Single-agent DTIC versus combination chemotherapy with or... : Melanoma Research (original) (raw)

Introduction

Cutaneous melanoma is an increasingly common tumour in the United States, with over 38,000 cases diagnosed in 1996. 1 The incidence of this tumour is increasing faster than any other solid tumour. In addition, the death rate from melanoma has doubled over the last three decades, further highlighting this disease as an important cause of cancer-related mortality in the USA. Some investigators have suggested that the true incidence may be two to three times greater than the above estimates. 2 It is also important to note that despite the fact that melanoma incidence is only 5% of that of non-melanoma skin cancer, it accounts for approximately three times as many deaths (roughly 7200 versus 2300 annually).

As with other solid tumours, disease stage is an important determinant of survival. Advanced melanoma, i.e. metastatic stage IV disease, is associated with an extremely poor median survival, i.e. of the order of 6 months. Few patients survive for more than 5 years (< 10%). Systemic chemotherapy is the mainstay of treatment for stage IV melanoma and is employed largely with palliative intent. Unfortunately, few agents have demonstrated substantial antitumour activity against metastatic melanoma. The alkylating agent dacarbazine (DTIC) is considered to be the most active drug for the treatment of this disease, with a response rate of the order of 20%. In order to improve the chemotherapy response rates in this disease, numerous clinical trials have been performed using combination chemotherapy (both DTIC-containing and non-DTIC-containing) as well as chemotherapy plus immunotherapy (i.e. interferons ± interleukin-2). Unfortunately, it remains unclear whether any of these combination therapies is superior to single-agent DTIC in this disease setting.

This report presents the results of a meta-analysis comparing the response rates of DTIC as single-agent therapy for metastatic melanoma with combination chemotherapy ± immunotherapy. Pooling the available clinical trial data using established meta-analytic techniques should provide a methodologically rigorous evaluation of the efficacy of DTIC in metastatic melanoma compared with the above cited combination regimens. These data may provide a basis for future clinical trial design.

Materials and methods

The methods employed in this analysis have been described previously. 3 Briefly, a study protocol was prospectively developed that outlined the purpose and methods of the study. Eligibility criteria for studies were determined prospectively, as were the specific data elements to be extracted from each trial. A plan for data analysis was also formulated as part of the study protocol. A data extraction form was designed for recording the relevant data from each published paper.

Literature retrieval was performed using previously described methods. 3,4 Briefly, a MEDLARS search was conducted covering the time period from January 1970 to January 1999. The CancerLit and EMBASE databases were also fully explored, as was the CD-ROM version of Current Contents. The search included all available languages. Electronic database searches were supplemented by manual searches of study bibliographies and review of relevant textbooks. If a series of papers were published, all data were retrieved from the most recent report.

The initial citations (in the form of abstracts) from this literature search were screened by a physician investigator to exclude those that did not meet the inclusion criteria specified by the protocol. Reasons for rejection included animal studies, in vitro studies, review articles, letters to the editor, and inclusion of patients with other than stage IV disease without stratification by stage. Citations selected from this initial search were subsequently screened for eligibility using the following criteria:

Citations meeting the above criteria were then entered onto an accept log and copies of the full papers obtained. Key data elements extracted from each trial included the year of publication, DTIC dosage and schedule, combination therapy drugs used in the comparison arm(s), number of patients randomized/analysed, percentage of female patients, prior chemotherapy (yes/no), mean/median Karnofsky or other measure of performance status, percentage with visceral metastases, presence/absence of brain metastases, number of metastatic sites, response rate (complete + partial), median time to tumour progression, and overall survival.

Statistical methods

Data analysis was performed according to meta-analysis procedures previously described by Peto et al. 5 This method is a modification of the Mantel-Haenzel method and is based on a fixed effects model. Study data were arranged in a 2 × 2 matrix and a summary, or Peto, odds ratio (ORp) and its 95% confidence interval (CI) were calculated. The primary outcome (event) of interest was the tumour response rate. The expected number of events in the ‘experimental arm’ (combination chemotherapy arm[s]) of each study was calculated and an estimate of the variance of the observed minus the expected number of events in each study was then determined, as well as the variance across all studies. Other outcomes examined were the time to tumour progression and overall survival.

Prior to the estimation of an ORp, a statistical test for homogeneity was performed (Q). This procedure tests the hypothesis that the effect sizes are equal in all of the studies. 5 If Q exceeds the upper tail critical value of χ2 (P < 0.05) at k − 1 degrees of freedom (where k equals the number of studies analysed), the observed variance in study effect size is significantly greater than what would be expected by chance if all studies shared a common population effect size. If the hypothesis that the studies are homogeneous is rejected, the studies are not measuring an effect of the same size and calculation of a pooled estimate of effect may be of questionable validity. In this setting, sensitivity analyses are performed to evaluate possible sources of heterogeneity.

Results

The literature search yielded a total of 175 citations. Full papers were obtained for the 22 of these that appeared to meet the inclusion criteria. 6–27 Of these, 20 were eligible for inclusion 8–27 and these are summarized in Table 1. All but three 10,14,26 of the accepted articles were published as full research papers. These 20 papers contained a total of 3273 randomized patients, of which 2706 (83%) were analysed for efficacy. Overall, 21 DTIC-alone arms (controls) were compared with 26 combination chemotherapy arms.

T1-9

Table 1:

Overview of included randomized trials

Data were collected on a number of known prognostic factors in this clinical setting in order to control for these parameters in the statistical analysis. These factors included performance status, the percentage of patients with more than three sites of metastatic disease, and the percentage of patients with visceral metastases. Unfortunately, only a minority of studies contained such information. For instance, only 10 of the 20 reports included information on patient performance status, while almost 75% did not detail information on the percentage of patients with visceral metastases. Most trials lacked information on the number of patients with three or more sites of metastases. This precluded analysis of the impact of these factors on the outcome measures of interest.

Based on the chemotherapeutic agents used in the combination therapy treatment arms, the available studies were stratified into three groups for analysis: DTIC- versus non-DTIC-containing combination regimens, DTIC- versus DTIC-containing drug regimens, and DTIC versus DTIC plus immunotherapy (see Table 1). Prior to combining all the studies in a meta-analysis, the overall survival and response rate (weighted by study sample size) for the DTIC control arms were calculated. Overall survival for patients treated with DTIC as a single agent was found to be 7.4 ± 0.6 months. The response rate (complete + partial) for DTIC alone was 16.9% (CI 14.7–19.1). Combining the 11 treatment arms that employed DTIC plus immunotherapy gave an overall response rate of 21.5%, while DTIC-containing combination regimens and non-DTIC-containing regimens demonstrated response rates of approximately 18%.

Although the above data suggest possible moderate differences in tumour response across the three chemotherapy classes, a meta-analysis was performed to definitively answer the question of whether any combination therapy is superior to single-agent DTIC. From the calculation of the ‘crude’ response rates it appears that DTIC is marginally active in stage IV melanoma. Since the toxicity of single-agent DTIC is generally less than that experienced with multi-drug therapy and since combination therapy may be associated with increased health care expenditures, it is important to determine whether any regimen demonstrates activity significantly beyond that seen with DTIC alone.

Initially, all 20 studies were combined in a meta-analysis using tumour response rate (complete + partial) as the outcome of interest (see Table 2). This analysis compared DTIC against all other combination drug regimens. The ORp was found to be 1.23, with a 95% CI of 1.02–1.48. An ORp of this magnitude suggests that combination therapy is associated with a better response rate than single-agent DTIC of the order of 23%. An analysis for heterogeneity showed Q to equal 23.6. With 26 degrees of freedom, this was a non-statistically significant result demonstrating a lack of heterogeneity across all studies.

T2-9

Table 2:

ORs and 95% CIs of 20 randomized trials (26 treatment arms) included in the meta-analysis

The 20 trials were next stratified into the three chemotherapy classes previously described, i.e. DTIC-containing, non-DTIC-containing and DTIC plus immunotherapy, and the response rates compared with DTIC alone. Combining the four trials (total of five treatment arms) using non-DTIC-containing combination regimens yielded a ORp of 0.77 (95% CI 0.45–1.32), a non-statistically significant result. Although the ORp is less than 1, favouring single-agent DTIC regimens, the wide CI that includes the null value (i.e. 1.00) does not support this contention. Therefore, the data show no statistically significant difference in the response rate between single-agent DTIC versus non-DTIC-containing multi-drug regimens.

The next analysis compared DTIC alone to DTIC-containing combination drug therapies. Ten studies representing 12 treatment arms were included in the analysis (total of 1273 patients). Five of these studies used identical doses of DTIC in both the experimental and control arm(s). 10,11,14,16,17 The data were combined in a meta-analysis, yielding a ORp of 1.33 with a 95% CI of 0.99–1.78, a marginally non-statistically significant result. Although non-significant, these data are suggestive of possible improved response rates with DTIC-containing combination regimens versus single-agent DTIC. That is, the resulting ORp suggests that combination therapies containing DTIC produce response rates 33% greater than DTIC alone.

Finally, single-agent DTIC arms were compared with all the DTIC plus interferon treatment arms (see Table 1); five studies employed interferon-α in six treatment arms. The ORp derived from this analysis was 1.53 (95% CI 1.10–2.13). The ORp suggests that the combination of DTIC and interferon-α is associated with, on average, a 53% greater clinical response compared with DTIC alone. This result is statistically significant (P = 0.01). Overall, the above data provide evidence that DTIC plus interferon-α may produce tumour response rates superior to those seen with DTIC alone. Survival curves were available for five of the seven DTIC plus interferon treatment arms. Combining these survival data showed no statistically significant difference in survival between the single-agent DTIC control arms and the combination therapy regimen (OR = 1.16, 95% CI 0.87–1.54).

Discussion

The prognosis for patients with disseminated malignant melanoma remains poor, with a median survival of the order of 6–12 months. Although a very small minority of patients may survive long-term (i.e. > 5 years), stage IV melanoma remains a rapidly fatal disease despite the availability of a multitude of chemotherapeutic and biological agents.

DTIC represents the cornerstone of chemotherapeutic management of this disease. Response rates for single-agent DTIC are often cited as ranging from 10 to 25%, with complete responses seen in fewer than 5% of patients. In addition, the response duration is often brief, i.e. 5–6 months. The medical literature contains numerous phase II and phase III trials employing various drug combinations, including the use of biological therapy, in an attempt to improve upon the very modest activity of single-agent DTIC. At present, it is unclear if any of these combination therapy regimens is superior to single-agent DTIC in terms of tumour response. Single-agent DTIC is a generally well tolerated regimen, with the major side effects being nausea and vomiting. Combination therapies are associated with greater morbidity (toxicity) that impacts not only on quality of life but health care expenditure. In this context, clearly demonstrating the superiority of multi-drug regimens is necessary to justify their routine use.

The present analysis is an attempt to provide a methodologically rigorous evaluation of this question. Using accepted meta-analytic methods, data from over 3000 patients with metastatic melanoma were pooled. Overall, the combination of DTIC with interferon-α demonstrated a statistically significant improvement in tumour response when compared with DTIC alone. The magnitude of the improvement in response rate may be of the order of 53%. Although the superior response rate was not associated with improved survival, the analysis was based on a relatively small subset of patients included in the overall meta-analysis (n = 549). In addition, improved clinical response may result in improvements in functional status, quality of life and possibly reduced medical expenditure. These endpoints should be addressed in future clinical trials.

Other DTIC-containing combination therapies also showed response rates that were, on average, 33% greater than those seen with single-agent DTIC. Although the CI for this latter comparison included the null value (i.e. a non-statistically significant result), the OR was only marginally non-significant, i.e. 0.99–1.78. This finding lends weight to the contention that combination chemotherapy protocols produce response rates superior to those seen using DTIC alone.

One of the important limitations of the existing clinical trial database is the lack of stratification of clinical trial data by known prognostic factors, i.e. site of metastasis (skin, lung and lymph nodes versus visceral metastases), number of organs involved with disease, and performance status. As pointed out earlier, many of the 20 trials included in this meta-analysis failed to include such information. This precluded evaluation of the possible impact of these factors on the outcome of interest. Future clinical trials must control for recognized prognostic factors in order to accurately characterize the clinical activity of the drugs under study.

The strengths of the present analysis include the comprehensive nature of the literature search, an explicit protocol indicating inclusion and exclusion criteria for clinical trials, the use of accepted meta-analytical techniques for pooling data from multiple randomized trials, and the large sample size analysed. Since DTIC is considered the ‘gold standard’ in metastatic melanoma, this analysis sought to summarize all the available clinical trial data comparing combination chemotherapy to this ‘standard'. Given the uncertainty regarding the most active regimen for this disease, comparison against the best studied and most widely accepted agent appears to be the most useful approach.

In summary, the present meta-analysis of over 3000 patients from 20 randomized trials demonstrates that the combination of DTIC plus interferon-α produces response rates 53% greater than DTIC alone in stage IV malignant melanoma. The data detailed above should serve as a basis for future clinical trial design and as a clinical ‘benchmark’ for the comparison of new chemotherapy regimens in this disease.

References

1. Ries LAG, Kosary CL, Hankey BF, et al. SEER Cancer Statistics Review: 1973–1994, NIH publication no. 97-2789. Bethesda: National Cancer Institute, 1997.

2. Rigel DS, Kopf AW, Friedman RJ. The rate of malignant melanoma in the United States: are we making an impact? J Am Acad Dermatol 1997; 17: 1050. 1050.

3. Hedges LV, Olkin J. Statistical Methods for Meta-Analysis. San Diego: Academic Press, 1985.

4. Cooper H, Hedges LV. The Handbook of Research Synthesis. New York: Russell Sage Foundation, 1994.

5. Yusuf S, Peto R, Lewis J. et al. Beta-blockade during and after myocardial infarction: an overview of the randomized trials. Prog Cardiovasc Dis 1985; 27: 335 –337.

6. Costanza ME, Nathanson L, Costello WG. et al. Results of a randomized study comparing DTIC with TIC mustard in malignant melanoma. Cancer 1976; 37 (4): 1654 –1659.

7. Rudolf Z, Strojan P. DTIC versus IFN-alpha plus DTIC in the treatment of patients with metastatic malignant melanoma. Neoplasma 1996; 43 (2): 93 –97.

8. Bajetta E, DiLeo A, Zampino MG. et al. Multicenter randomized trial of dacarbazine alone or in combination with two different doses and schedules of interferon alfa-2a in the treatment of advanced melanoma. J Clin Oncol 1994; 12 (4): 806 –810.

9. Bellett RE, Mastrangelo MJ, Laucius JF. et al. Randomized prospective trial of DTIC alone versus BCNU plus vincristine in the treatment of metastatic malignant melanoma. Cancer Treat Rep 1996; 60 (5): 595 –600.

10. Buzaid AC, Legha S, Winn R. et al. Cisplatin, vinblastine and dacarbazine vs dacarbazine alone in metastatic melanoma: preliminary results of a Phase III cancer community oncology (CCOP) trial. Proc Annu Meet Am Soc Clin Oncol 1993; 12: A1328. A1328.

11. Carter RD, Krementz ET, Hill GJ. et al. DTIC and combination therapy for melanoma. : I. Studies with DTIC, BCNU, CCNU, vincristine and hydroxyurea. Cancer Treat Rep 1976; 60 (5): 601 –609.

12. Clunie GJ, Gough IR, Dury M. et al. A trial of imidazole carboxamide and Corynebacterium parvum in disseminated melanoma: clinical and immunologic results. Cancer 1980; 46 (3): 475 –479.

13. Chauvergne J, Bui NB, Cappelaere P. et al. Chemotherapy in advanced malignant melanoma. : Results of a controlled trial comparing a combination of dacarbazine and dacarbazine alone. Semin Hosp 1982; 58 (46): 2697 –2701.

14. Chiarion Sileni V, Nortilli R, Medici M. et al. BCNU, cisplatin, dacarbazine and tamoxifen in metastatic melanoma: results of a randomized Phase II study. Proc Annu Meet Am Soc Clin Oncol 1997; 16: A1782. A1782.

15. Cocconi G, Bella M, Calabresi F. et al. Treatment of metastatic malignant melanoma with dacarbazine plus tamoxifen. N Engl J Med 1992; 327: 516 –523.

16. Costanza ME, Nathanson L, Lenhard R. et al. Therapy of malignant melanoma with an imidazole carboxamide and bis-chloroethyl nitrosourea. Cancer 1972; 30: 1457 –1461.

17. Costanza ME, Nathanson L, Schoenfeld D. et al. Results with methyl-CCNU and DTIC in metastatic melanoma. Cancer 1977; 40 (3): 1010 –1015.

18. Falkson CI, Ibrahim J, Kirkwood JM. et al. Phase III trial of dacarbazine versus dacarbazine with interferon alpha-2 versus dacarbazine with tamoxifen versus dacarbazine with interferon alpha-2 and tamoxifen in patients with metastatic malignant melanoma: an Eastern Cooperative Oncology Group study. J Clin Oncol 1998; 16 (5): 1743 –1751.

19. Falkson CI. Experience with interferon alpha-2b combined with dacarbazine in the treatment of metastatic malignant melanoma. Med Oncol 1995; 12: 35 –40.

20. Kongoniia LM, Moroz LV, Perevodchikova NI. et al. Comparison of the efficacy of imidazole-carboxamide and of a combination of nitrosomethylurea, vincristine and dactinomycin in disseminated melanoma. Vopr Onkol 1981; 27 (4): 16 –21.

21. Lopez M, Perno CF, Lauro L. et al. Controlled study of DTIC versus DTIC plus epirubicin in metastatic malignant melanoma. Invest New Drugs 1984; 2 (3): 319 –322.

22. Luikart SD, Kennealey GT, Kirkwood JM. Randomized phase III trial of vinblastine, bleomycin and cis-dichlorodiammine-platinum versus dacarbazine in malignant melanoma. J Clin Oncol 1984; 2 (3): 164 –168.

23. Moon JH, Gailani S, Cooper R. et al. Comparison of the combination of BCNU and vincristine with two dose schedules of DTIC in the treatment of disseminated malignant melanoma. Cancer 1975; 35: 368 –371.

24. Ringborg U, Rudenstam CM, Hansson J. et al. Dacarbazine versus dacarbazine-vindesine in disseminated malignant melanoma: a randomized phase II study. Med Oncol Tumor Pharmacother 1989; 6 (4): 285 –289.

25. Thomson DB, Adena M, McLeod GR. et al. Interferon alpha-2a does not improve response or survival when combined with dacarbazine in metastatic melanoma: results of a multi-institutional Australian randomized trial. Melanoma Res 1993; 3 (2): 133 –138.

26. Vorobiof DA, Falkson G, Voges CW. DTIC versus DTIC and recombinant interferon alpha-2b in the treatment of patients with advanced malignant melanoma. Proc ASCO 1989; 8: 284. 284.

27. WHO Collaborating Centres for Evaluation of Methods of Diagnosis and Treatment of Melanoma. Controlled study with DTIC + BCG and DTIC +Corynebacterium parvum in advanced melanoma. Tumori 1984; 70: 41 –48.

Keywords:

dacarbazine; interferon; meta-analysis; response rate

© 2001 Lippincott Williams & Wilkins, Inc.