Update on Adjuvant Interferon Therapy for High-Risk Melanoma

September 1, 2002

Two of the most important predictors of relapse (and, therefore, survival) in patients with melanoma are the Breslow thickness of the primary melanoma and regional lymph node involvement. Patients with melanomas greater than 4 mm in thickness have approximately a 50% risk of recurrence, and those with lymph node involvement have a 50% to 85% risk of recurrence depending on the number of lymph nodes involved. Thus, a group of patients can be identified who are at high risk of death from melanoma and are, therefore, appropriate candidates for postsurgical adjuvant therapy.

Two of the most important predictors of relapse (and, therefore, survival) inpatients with melanoma are the Breslow thickness of the primary melanoma andregional lymph node involvement. Patients with melanomas greater than 4 mm inthickness have approximately a 50% risk of recurrence, and those with lymph nodeinvolvement have a 50% to 85% risk of recurrence depending on the number oflymph nodes involved. Thus, a group of patients can be identified who are athigh risk of death from melanoma and are, therefore, appropriate candidates forpostsurgical adjuvant therapy.

Over the past 20 years, numerous agents have been evaluated in a series ofnonrandomized and randomized adjuvant therapy trials in melanoma patients. Manyof these trials suffered from serious methodologic problems, (such as inadequatestatistical power, use of inappropriate controls, and lack of stratification forknown prognostic factors). However, the major obstacle to the success ofadjuvant therapy for melanoma has been the lack of active agents. Agents tested,with little or no benefit, include bacillus Calmette-Guérin (BCG), levamisole (Ergamisol),interferon gamma-1b (Actimmune), interleukin-2, retinoids, dacarbazine (DTIC-Dome),and megestrol acetate.

Drs. Agarwala and Kirkwood thoroughly review the completed and ongoing trialsof interferon-alpha as postsurgical adjuvant treatment of melanoma patients.Although clinicians generally agree with most of the conclusions reached by theauthors, the role of high-dose interferon in patients with high-risk melanomaremains controversial. Drs. Agarwala and Kirkwood conclude that high-doseinterferon is the standard of care for patients with high-risk melanoma and,therefore, is the most appropriate control for future adjuvant clinical trials.However, there is considerable discussion and debate as to whether high-doseinterferon improves overall survival in patients with high-risk melanoma.

ECOG Trial E1684

The data from the clinical trials reviewed by Drs. Agarwala and Kirkwoodconsistently demonstrate that high-dose interferon is associated withimprovement in disease-free survival. Multiple randomized trials have shown thathigh-dose interferon alfa-2b for 1 year improves relapse-free survival comparedwith observation. However, in terms of overall survival, the only clinical trialof high-dose interferon that has demonstrated an improvement in overall survivalis the Eastern Cooperative Oncology Group (ECOG) study E1684. In that trial,patients with high-risk melanoma were randomized to either high-dose interferonalfa-2b (Intron A) therapy for 1 year or observation after surgery. Asreported, the median survival was increased by 1 year in the treatment arm,and there was a 10% increase in the 5-year overall survival rate (46% forinterferon vs 37% for observation, one-sided P = .0237).

Does this result justify the use of interferon, and are the differencesbetween the two arms statistically significant? Only a one-sided test wasreported: Had the results been reported with a more standard and appropriatetwo-sided test, the resulting P value would have been greater and barelystatistically significant. Use of the less stringent measure is appropriate whenone assumes only a benefit for the experimental therapy and no possible harm. Infact, when a two-sided test was performed by Cole et al in a laterquality-of-life study using the same data set, there was no statisticallysignificant difference between the treatment and observation arms in terms ofoverall survival (P = .07).[1]

Moreover, an updated analysis of E1684 by Kirkwood no longer showed asurvival advantage at a median 12.6 years of follow-up: The high-doseinterferon arm had 93 deaths in 146 patients, whereas the observation arm had 95deaths in 140 patients. Although a trend toward improved overall survival in theinterferon-treated patients remains, the one-sided log-rank test showed a Pvalue of .09.[2]

Trial E1690

The second clinical trial of high-dose interferon designed to confirm theresults of E1684 was a three-arm trial comparing high-dose interferon, low-doseinterferon for 2 years, and observation (Intergroup E1690). The results of thisstudy confirmed an advantage in disease-free survival with high-dose interferontherapy and showed that the low-dose, longer-duration regimen was lesseffective. However, the estimated 5-year overall survival rate for the high-doseinterferon group was 52%, compared to 55% for the observation group.

Drs. Agarwala and Kirkwood offer many potential explanations as to why noadvantage in overall survival was seen in E1690 and suggest that a"crossover" effect and additional treatment received by patients inthe observation arm accounts for this lack of benefit. Although crossover cantheoretically bias results, the fact is that this study did not confirm theoverall survival results of E1684. The high-dose interferon regimenunequivocally showed no benefit in overall survival; indeed, the survival curvesare superimposable.

Trial E1694

The results of the most recent adjuvant study have also been cited asconfirmation of an improvement in overall survival with high-dose interferon.This study, E1694, compared 1 year of high-dose interferon with 2 years of aganglioside vaccine (GMK). Previous studies had shown that patients with IgMantibodies to this antigen had a better prognosis than those who did not.[3] Asreviewed by Drs. Agarwala and Kirkwood, survival and overall survival in theinterferon-alpha arm were equivalent to that in the vaccine arm. Again, using aone-sided log-rank test, the estimated 2-year overall survival rate was 78% inthe interferon alfa-2b arm vs 73% in the GMK arm.

The authors have suggested that the GMK vaccine was equivalent to anuntreated control arm. While it is possible that the GMK vaccine is similar toan observation arm, an alternative possibility is that the GMK vaccine has anegative impact on survival. Preclinical and clinical data suggest that certaintypes of immunologic responses to vaccines may, in fact, be deleterious.Specifically, IgG (rather than IgM) antibody responses to melanoma-associatedantigens have been associated with decreased overall survival in patients andmore aggressive melanoma tumor growth in some in vitro models.[4,5]

In E1694, 88% of the vaccinated patients were found after 1 month to have IgMantibodies to GM2. By the end of 1 year, only 35% of patients had significantIgM titers. The percentage of patients with IgG antibody titers increased overthe course of treatment from 26% of the vaccine arm at 1 month to 50% at1 year. Therefore, it is possible that vaccination with GMK stimulated anunintended and deleterious immune response, making it also possible thattreatment with the GMK vaccine, in fact, harmed patients. This may account forthe superiority of interferon over GMK in the study.

Drs. Agarwala and Kirkwood state that there was no evidence of an adverseeffect of the GMK vaccine, and that the outcome of patients randomized to theGMK vaccine, compared to the observation arm of previous adjuvant studies, wassimilar. This type of cross-study comparison is fraught with difficulties, ashighlighted by the authors when attempting to explain the lack of an overallsurvival benefit in E1690. In that study, the survival of patients in theobservation arm was 54% vs 37% in the observation arm of E1684. This substantialimprovement in survival occurred without any obvious therapeutic intervention.Therefore, it is difficult to conclude that, because survival of the untreatedcontrol arm in E1690 was similar to that of the vaccine arm in E1694, there wereno deleterious effects from treatment with the vaccine.

Thus, although data for high-dose interferon from several prospectiverandomized clinical trials consistently show improvement in disease-freesurvival, it is less clear that the agent improves overall survival in patientswith high-risk melanoma.

These efficacy results need to be considered carefully, in light of thesubstantial toxicity associated with 1 year of treatment with high-doseinterferon. Drs. Agarwala and Kirkwood thoroughly review the toxicities reportedin the randomized clinical trials conducted to date. Significant side effectsare common and often severe, including fatigue, fever, chills, myalgias,headache, nausea, and elevated hepatic transaminases. Depression is also afrequent side effect; in a recent study, 45% of patients had symptoms of majordepression, which may be ameliorated by pretreatment with antidepressivemedications.[6]


The optimal care for a patient with high-risk melanoma in 2002 is not clear.It requires integration of the existing evidence, the judgment of experiencedclinicians, and the informed input of patients. Interferon is an appropriateoption for some melanoma patients, but it is not the standard of care. It shouldremain under investigation for its place in therapy, but it should not inhibitthe search for new and better agents. Treatment decisions need to beindividualized. The decision to use interferon must be based on considerationssuch as the relative importance of increased disease-free survival compared tooverall survival and impact on quality of life.

The US Food and Drug Administration (FDA) approval of interferon-alpha foradjuvant treatment of high-risk melanoma has substantially changed the clinicalmanagement of these patients. The challenge now is to go beyond these initialresults. This will require more innovative and efficient clinical trials thatshould not be constrained by a mandate that interferon-alpha be used as acontrol. Increasingly, studies will employ more accurate identification of thoseat risk for disseminated metastatic disease, which becomes more vital as lesstoxic and more effective therapies are identified. Continued participation byphysicians and patients in well-designed, randomized clinical trials willfacilitate continued progress in the treatment of melanoma.


1. Cole BF, Gelber, R, Kirkwood JM, et al: Quality-of-life-adjusted survivalanalysis of interferon-alfa2b adjuvant treatment of high-risk resected cutaneousmelanoma: An Eastern Cooperative Oncology Group study. J Clin Oncol14:2666-2673, 1996.

2. Kirkwood JM, Manola J, Ibrahim J, et al: Pooled-analysis of four ECOG/Intergrouptrials of high-dose interferon alfa-2b (HDI) in 1916 patients with high-riskresected cutaneous melanoma (abstract 1395). Proc Am Soc Clin Oncol 20:350a,2001.

3. Livingston PO, Wong GYG, Adluri S, et al: Improved survival in stage IIImelanoma patients with GM2 antibodies: A randomized trial of adjuvantvaccination with GM2 ganglioside. J Clin Oncol 12:1036-1044, 1994.

4. Hsueh EC, Gupta RK, Qi K, et al: Correlation of specific immune responseswith survival in melanoma patients with distant metastases receiving polyvalentmelanoma cell vaccine. J Clin Oncol 16(9):2913-2920, 1998.

5. Ravindranath MH, Kelley MC, Jones RC, et al: Ratio of IgG:IgM antibodiesto sialyl Lewis(x) and GM3 correlates with tumor growth after immunization withmelanoma-cell vaccine with different adjuvants in mice. Int J Cancer75(1):117-124, 1998.

6. Musselman DL, Lawson DH, Gumnick JF, et al: Paroxetine for the preventionof depression induced by high-dose interferon alfa. N Engl J Med344(13):961-966, 2001.