime-boost vaccination setting, induced humoral and/or cellular responses in a majority of patients.31 Kaufman and colleagues have examined several combinations of a recombinant vaccinia virus in patients with unresectable melanoma,32-34 evaluating the vaccines’ toxicity, therapeutic efficacy, and ability to generate melanoma-specific immunity. Despite using a plethora of viral vectors that encoded various melanoma-associated antigens and co-stimulatory molecules, results have been disappointing. Rosenberg and associatesdescribed using various viral vaccines to treat 160 patients, with only 2 demonstrating any clinical response.22

Together, the various viral vaccination strategies have an ORR of only 1.9%. A recent phase I trial showed that intra-tumoral plasmid IL-12 gene electroporation was highly efficient at delivering plasmid-DNA across 7 dosing groups. IL-12 protein levels increased up to 18-fold above mean baseline in patients with stages IIIB/C and IV melanoma.35 Daud and colleagues used electroporation with a plasmid vector containing the IL-12 gene to treat accessible metastatic lesions in 24 patients. Of the evaluable patients, 10 (53%) exhibited stable disease (SD) or distant-tumor regression; 4 of the 19 patients (21%) with distant lesions showed distant clinical response. A total of 3 patients had complete remission (CR) lasting beyond study follow-up, and 2 of these received no subsequent systemic therapy; the third received dacarbazine after plasmid IL-12 treatment but prior to exhibiting complete tumor regression. In addition, 6 patients exhibited SD at distant sites that lasted 4 to 20 months.

Neoadjuvant Vaccination

Grinshtein and associates recently examined the effectiveness of neoadjuvant vaccination in a murine melanoma model at improving tumor-free survival following surgery.36 They found that vaccinating mice in the neoadjuvant setting was far superior to vaccination in the adjuvant setting, providing nearly 100% protection against relapse. Further, neoadjuvant vaccination appeared to increase the overall number of antigen-specific T-cells and tumor-infiltrating lymphocytes within the tumor.

Adoptive Cell Transfer Therapy

ACT therapy for melanoma patients initially involves isolating highly active, melanoma-antigen—specific autologous T-cells from a patient and growing large numbers of the cells in vitro. The reactive T-cells are then infused back into the patient. ACT therapy is usually administered along with high-dose IL-2, which appears to be a promising approach for advanced melanoma.37 Initial studies using cloned melanoma-antigen—specific T-cells, with or without IL-2, failed to observe an objective anti-tumor response.38,39 Another method uses tumor-infiltrating lymphocytes (TIL), which has the hypothetical advantage of containing a diverse effector population comprised of CD4 and CD8 T-cells. In clinical studies using TIL in conjunction with high-dose IL-2, 33% of patients experienced objective tumor regression. Unfortunately, these clinical responses were transient, with the transferred TIL cells failing to persist.40,41

To enhance the efficacy of ACT therapy using TILs, a lymphodepleting, non-myeloablative preconditioning chemotherapy regimen consisting of cyclophosphamide and fludarabine was administered prior to the highly reactive melanoma-antigen—specific TIL cells in combination with high-dose IL-2.42 The initial study involved 13 patients whose melanoma had progressed despite multiple treatments, including high-dose IL-2. Objective clinical response was observed in 6 patients, and 4 demonstrated mixed response, with some lesions regressing and others progressing.42 In a recent follow-up study, cancer regression in patients with refractory metastatic melanoma with large, vascularized tumors was noted in a striking 18 of 35 patients (51% response rate); 4 patients had complete regression of all metastatic disease.43 It is clear that highly reactive cytotoxic T-cells can mediate the destruction of large, bulky tumors and render stage IV melanoma patients disease-free, albeit rarely.

Dudley and colleagues recently reported on two different sequencing trials in patients with stage IV melanoma that involved administering non-myeloablative chemotherapy with whole body radiotherapy. They used the previously described lymphodepleting preparative regimen of cyclophosphamide and fludarabine plus either 2 Gy or 12 Gy of total body irradiation (TBI) prior to ACT therapy.44 While non-myeloablative chemotherapy alone was associated with an ORR of 49%, the addition of 2 Gy of TBI increased the ORR to 52%, and ORR increased to 72% with the use of 12 Gy of TBI. This study yields the highest ORRs to date for the immunotherapeutic treatment of patients with stage IV melanoma.


Targeted therapy for patients with metastatic melanoma has expanded rapidly over the past decade, with several novel agents currently being examined in early-phase clinical trials. Regulatory T-cells seem to play an important role in host immune tolerance, autoimmunity, and suppression of the immune response to cancer.45-52 The use of targeted therapeutics against melanoma remains an active area of research, and we continue to learn about these agents through important clinical trials.

Denileukin Diftitox (Ontak)

The development of denileukin diftitox, an immunotoxin specific for the IL-2 receptor, has become standard therapy for patients with cutaneous T-cell lymphoma, producing an ORR of 30% to 40%.53 Recently, Dannull and associates showed that denileukin diftitox induced elimination of regulatory T-cells (Tregs), and, followed by vaccination with RNA-transfected DC, significantly improved the stimulation of tumor-specific T-cell responses in renal cell carcinoma patients.54 In contrast, a study by Attia and colleagues concluded that denileukin diftitox does not eliminate the absolute number of Tregs or decrease their suppressive activity.55 Another more recent report showed that the transient depletion of T-cells with denileukin diftitox contributed to significant tumor regression and partial response (PR) in 5 (31%) out of 16 patients.56

Sorafenib (Nexavar)

Numerous studies have substantiated the central role of cellular signaling through the RAF/MEK/ERK/MAPK pathway and its involvement in melanocytic tumorigenesis. Sorafenib is an orally administered multikinase inhibitor that targets not only BRAF serine/threonine kinases, which are frequently mutated in melanoma, but also several tyrosine kinase receptors associated with tumor angiogenesis and progression. Despite strong pre-clinical data, a recent randomized phase II trial found that sorafenib monotherapy demonstrated little or no anti-tumor activity in advanced-melanoma patients.57 There appeared to be no correlation between modest anti-tumor activity and the presence of an oncogenic V600E BRAF mutation within exon 15.

Other studies investigated sorafenib along with chemotherapeutic agents that had previously demonstrated some activity against melanoma, such as carboplatin and paclitaxel, in patients with advanced melanoma. This combination induced 1 CR (<1%), 27 PRs (26%), and 61 SDs (58%).58 In a phase I study of patients with advanced melanoma, sorafenib combined with dacarbazine induced PR in 3 of 10 evaluable patients and SD in 5 patients.59 Further investigations are underway that combine sorafenib with other cytotoxic agents.

Antisense Oligonucleotides

Antisense oligonucleotides are short sequences of cDNA engineered to bind to specific gene sequences and prevent the translation of mRNA, with resultant abrogation of protein synthesis. It has been shown that one mechanism of melanoma cell survival depends on overexpression of the BCL2 gene, which produces an anti-apoptotic protein that blocks the release of cytochrome C. Down-regulation of the BCL2 protein results in a higher level of chemotherapy-induced apoptosis in human cancer xenografts.60

A recent randomized phase III trial accruing 771 patients compared dacarbazine alone to dacarbazine plus oblimersen (Genasense), a BCL2 antisense oligonucleotide.61 At 24 months of follow-up, a trend toward improved overall survival (OS) was observed for the combination arm compared with the dacarbazine monotherapy arm (Table 1). Dacarbazine plus oblimersen produced significant improvement in median progression-free survival, and rates of overall response, CR, and durable response.

Interestingly, the 508 patients in the study with normal levels of lactate dehydrogenase (LDH) demonstrated significantly improved OS with the combined therapy. This indicates that LDH is an effective pretreatment biomarker and confirms an overall poor prognosis for patients with elevated LDH levels prior to treatment. The authors also suggest obtaining serum LDH levels to identify patients that are unlikely to benefit from oblimersen-dacarbazine treatment. Although the survival differences are marginal, they show that oblimersen demonstrates significant clinical activity in melanoma. To maximize the likelihood that oblimersen and similar agents will show statistically significant activity or improvements in survival, future trials with these drugs should probably be limited to patients with normal LDH levels.

Anti-CTLA-4 Antibodies

Another novel approach to the treatment of patients with metastatic melanoma involves using different antibodies to reduce the sensitivity of activated T-cells to negative regulatory signals. One mechanism of immune regulation is via CTLA-4 protein expression on activated T-cells, which transduces signals that inhibit T-cell activation and proliferation, as well as IL-2 production, back to the T-cell. It has been shown that the CTLA-4 receptor is present on Tregs, though recent data is contradictory as to whether CTLA-4 blockade leads to the depletion or impairment of Tregs.

CP-675206 (tremelimumab). CP-675206 is an IgG2 antibody with high affinity for CTLA-4 on the surface of T-cells. Ribas and colleagues recently conducted a phase I clinical trial involving 34 patients with stage IV melanoma.62 Patients received a single dose of CP-675206;4 (12%) had objective responses, 2 showed CR, 2 had PR, and 4 experienced SD. Autoimmune side effects were considerable and included vitiligo, dermatitis, thyroiditis, hypophysitis, and colitis.

More recently, a phase I/II trial investigated two dosing schedules of CP-675206 in patients with stage IV melanoma.63 One group (n = 43) received 10 mg/kg per month, and the other group (n = 41) received 15 mg/kg per month. ORR was similar in both groups, at 9.3% in the low-dose group and 9.8% in the high-dose arm (Table 2). Each arm produced 1 CR and 3 PRs. Median survival was 9.97 months in the low-dose arm and 11.53 months in the high-dose arm, and 12-month survival rates were 32% and 46%, respectively. The most common adverse effects were minor to moderate diarrhea, rash, and pruritis. The relative ease of administration and the reduced frequency of adverse effects for patients in the 15 mg/kg arm led investigators to recommend the regimen for future studies.

MDX-010 (ipilimumab). Another anti-CTLA-4 antibody, MDX-010 (ipilim-umab), has also been tested in early-phase clinical trials. Hersh and colleaguespresented preliminary results from a phase II trial of 76 patients randomized to MDX-010 injections once a month for 4 months, alone or with dacarbazine.64 In the group treated only with MDX-010, PR was observed in 2 patients compared with 5 patients in the combination arm. These early results suggest that the MDX-010 antibody has activity in melanoma as a single agent and possibly improved efficacy combined with other therapies, such as vaccines, cytokines, or growth factors.

A recent phase I/II study administered MDX-010 with concomitant high-dose IL-2 therapy to 36 heavily pre-treated patients with stage IV melanoma.65 Objective tumor response was observed in 8 (22%) patients, comprising 3 CR and 5 PR. Although there was no evidence of a synergistic effect between the antibody and IL-2, investigators observed a correlation between tumor regression and the development of autoimmune phenomenon, such as vitiligo, a finding supported by previous studies.66-68

Weber and colleagues have also shown MDX-010 monotherapy to be effective in a trial of 88 patients with stage IV melanoma.69 They investigated 3 dosing regimens: multiple doses of 5 mg/kg, multiple doses of 10 mg/kg, or single doses of 20 mg/kg. Across all arms, ORR was 4.6%, with 1 CR and 3 PR, and 14 SD, 9 of which lasted ≥24 weeks. The more rigorous 10 mg/kg schedule was associated with an increase in severe adverse effects, most commonly colitis with diarrhea. In the 10 mg/kg arm, 2 patients with SD regressed to PR by the end of the study and 2 patients improved from SD to CR by week 48; another patient had a sustained CR lasting >57 weeks. The disease control rate (CR PR SD) was 19% for the study overall, with the greatest rate of disease control observed in the 10 mg/kg dosing group, at 39%, compared with 14.4% for the 5 mg/kg arm and 13.3% for the 20 mg/kg arm. OS at 18 months was 39%, and at 2 years, it dropped to 30%. (Table 3)


In a phase I proof-of-concept study, 55 previously treated patients with advanced melanoma received PLX4032, a small molecule oral inhibitor of BRAF.70 Approximately 60% of patients with melanoma have the BRAF mutation, believed to drive cell division in melanoma. PLX4032 induces cell cycle arrest and apoptosis exclusively in cells positive for the V600E BRAF mutation and does not affect healthy cells, minimizing adverse effects. Patients without the BRAF mutation did not respond to PLX4032.

After establishing a maximum tolerated dose (MTD) of 960 mg twice daily, researchers proceeded with a phase I extension study, in which 31 patients—all with the BRAF mutation—received the MTD until progression. Data for 27 evaluable patients showed a 70% ORR, including 1 CR after 3 cycles.71 Another 18 patients had PR, with 15 demonstrating >50% response. One patient with PR achieved CR after the data cutoff date. Minor response (10%-30%) was observed in 6 patients. Sites of tumor regression included subcutaneous tissue, liver, lung, bladder, gastrointestinal tract, and bone, and several patients had resolution of most symptoms. Although the study was not designed to measure PFS, investigators estimated PFS at ~8.5 months, noting that the median was not reached. Grade 3 dose-limiting toxicities included fatigue (7%), arthralgias (3%), photosensitivity (3%), and rash (3%).72 Squamous cell skin cancer developed in 23% of patients, and one-quarter of patients required treatment interruption or dose reductions. Although initial response rates are high, this may be partly attributable to the fact that all patients were prescreened and only those with the BRAF mutation were enrolled. Responses were transient, with the authors reporting that many patients with initial response subsequently relapsed and became resistant to PLX4032.72

Two multinational studies of PLX4032 are now underway, including the phase II BRIM2 trial, which began recruiting in September 2009, and the phase III randomized, controlled BRIM3 study, initiated in January 2010. BRIM2 is giving the MTD of PLX4032 to 100 patients with previously treated advanced melanoma. BRIM3 plans to recruit 700 previously untreated melanoma patients, who will receive PLX4032 or dacarbazine, with OS as the primary endpoint. Patients will be selected for both studies using an investigational diagnostic test developed to detect the BRAF mutation.


The overall prognosis for patients with advanced melanoma (stage III or IV) is poor.73,74 Most patients who experience a locoregional recurrence subsequently undergo a complete lymphadenectomy, with a 5-year OS rate of ~25%; any distant recurrence carries a much poorer prognosis, with a 5-year OS of 5% to 10%.75 Surgical options for patients with metastatic melanoma generally fall into three categories: curative, palliative, and immunotherapeutic.

Curative Surgery

The risks of attempting curative surgery should be weighed carefully against potential benefits. Recent data on the surgical management of metastatic melanoma indicate that certain factors are associated with improved OS, such as the ability to achieve a complete resection with negative margins, location of the initial site of metastasis, extent of metastatic disease, disease-free interval after excision of the primary melanoma, and stage of disease at initial diagnosis.76 Favorable sites of resection include the skin, subcutaneous tissue, lymph nodes, lung, and gastrointestinal tract; unfavorable sites of metastases are the brain, kidneys, and liver.77 Wong and associates found that skin and subcutaneous sites have the best long-term outcomes, with 5-year OS of 20% to 30% and median OS of 48 months. They had no 5-year survivors among patients with metastases to unfavorable sites, and the median OS was only 18.2 months. OS and possible long-term outcomes for most patients with stage IV disease seem ultimately to depend on the intrinsic biologic activity and nature of the melanoma (aggressive vs non-aggressive).

Palliative Surgery

The primary goal of palliative surgery is to relieve identifiable symptoms, such as chronic pain resulting from the progressive growth of metastatic tumor deposits. For instance, patients often present with bulky, recurrent lymphadenopathy of the axilla or groin, which can cause severe pain and partial sensory and motor limitations of the involved extremity. Other sites for successful surgical palliation include the gastrointestinal tract, where single metastatic deposits can cause bleeding or a high-grade obstruction. Single metastatic deposits within the brain can be resected for palliative intent; waiting may lead to surrounding brain edema, which can be fatal if left untreated. However, the outcome for such patients has been uniformly dismal, with a median survival of <1 year.77 Careful consideration should be given to palliative surgical intervention, with the benefits of alleviating the symptoms weighed against the likely morbidity and the complexity of the procedure.

Surgical Cytoreductive Immunotherapy

The last type of surgical intervention for stage IV melanoma is best described as oncologic surgical cytoreductive immunotherapy:removal of the bulk of metastatic disease to leave what is often referred to as minimal residual disease, in hopes of improving the overall function and dynamic nature of the host anti-tumor immune response.2 It is thought that complete resection of the tumor burden (or “tumor cell factory,” as Morton and associates describe it) may allow the host immune system to overcome tumor-induced immunosuppression and stimulate a more robust host immune response.2

As previously mentioned, the design of the onamelatucel-L vaccine trial for stage IV patients required all participants to undergo definitive surgical removal of all metastatic disease prior to enrollment. Although no advantage was seen for onamelatucel-L over placebo, a remarkably high 40% of patients in both arms were alive at 5 years. This suggests that prolonged survival was due to resection of metastatic disease.78


Many new agents are being investigated in melanoma, as are existing agents approved for other malignancies. Most are a long way from approval, and none appear to be curative. While the targeted novel drug PLX4032 has received a lot of recent attention, nearly all patients in the phase I study who demonstrated response later relapsed, suggesting it will need to be investigated with other agents. In addition, PLX4032 is active only in the 60% of patients with a BRAF mutation.

Our understanding of the tumor biology of melanoma clearly remains limited, but there is evidence that complete surgical resection of metastatic disease can play an important role in long-term survival in selected stage IV patients. The risks of surgery must be carefully weighed against the possible benefits, however. The continuing poor prognosis and lack of effective treatments for late-stage melanoma indicates the need for more patients to participate in clinical trials.