Summary of findings
The management of people with lymph node metastases from cutaneous melanoma is one of the most challenging issues for surgical and medical oncologists. In fact, the therapeutic role of sentinel node biopsy, radical lymph node dissection, and adjuvant therapies has never been clearly proven (Australian Cancer Network 2008; Thompson 2009).
After surgery, the only approved medical adjuvant therapy for people with lymph node metastases is interferon (IFN) alpha (Kirkwood 1996). After the first trial reporting an overall survival benefit (Kirkwood 1996), the FDA (the United States Food and Drug Administration) and the EMEA (European Medicines Agency) approved high-dose interferon for the adjuvant treatment of high-risk melanoma patients (i.e. TNM stage IIb-IIIC) (Eggermont 2009). On the basis of this randomised controlled trial (RCT), the standard regimen is currently one month of high-dose intravenous interferon, followed by one year of low-dose IFN.
Following the RCT published by Kirkwood and Colleagues in 1996, many trials have been conducted to compare interferon versus observation, or different interferon schedules and dosages (Ascierto 2008). The results appear to confirm a benefit in terms of disease-free survival, but the impact of adjuvant interferon on overall survival is quite unclear (Kirkwood 2004; Lens 2002; Pirard 2004; Wheatley 2003). The results of an individual patient data meta-analysis on this subject have been published in the form of an abstract: They confirm longer disease-free survival for those treated with IFN and, interestingly, pinpoint a small but significant advantage in terms of overall survival (Wheatley 2007).
Based on the disease-free survival advantage, the small overall survival benefit, and the high toxicity, the role of interferon alpha in melanoma remains debatable (Ascierto 2008; Bajetta 2008; Eggermont 2008a; Janku 2010; Kirkwood 2009; Thirlwell 2008; Verma 2006). Accordingly, because the overall survival benefit may be small when weighed against possible toxicity, the international guidelines do not routinely recommend interferon. However, in those participants with metastatic regional lymph nodes who have had a lymphadenectomy, some guidelines suggest its use may be considered (Australian Cancer Network 2008; Dummer 2009; Fecher 2009; Saiag 2007).
We present a glossary of terms used in this review in Table 1.
Description of the condition
Although melanoma is not the most common form of skin cancer, it accounts for 80% of all skin cancer deaths (Jemal 2010; Lens 2004; Lui 2007; Tawbi 2007; Thompson 2005; Tsao 2004; Wang 2007). During the 20th century, the incidence of melanoma in white populations rose faster than the incidence of any other solid tumour, barring lung cancer (Ferlay 2010; Jemal 2010; Mocellin 2011). Data from the United States indicates that in 2001, 47,700 new cases were recorded compared to 59,940 new cases in 2007. Worldwide, 160,000 new cases and 41,000 deaths were reported in 2002. In 2007, the American Cancer Society estimated that the lifetime risk of melanoma is higher for men (1/49) than for women (1/73), the death rate from melanoma reaching 8110 cases in that same year (Wang 2007). These data clearly indicate that cutaneous melanoma represents a growing public health problem.
Early diagnosis and surgical removal of the primary tumour still represent the approach with the highest likelihood of cure for participants with early stage melanoma (Mocellin 2011a; Sladden 2009). In this setting, sentinel node biopsy provides patients with a significant benefit in terms of disease-free survival, and when there is metastatic disease present in the regional lymph nodes, there is a possible overall survival advantage (Morton 2006; Pasquali 2010a).
In people with advanced melanoma, resection is recommended for selected cases with limited metastatic disease (Thompson 2005; Tsao 2004). Anti-CTLA-4 monoclonal antibodies and inhibitors of v-Raf murine sarcoma viral oncogene homologue B1 (BRAF) recently gained the approval of drug agencies for AJCC stage IV disseminated melanoma (Chapman 2011; Hodi 2010), a condition considered as one of the most resistant to conventional chemotherapy (Crosby 2000; Gogas 2007; Lui 2007; Sasse 2007; Tawbi 2007).
In those with high-risk melanoma, AJCC stage II (T2-4N0M0) and AJCC stage III (TanyN+M0) disease, the 5-year overall survival rate varies between 30% and 70% (Balch 2009), and the only agents currently approved for the treatment of these patients in an adjuvant setting after radical surgery are interferon alpha and pegylated interferon (Eggermont 2009; Garbe 2010; Kirkwood 2008).
Description of the intervention
Interferon alpha is a type I interferon mainly produced by macrophages (Theofilopoulos 2005). The interferon cluster region of chromosome 9p22 encodes 13 different interferon genes: Among them, the interferon-2 gene presents 3 polymorphic variants, known as interferon alpha 2a, interferon alpha 2b, and interferon alpha 2c (Pestka 2007). Interferon alpha demonstrated anticancer effects in both experimental models and in the clinical setting, although its exact mechanism of action is still unclear (Moschos 2007; Pasquali 2010). With regard to the treatment of melanoma, only the proteins encoded by the interferon alpha 2a and interferon alpha 2b genes, which differ in a single amino acid at position 23 (lysine > arginine), have been tested as therapeutic agents in the clinical setting, and human recombinant interferon alpha 2b has been approved for the adjuvant treatment of this type of skin cancer.
After apparently radical surgery of primary melanoma without clinical evidence of distant metastases, the only approved medical adjuvant therapy for patients with high-risk melanoma is interferon alpha (Kirkwood 2008). Based on the results from the first trial reporting an overall survival benefit (Kirkwood 1996), the US Food and Drug Administration (FDA) and the European Medicines Agency approved the high-dose interferon (IFN) for the adjuvant treatment of high-risk melanoma patients (i.e. AJCC TNM stage IIB-IIIC) (Eggermont 2009). Recently, based on the EORTC (European Organisation for Research and Treatment of Cancer) trial 18991 (Eggermont 2008), the FDA approved pegylated interferon. Currently, the standard regimen is one month of high-dose intravenous IFN, followed by one year of low-dose interferon. However, different schedules of interferon therapy have also been tested (Lens 2006). There are four main approaches to interferon therapy:
- high-dose interferon alpha administered for one month (and followed by one year at a low dose),
- interferon alpha at an intermediate dose (tested for one or two years),
- interferon alpha at a low dose (tested for one or three years), or
- pegylated interferon.
High-dose interferon is burdened by the highest toxicity; however, it is believed to be the most active regimen (Kirkwood 2008).
How the intervention might work
The best known physiological activity of interferons is their antiviral function through their ability to inhibit the replication of different types of viruses, both by blocking the cell cycle machinery and by stimulating the immune response. Interferon has also long been recognised as an anticancer factor in both experimental and clinical models (e.g. myeloid leukaemia, renal cell carcinoma, and melanoma). However, the cellular and molecular mechanism underlying this activity is still incompletely elucidated (Borden 2007; Pasquali 2010). Interesting findings about the mechanism of action of interferon in melanoma came from an observational study in which interferon was administered as neoadjuvant treatment in participants with clinically detectable lymph node metastases (Moschos 2006). The subsequent radical lymph node dissection allowed the investigators to study the effects of interferon on tumour tissue and to examine immunologic and molecular biomarkers of tumour response. Interferon did not influence tumour cell phenotype, proliferation rate, apoptosis, or angiogenesis. On the other hand, a higher grade of tumour-infiltrating-mononuclear-immune cell was observed in cases showing response to interferon treatment. Therefore, the main anticancer effect of interferon is currently believed to be linked to its immunostimulatory effects.
Why it is important to do this review
Several randomised studies have been designed and conducted on the use of interferon alpha as an adjuvant treatment for melanoma (Cascinelli 2001; Eggermont 2008; Garbe 2008; Kirkwood 2004). While the disease-free survival benefit for interferon treatment was consistent across the trials, a meaningful overall survival benefit for interferon therapy has been reported in some studies only (Garbe 2008; Kirkwood 1996). According to 2 meta-analyses published more than 6 years ago on the results of 12 (Wheatley 2003) and 9 (Pirard 2004) randomised controlled trials, respectively, interferon alpha appears to provide a statistically significant disease-free survival advantage (mainly over observation) in participants with high-risk cutaneous melanoma, whereas no impact on overall survival was demonstrated. Since then, 2 other large randomised controlled trials have been published that compared interferon and observation. These enrolled 1700 participants, but the findings have been conflicting in terms of overall survival benefit (1 trial is positive: Garbe 2008, and the other one showed no differences: Eggermont 2008).
With the aim of summarising the evidence on interferon, we carried out a meta-analysis that compared interferon treatment versus any other comparator than interferon in participants with high-risk cutaneous melanoma (Mocellin 2010). Our findings confirmed the disease-free survival benefit reported by the previous meta-analyses, and supported the significant 3% overall survival advantage reported by the individual patient data meta-analysis of Wheatley (Wheatley 2007). These meta-analyses did not include results for at least two more trials (the Nordic Trial and the Sunbelt Melanoma Trial), which are currently only available in abstract form (Hansson 2011; McMasters 2008).
The most recent meta-analyses on interferon as adjuvant treatment for melanoma have not investigated adverse events and quality of life (Mocellin 2010; Wheatley 2007). Even if these topics are relevant in oncology, these end points are not suitable for a meta-analysis. In fact, almost all trials compare interferon to observation. Interferon arms are intrinsically associated with lower quality of life and presence of adverse events when compared to control arms; thus, quality of life and adverse events data are hardly reported in terms of summary data suitable for meta-analysis.
Therefore, we have performed this systematic review and meta-analysed the available evidence in order to determine the true benefit of interferon in the adjuvant treatment of melanoma.
To assess the disease-free survival and overall survival effects of interferon alpha as adjuvant treatment for people with high-risk cutaneous melanoma.
We have not addressed adverse events and quality of life, even if they represent a crucial issue in interferon alpha treatment, in this systematic review, as virtually all trials have compared interferon treatment with observation (i.e. no treatment).
Criteria for considering studies for this review
Types of studies
We considered as eligible all randomised controlled trials that compared interferon alpha with observation (or any regimen other than interferon alpha) for the adjuvant treatment of skin melanoma.
Types of participants
We included people with high-risk skin melanoma, that is, those with regional lymph node metastasis undergoing radical lymph node dissection (AJCC stage III), or people without nodal disease but with primary tumour thickness greater than 1 mm (AJCC stage II).
Types of interventions
We included adjuvant (i.e. postoperative) interferon (experimental arm) versus observation or any treatment other than interferon (control arm).
Types of outcome measures
The main outcome measure was the hazard ratio (HR), defined as the ratio between the risk of event (disease recurrence or death) in the treatment arm (adjuvant interferon) and the same risk in the control arm (no adjuvant interferon).
Moreover, for overall survival data, we calculated the number needed to treat (NNT), defined as the number of participants to be treated with IFN to avoid one event (death).
The primary outcomes considered were disease-free survival and overall survival. The survival time was calculated from the date of randomisation.
Search methods for identification of studies
We aimed to identify all relevant randomised controlled trials regardless of language, publication status (published, unpublished, in press, or in progress), or publication format (abstract, meeting presentation, full-text article).
We searched the following databases up to 22 August 2012:
- the Cochrane Skin Group Specialised Register using the following terms: melanoma* and (interferon* or interferon*);
- the Cochrane Central Register of Controlled Trials (CENTRAL) 2012, issue 8, in The Cochrane Library using the search strategy in Appendix 1;
- MEDLINE via OVID (from 2005) using the strategy in Appendix 2;
- EMBASE via OVID (from 2010) using the strategy in Appendix 3;
- AMED via OVID (Allied and Complementary Medicine, from 1985) using the strategy in Appendix 4;
- LILACS (Latin American and Caribbean Health Science Information database, from 1982) using the strategy in Appendix 5; and
- ASCO (American Society of Clinical Oncology) Annual Meeting Abstracts 2000 to 2011.
The UK and US Cochrane Centres have ongoing projects to systematically search MEDLINE and EMBASE for reports of trials that are then included in the Cochrane Central Register of Controlled Trials. Searching has currently been completed in MEDLINE to 2004 and in EMBASE to 2009. Further searches of these two databases were undertaken for this review by the Cochrane Skin Group to cover the years not searched by the UK and US Cochrane Centres for CENTRAL.
We searched the following trials registers using the terms "melanoma" and "interferon" during the second half of 2011:
- The metaRegister of Controlled Trials (www.controlled-trials.com).
- The US National Institutes of Health Ongoing Trials Register (www.clinicaltrials.gov).
- The Australian and New Zealand Clinical Trials Registry (www.anzctr.org.au).
- The World Health Organization International Clinical Trials Registry platform (www.who.int/trialsearch).
Searching other resources
We checked the bibliographies of selected studies for further references to relevant trials.
Data collection and analysis
We performed meta-analysis following the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) and PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (Moher 2009). We applied standard meta-analysis methods (Sutton 2000) to evaluate the summary effect of postoperative interferon on disease-free and overall survival.
Selection of studies
We assessed eligibility of the retrieved articles by title and abstract (SP and SM). If this information was insufficient for eligibility assessment, we reviewed the full article.
Data extraction and management
Two authors extracted data (SP and ML), and a third author independently verified the extracted data (SM). After all eligible studies had been identified, two authors independently performed a quality evaluation of the heterogeneity between studies, randomisation, blinding, and follow-up, as well as data extraction (SM and SP). The authors achieved consensus on all results considered for the final analysis.
The following data were retrieved (if reported): demographics, study design, eligibility criteria, randomisation method, allocation method, disease stage at enrolment, type of interferon alpha, interferon alpha dosage, treatment duration, number of participants enrolled, number of participants randomised, number evaluated for data analyses, cross-over allowance and percentage, intention-to-treat versus per-protocol findings, follow-up period, disease-free survival, overall survival, and loss to follow up.
Assessment of risk of bias in included studies
We assessed risk of bias according to The Cochrane Collaboration checklist (Higgins 2011):
(a) the method of generation of the randomisation sequence;
(b) the method of allocation concealment;
(c) the blinding of participants, clinicians, and outcome assessors;
(d) the presence of incomplete outcome data; and
(e) selective outcome reporting.
This information was recorded in a 'Risk of bias' table, which is part of the 'Characteristics of included studies' table for each study.
Measures of treatment effect
The main outcome measure was the hazard ratio (HR), defined as the ratio between the risk of event in the treatment arm (adjuvant interferon) and the same risk in the control arm (no adjuvant interferon). Hazard ratios were reported with their 95% confidence intervals (CIs). Survival data (HR) were either entered directly in RevMan or extrapolated from Kaplan-Meier plots using dedicated methods (Parmar 1998; Tierney 2007). Moreover, for overall survival data, we calculated the number needed to treat (NNT), defined as the number of participants to be treated with IFN to avoid one event (death).
Unit of analysis issues
For multiple-arm trials in which two (or more) interferon arms were compared with the same control arm, we took within-study correlation into consideration. As suggested by Borenstein and colleagues (Borenstein 2009), we first computed a composite effect size for the comparison of each interferon arm versus the control arm. Next, we calculated the correlation factor (r) based on the number of cases in each arm, which allowed us to compute the variance (V) of the composite effect size according to the following formula: V = 1/4 [(V1 + V2) + (2r*√ V1*√ V2)], where V1 and V2 are the variances of the original comparisons between each treatment arm and the control arm. Using this variance, we finally computed the standard error and then the 95% CI of the composite effect.
Dealing with missing data
We contacted trial authors of included studies whenever data essential for the meta-analysis were missing or unclear.
Assessment of heterogeneity
We assessed the consistency of results (effect sizes) among studies using the two standard heterogeneity tests: the Chi²-based Cochran Q-test and the I² statistic (Higgins 2002). To be more conservative, we considered that heterogeneity was statistically significant when the Cochran Q-test P value was less than 0.1 (i.e. the alpha level of significance for this test was set at 10%). In addition, inconsistency across studies was considered low, moderate, and high for I² statistic values lower than 25%, between 25% and 50%, and greater than 50%, respectively. We considered heterogeneity significant either when the I² statistic was greater than 50%, the Q-test P value was smaller than 0.1, or both: In this case, we applied the random-effects model to calculate the overall effect (which assumes that studies do not share the same common effect, and assigns a weight to each study taking into account both within- and between-study variance), using the inverse variance method of DerSimonian and Laird (DerSimonian 1986). In all other cases, we applied the fixed-effect model.
Assessment of reporting biases
We used a funnel plot to detect the so-called "small study effect". In fact, publication and selection biases might burden small studies, which may show a lower methodological quality, possibly leading to a small effect in a meta-analysis, the smaller studies having larger effects. Furthermore, between-trial heterogeneity, which is observed in studies with a relatively small number of participants, might lower the study effect (Sterne 2000). We formally investigated funnel plot asymmetry with the Egger linear regression approach and the Begg rank correlation test: To be more conservative, we considered these tests statistically significant when the P value was less than 0.1. We formally assessed the impact of a small study effect bias on the summary effects by means of the 'trim & fill' method described by Duval and Tweedie (Duval 2000).
In order to assess the impact of "small" studies, we also performed a sensitivity analysis (see Sensitivity analysis) by excluding these trials. To this aim, we defined "small" as any study with a sample size smaller than 400 based on the following parameters of an ideal trial: A) minimum hazard ratio (HR) = 1.5; alpha level of significance = 5%; statistical power = 80%; median survival in controls = 60 months; accrual time = 36 months; follow-up time = 48 months; randomisation ratio = 1:1.
We used hazard ratios (HR) and 95% confidence intervals (CI) to synthesise the summary effect of interferon in terms of both disease-free survival and overall survival.
Moreover, based on the summary HR (calculated with the meta-analysis) and the 5-year OS rate in the control population of participants with AJCC stage II-III melanoma (roughly 60%), we calculated the number needed to treat (NNT), which provides readers with the number of participants to be treated with interferon in order to avoid one event (i.e. death); to this aim, we followed the method proposed by Altman (Altman 1999).
Subgroup analysis and investigation of heterogeneity
In order to investigate potential sources of heterogeneity, we used subgroup analysis and metaregression (method of moments) for the following factors: interferon dosage, treatment duration, type of interferon alpha used, participants' AJCC TNM stage (studies have been conducted with the 5th and 6th edition of the TNM staging manual), type of comparator, and year of publication.
We used the 'leave-one-out' sensitivity analysis to systematically ascertain the impact of individual randomised studies on the overall effect. We employed the exclusion of randomised studies for specific reasons (e.g. differences in control arms or sample size) as a further type of sensitivity analysis.
Description of studies
Since we systematically reported the details of the included and excluded studies in the dedicated 'Characteristics of included studies' and 'Characteristics of excluded studies' tables, here we discuss some general aspects of the trials included in the meta-analysis, as well as some peculiarities of specific RCTs.
Results of the search
The literature searches yielded 640 papers plus 4 papers from other sources (we report the review's flow diagram in Figure 1). Of these 644 references, we excluded 609 and considered the full text of the remaining 35. We then excluded 17 and included 18.
|Figure 1. Study flow diagram|
We included 18 RCTs, with a total of 10,499 participants. These were published between 1995 and 2011. Table 2 summarises design, sample sizes, setting, participants, interventions, and outcomes of eligible studies. All but one study (a phase II RCT) were phase III RCTs.
Sample sizes ranged between 96 and 1388 participants.
Studies were conducted across Europe (mainly within centres belonging to the European Organisation for Research and Treatment of Cancer (EORTC)), the United States (mainly within centres belonging to the Eastern Cooperative Oncology Group (ECOG)), and Australia.
Included studies investigated the impact of interferon alpha on disease-free survival and overall survival of people with cutaneous melanoma after surgery. All participants underwent radical surgery for AJCC stage II (n = 810; 7.8%), AJCC stage III (n = 3512; 33.9%), or AJCC stages II–III (n = 6023; 58.3%) cutaneous melanoma. Participants were both men and women, generally aged 18 to 70 years.
Interferon regimens varied in terms of dosage (high-dose (20 mega-units (MU)/m²), intermediate-dose (10 MU/m²), and low-dose (1 to 3 MU/m²)), administration route (subcutaneously (s.c.), intramuscularly (i.m.), and intravenously (i.v.)), and duration of treatment (4 months to 5 years), as detailed in the dedicated section: 'Characteristics of included studies'.
The two main causes of interruption to interferon treatment were disease progression or toxic effects: In the latter case, rates of treatment discontinuation (or dose reduction/delay) ranged from 0% to 58% (median 15%).
In terms of surgery, all participants received primary tumour excision in both interferon and control arms. Radical lymph node dissection was performed upon clinical and pathological evidence of lymph node metastasis in all RCTs, except for the E1684 trial (Kirkwood 1996), in which the enrolled participants underwent elective lymphadenectomy. Moreover, in trial EORTC18871 (Kleeberg 2004), the physician in charge performed elective lymph node dissection at their own discretion. The spread to regional lymph nodes was assessed mainly by means of physical examination (clinically evident metastatic disease) until the year 2000. Thereafter, sentinel node biopsy for the detection of subclinical metastatic disease was allowed by trial protocols. Many study designs did not consider sentinel node biopsy, thus, precluding the performance of subgroup meta-analysis based on the type of lymph node positivity (i.e. macrometastasis (clinically palpable lymph nodes) versus micrometastasis (i.e. revealed by pathological assessment of the sentinel node biopsy)).
In one RCT (Kirkwood 2001a), the treatment arm regimen consisted of interferon combined with GMK (a ganglioside-based anticancer vaccine). Because GMK alone was used in the comparison treatment arm, this study allowed the evaluation of any therapeutic benefit of IFN treatment; thus, we included it in the meta-analysis. In addition, we used the 'leave-one-out' procedure as a sensitivity analysis to determine whether this trial had any impact on the overall effect estimated by the meta-analysis (see results in Effects of interventions).
Since the findings from 2 trials, EORTC18871 and DKG80-1, were reported in a single article as pooled results (and consequently as pooled HRs) (Kleeberg 2004), we included these 2 trials as a single study in the present meta-analysis.
We reported the reasons for the exclusion of 17 studies in the 'Characteristics of excluded studies' tables.
In one trial (Kokoschka 1990), although the authors mentioned that participants were randomly selected to receive interferon therapy, there was no mention of a corresponding control arm, and it seemed that surgically treated melanoma participants who were diagnosed at the time of the study were considered as a non-randomised control group. In addition, participants with both AJCC TNM stage II and stage III melanoma were enrolled without mentioning any randomisation, and the results were represented separately as Kaplan–Meier curves for disease-free survival (DFS), with no log-rank P values, for the two stages. Also, the authors of this trial claimed that both stage I and stage II participants benefited from IFN therapy, although no difference in DFS or OS was detected between treatment and observation groups. All attempts to contact principal investigators of this trial were unsuccessful.
For two RCTs, the results were partially excluded. In a three-arm RCT (Garbe 2008), the observation group was compared with both interferon alone and a combination of interferon and dacarbazine. In this case, we included only the comparison of IFN alone versus observation. The IFN plus dacarbazine arm was designed specifically to answer whether dacarbazine adds any therapeutic advantage to IFN; thus, we did not include results from this arm in the present meta-analysis. One RCT (McMasters 2008) included two types of participants with AJCC TNM stage III disease: those with positive sentinel lymph node based on standard pathological evaluation and those classified as positive based on the results of polymerase chain reaction (PCR) for melanoma-associated antigens. Although sentinel lymph node ultra staging by means of PCR methods is claimed to be associated with participant prognosis (Mocellin 2007), we decided to exclude the PCR arm since this was the only RCT including participants with this type of classification (that is, all the other RCTs enrolled TNM stage III participants exclusively on the basis of pathological evaluation).
Studies awaiting classification
There are no studies awaiting classification.
Two ongoing studies are comparing interferon treatment with new drugs: ipilimumab (ECOG E1609), an anti-CTLA-4 monoclonal antibody, and imatinib (NCT01782508), a c-KIT inhibitor. Another trial is investigating the effect of interferon in participants with ulcerated primary melanomas (EORTC 18081).
Risk of bias in included studies
Figure 2 summarises the risk of bias for each included study. Although detailed descriptions of random sequence generation and allocation concealment were often missing, the quality of the available evidence was generally good except for Rusciani 1997.
|Figure 2. 'Risk of bias' summary: review authors' judgments about each 'Risk of bias' item for each included study|
Random sequence generation (selection bias)
Information regarding random sequence generation was unclear in nine studies. In the other eight studies, we judged this item at 'low risk of bias'. In Rusciani 1997, the risk of selection bias was high as it was unclear whether participants were randomly assigned to control or treatment. Because of this lack of clarity, the authors were contacted by phone and confirmed the randomised design, but omitted information regarding the possible risk of selection bias.
Seven of the eligible studies under-reported allocation concealment, so we judged the risk of this type of bias as 'unclear'. In the other 10 studies, we judged this item at 'low risk of bias'. In Rusciani 1997, the risk of selection bias was high, as we have reported above.
Trials were not conducted in a blinded fashion, neither for the participants nor for the investigators. However, this is not expected to affect the primary aims of the study, i.e. disease-free and overall survival, although some concerns may arise for other end points, such as toxicity and quality of life.
Incomplete outcome data
The risk of attrition bias was low for the majority of the studies. Missing outcome data were balanced in numbers across the intervention and observation groups, with similar reasons for missing data across groups.
The main outcomes of the studies were disease-free survival, overall survival, or both: One or both of these outcomes have been investigated according to the study design of each trial. In particular, two trials did not report data about the effect of interferon alpha on participant overall survival as they had been designed by the investigators to address only the issue of disease-free survival (Kirkwood 2001a; Pehamberger 1998). Finally, the above-mentioned Rusciani 1997 did not report any survival data (i.e. time-to-event data).
Other potential sources of bias
Since we did not identify any other potential sources of bias, we judged all the trials as at low risk of bias for this domain.
Effects of interventions
Disease-free survival (DFS)
All 17 eligible trials (Agarwala 2011; Cameron 2001; Cascinelli 2001; Creagan 1995; Eggermont 2005; Eggermont 2008; Garbe 2008; Grob 1998; Hancock 2004; Hansson 2011; Kirkwood 1996; Kirkwood 2000; Kirkwood 2001; Kirkwood 2001a; Kleeberg 2004; McMasters 2008; Pehamberger 1998), enrolling 10,345 participants, reported data on DFS.
The meta-analysis of these 17 RCTs ( Analysis 1.1) showed a significant risk reduction (17%) in participants undergoing postoperative interferon compared to those undergoing observation (15 trials) or other adjuvant therapy (HR = 0.83; CI 0.78 to 0.87; Z-test P value < 0.0001). Between-study heterogeneity was not statistically significant (I² statistic = 16%; Q-test P value = 0.27).
The 'leave-one-out' sensitivity analysis showed no dominant RCT, which demonstrates that the overall effect estimate is not affected by the findings of a single RCT.
Upon exclusion of the 2 RCTs with an active control arm (Kirkwood 2001; Kirkwood 2001a), the results were very similar to those obtained including all 17 RCTs (HR = 0.84; CI 0.79 to 0.89; Z-test P value < 0.0001; I² statistic = 2%, Q-test P value = 0.43).
As shown in Figure 3, a small study effect was likely to be present (Begg test P = 0.17; Egger test P = 0.03), and the 'trim & fill' procedure revealed that 6 studies might be missing: Their addition to the meta-analysis would reduce, but not nullify, the overall effect of interferon on DFS (adjusted HR = 0.87; CI 0.83 to 0.92). On the other hand, upon exclusion of RCTs enrolling fewer than 400 participants (Cameron 2001; Creagan 1995; Kirkwood 1996; Kirkwood 2001a; McMasters 2008; Pehamberger 1998), the meta-analysis yielded results very similar to those obtained including all RCTs (HR = 0.85; CI 0.80 to 0.90; Z-test P value < 0.0001; I² statistic = 20%; Q-test P value = 0.25).
|Figure 3. Funnel plot of comparison: 1 Interferon alpha versus any other comparator; outcome: 1.1 Disease-free survival (DFS)|
Finally, although we found no significant heterogeneity, we used subgroup analysis and metaregression to investigate whether some study design features might have affected trial outcomes. We report the results of all these analyses in Table 3 (for metaregression, see also Figure 4 and Figure 5). Overall, we found that none of the assessable factors we considered (interferon dose, TNM stage, year of publication, and treatment duration) significantly affected the impact of interferon on participants' DFS.
|Figure 4. Metaregression: interferon effect (log hazard ratio, Y-axis) on disease-free survival versus treatment duration (months, X-axis). See Table 3 for statistical details. Each circle represents a randomised controlled trial (the circle size is proportional to the trial weight)|
|Figure 5. Metaregression: interferon effect (log hazard ratio, Y-axis) on disease-free survival versus publication year (year, X-axis). See Table 3 for statistical details. Each circle represents a randomised controlled trial (the circle size is proportional to the trial weight)|
Overall survival (OS)
Fifteen out of 17 eligible trials (Agarwala 2011; Cameron 2001; Cascinelli 2001; Creagan 1995; Eggermont 2005; Eggermont 2008; Garbe 2008; Grob 1998; Hancock 2004; Hansson 2011; Kirkwood 1996; Kirkwood 2000; Kirkwood 2001; Kleeberg 2004; McMasters 2008), enrolling 9927 participants, reported data on OS.
The meta-analysis of these 15 RCTs ( Analysis 1.2) showed a significant risk reduction (9%) in participants undergoing postoperative interferon compared to those undergoing observation (14 trials) or other adjuvant therapy (HR = 0.91; CI 0.85 to 0.97; Z-test P value = 0.003). Between-study heterogeneity was not statistically significant (I² statistic = 6%; Q-test P value = 0.38).
Considering a 5-year OS rate of 60% for participants with TNM stage II–III cutaneous melanoma (Balch 2009), the number needed to treat (NNT) is approximately 35 participants (95% CI = 21 to 108 participants). This means that, given a group of 35 participants with high-risk melanoma treated with interferon, after 5 years of follow-up, 22 participants would be alive; in contrast, in a group of 35 participants not treated with interferon, 21 participants would be alive.
The 'leave-one-out' sensitivity analysis showed no dominant RCT, which demonstrates that the overall effect estimate is not affected by the findings of a single RCT.
Upon exclusion of the RCT with an active control arm (Kirkwood 2001), the results were very similar to those obtained including all 15 RCTs (HR = 0.92; CI 0.86 to 0.98; Z-test P value = 0.01; I² statistic = 0%; Q-test P value = 0.46).
As shown in Figure 6, a small study effect was likely to be present (Begg test P = 0.07; Egger test P = 0.06); however, the 'trim & fill' procedure suggested that no study is missing. On the other hand, upon exclusion of RCTs enrolling fewer than 400 participants (Cameron 2001; Creagan 1995; Kirkwood 1996; McMasters 2008), the meta-analysis yielded results very similar to those obtained including all RCTs (HR = 0.92; CI 0.86 to 0.98; Z-test P value = 0.012; I² statistic = 19%; Q-test P value = 0.26).
|Figure 6. Funnel plot of comparison: 1 Interferon alpha versus any other comparator; outcome: 1.2 Overall survival (OS)|
Finally, although we found no significant heterogeneity, we used subgroup analysis and metaregression to investigate whether some study design features might have affected trial outcomes. We report the results of all of these analyses in Table 4. According to subgroup analysis, interferon is not associated with an OS benefit in participants with AJCC TNM stage III melanoma (that is, in participants with regional lymph node metastasis), whereas the OS advantage is maintained in trials including both participants with stage III and those with stage II disease (that is, without lymph node metastasis). Only one trial enrolled exclusively AJCC TNM stage II participants and reported OS data, which did not enable us to perform subgroup analysis in this subset of participants. It should also be noted that the heterogeneity analysis did not show any significant difference between the summary hazard ratios (HR) of the three groups of trials (i.e. those enrolling participants with AJCC TNM stage II, stage III, and stage II-III, respectively). With regard to interferon dose, subgroup analysis suggested that interferon is not associated with an OS benefit in participants treated with high- or intermediate-dose, whereas the OS advantage is maintained in trials enrolling participants treated with low-dose interferon. Also in this case, the heterogeneity analysis did not show any significant difference between the summary HR of the three groups of trials (i.e. those treating participants with high-, intermediate-, or low-dose interferon, respectively). A metaregression showed neither year of publication nor treatment duration appeared to have any significant impact on the magnitude of treatment effect ( Table 4).
Summary of main results
The results of this meta-analysis support the efficacy of adjuvant interferon alpha (interferon) for the treatment of people with high-risk (AJCC TNM stage II-III) cutaneous melanoma in terms of both disease-free survival (DFS) and - to a lesser extent - overall survival (OS). In fact, the risk reduction associated with interferon administration was statistically significant for both DFS (17%, 95% CI 13 to 22%) and OS (9%, 95% CI 3 to 15%) ( Summary of findings for the main comparison).
The results of subgroup analysis failed to answer the questions of whether some treatment features (i.e. dosage, duration) might have an impact on interferon efficacy or whether some participant subgroups (i.e. with or without lymph node positivity) might benefit more from this postoperative adjuvant therapy.
High-grade toxicity (i.e. Grades 3 and 4) was observed in a minority of participants: In some trials, no-one had fever or fatigue of Grade 3 severity, but in other trials, up to 8% had fever and up to 23% had fatigue of Grade 3 severity. Less than 1% of participants had fever and fatigue of Grade 4 severity. Although it impaired quality of life, toxicity disappeared after treatment discontinuation.
Overall completeness and applicability of evidence
Although this meta-analysis showed that interferon is associated with a better survival, it failed to identify the optimal interferon dosage and schedules. The available evidence is unsuitable to fully address the issue of the impact of interferon on the quality of life of the treated participants.
Interferon dosage and schedules
While the studies included in this meta-analysis are sufficient to prove the effect of interferon in reducing risk of disease progression and death among high-risk people, it is still necessary to identify optimal treatment dosages and schedules and to improve patient selection. With regard to interferon dosage, findings from our subgroup analyses (see Table 3 and Table 4) indicate that interferon is equally effective in terms of DFS when different doses (i.e. high versus low versus intermediate) are administered.
In contrast, when it comes to OS, the therapeutic benefit remains statistically significant only when considering trials that used low-dose interferon. This finding is in contrast with the common belief that high-dose interferon is the most effective, which has led the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) to approve high-dose interferon and pegylated interferon. This is supported by the results of a RCT addressing this issue, which compared high-dose versus low-dose interferon, and showed that the low-dose interferon regimen was associated with advantage in DFS, but not in OS (Kirkwood 2000). Our findings appear to challenge this but leave the debate open, since the low statistical power might be the reason why the subgroup analysis of our meta-analysis failed to detect a significant risk reduction: In other words, the number of RCTs needed to detect a relatively small OS advantage (such as that observed in this meta-analysis) might be higher than the number of RCTs currently available.
It is important to remember that some RCTs (Hansson 2011; Kirkwood 1996) directly addressed this issue by randomly assigning participants to different interferon doses, but we had to exclude the following from this review as they did not fulfil the inclusion criteria: Chiarion-Sileni 2011; Grob 2010; Hauschild 2009; Pectasides 2009. However, none of them demonstrated that the classical E1684 regimen (i.e. high-dose induction + low-dose maintenance) (Kirkwood 1996) yields results different from the E1684 maintenance phase alone (Hauschild 2009) or that an intensified high-dose interferon regimen (4 courses of induction phase every other month) is any better than the classical E1684 full regimen (1-month induction + 11-month maintenance) (Chiarion-Sileni 2011), or that the E1684 full regimen is superior to the E1684 induction phase alone (Pectasides 2009). The results of these trials cannot be pooled in a meta-analysis since the study designs are too heterogeneous.
Analogously, we could not define whether the duration of interferon treatment impacts on its efficacy: In fact, metaregression analysis did not show any significant relationship between interferon treatment duration and DFS or OS benefit (see Table 3 and Table 4). The results of the EORTC trial published in 2005 (Eggermont 2005) showed that intermediate-dose interferon prolonged DFS (as compared to observation) only when administered for 2 years (and not for 1 year), although no effect on OS was observed. Nevertheless, four other RCTs that addressed this issue did not detect any statistically significant difference between the following:
- the original E1684 full regimen (high-dose for 1 month + low-dose for 11 months) versus an intermittent schedule (3 courses of high-dose for 1 month repeated every 4 months) (Mohr 2008);
- intermediate-dose interferon for 1 year compared to 2 years (Hansson 2011);
- a low-dose interferon regimen for 18 months versus the same for 60 months (Hauschild 2010); or
- low-dose interferon for 18 months versus a similar regimen (pegylated interferon 100 mcg subcutaneously once weekly) for 36 months (Grob 2010).
Also, the results of these trials cannot not be pooled in a meta-analysis since the study designs are too heterogeneous.
Finally, we could not exhaustively answer the question of whether interferon treatment is more effective in a specific disease stage (i.e. AJCC TNM stage II (negative lymph nodes) versus stage III (positive nodes)). In fact, considering disease-free survival as the end point, interferon appeared to be equally effective. However, when overall survival was taken into consideration, RCTs enrolling exclusively stage III participants showed no survival advantage ( Table 4), whereas RCTs indiscriminately enrolling participants with stage II or stage III disease did demonstrate an overall survival benefit ( Table 4), and the only RCT exclusively enrolling stage II participants was positive in terms of overall survival (Grob 2010). Overall, our findings might suggest that interferon adjuvant therapy is more effective in stage II melanoma patients, but the evidence supporting this hypothesis is too scarce to draw any definitive conclusion.
Toxity and quality of life
Severe adverse events (Grades 3 and 4 according to the scale adopted by the World Health Organization) after treatment with interferon occurred in the minority of the included studies ( Table 5), with Grade 3 fever (0% to 8%) and fatigue (0% to 23%) being the most frequently observed adverse events.
Side-effects such as these impair quality of life (QoL) as reported by some other studies that investigated quality of life after treatment with interferon (Bottomley 2009 based on Eggermont 2008; Ziefle 2011a based on Hauschild 2010; and Brandberg 2012 based on Hansson 2011). Results favoured the treatment with interferon over observation in terms of quality of life-adjusted survival (Cole 1996 based on Kirkwood 1996; Kilbridge 2001 and Kilbridge 2002 based on both Kirkwood 1996 and Kirkwood 2000; and Caraceni 1998 based on Cascinelli 2001). In the pivotal trial of the Eastern Cooperative Oncology Group, the E1684 (Kirkwood 1996), Cole et al studied the effect of quality of life on patient survival and showed that although interferon improved disease-free and overall survival by 9 and 7 months, respectively, it caused 6 months of significant toxicity (Cole 1996). However, participants treated with interferon had a better quality of life-adjusted survival than participants who were observed only, regardless of their relative valuations for toxicity and disease-free survival. This difference was statistically significant only for participants who considered toxicity to have a high relative value and disease relapse to have a low relative value, but not for participants who valued toxicity the same as relapse. Another study conducted on participants of the E1684 trial compared the quality of life of participants who experienced severe toxicity during treatment with interferon (Grade 4) with that of participants who had a recurrence in the observation arm, showing that participants rated quality of life with melanoma recurrence lower than even severe toxicity (Kilbridge 2001).
A quality of life-adjusted survival analysis conducted in a subgroup of participants enrolled in 2 of the included studies (Kirkwood 1996; Kirkwood 2000) suggested that interferon may improve quality of life-adjusted survival for some, but only a minority (16%) showed a statistically significant benefit. The change in quality of life-adjusted survival depends more on the measure for interferon toxicity than on the measure for melanoma recurrence (Kilbridge 2002). Taken together, these data showed that although interferon alpha negatively affects quality of life, there is an improvement in quality of life-adjusted survival for those who undergo treatment with interferon. The occurrence of toxicity and its negative effect on quality of life can be managed as suggested by expert authors (Bottomley 2009; Hauschild 2008), and it is reversible when the treatment is stopped (Brandberg 2012).
Of note, differences in assessment of quality of life prevented meta-analysis of quality of life measures. Currently, there is still a lack of a melanoma-oriented questionnaire that could account for changes in quality of life in subsets of patients (Cormier 2012). Beside the cancer questionnaire QLQ-C30 of the European Organisation for Research and Treatment of Cancer (EORTC), only two QoL questionnaires have been designed specifically for people who have melanoma: the FACT-Melanoma, which does not include evaluation of psychological domains, and the Malignant Melanoma Module, which has not been validated yet (Cormier 2011; Winstanley 2013). A new instrument for assessing quality of life of people with melanoma, but which is not addressing interferon and adjuvant treatment specifically, is under evaluation within the Quality of Life Group of the EORTC.
Quality of the evidence
This body of evidence allows us to draw robust conclusions regarding the impact of interferon alpha on disease-free and overall survival of high-risk melanoma patients. We included 18 studies, with 10,499 participants: Of these, we were able to perform meta-analyses for disease-free survival on 10,345 participants from 17 studies and overall survival on 9927 participants from 15 studies; 2 studies had not been designed to investigate overall survival, and 1 included study of 154 participants did not assess either outcome. The evidence we collected comes from well-designed randomised controlled trials (RCTs) most often conducted by international organisations, e.g. the European Organization for Research and Treatment of Cancer (EORTC), and the Eastern Cooperative Oncology Group (ECOG). Therefore, the quality of the evidence meta-analysed in the present work can be considered high ( Summary of findings for the main comparison).
Potential biases in the review process
There was a generally low risk of bias in all the eligible studies, but one (Rusciani 1997) that was not suitable for inclusion in the meta-analysis. Selection bias was difficult to ascertain in roughly half of the studies as no information was reported in the manuscripts. Finally, no trial was conducted in a blinded fashion. Although this is not supposed to affect survival, which is the primary end point of all studies, issues may arise regarding toxicity and quality of life.
Agreements and disagreements with other studies or reviews
Our findings support the efficacy of adjuvant interferon alpha (interferon) for the treatment of people with high-risk (AJCC TNM stage II-III) cutaneous melanoma in terms of both disease-free survival (DFS) and, to a lesser extent, overall survival (OS). These results represent a significant update of the existing literature on this subject: In fact, compared to previous meta-analyses, the present work includes 3 new RCTs (Agarwala 2011; Hansson 2011; McMasters 2008) and can count on the survival data of 10,345 participants.
Out of three previous meta-analyses of adjuvant interferon for melanoma patients (Pirard 2004; Mocellin 2010; Wheatley 2003), the first two reported a disease-free survival advantage but not an overall survival benefit (Pirard 2004; Wheatley 2003). In particular, Pirard and colleagues (Pirard 2004) considered nine trials including the two studies with severe drawbacks in their design, which we excluded from the present meta-analysis (see Excluded studies). These investigators measured the risk of both disease recurrence and death by calculating odds ratios (OR: the ratio between odds, where "odds" indicates the ratio between events and non-events in a given participant group): This is an inappropriate method to study survival data (also called time-to-event data), which are better described by hazard ratios (HR: the ratio between the risks of an event occurring over a given time span, where "hazard" indicates the risk in a given participant population). This is especially true when considering the OS data because in the long-term, all participants of both comparison arms will die of disease or any other cause; thus, no therapy-induced benefit could ever be detected using statistics based exclusively on the number of events.
In the second full-text meta-analysis, Wheatley and colleagues (Wheatley 2003) used time-to-event data and found a DFS (but not OS) advantage including 12 RCTs, but not the 2 trials we judged ineligible (Kokoschka 1990; Rusciani 1997). For three included RCTs (Eggermont 2005; Hancock 2004; Kleeberg 2004), these authors could not use the definitive data because the articles reporting them were published after their meta-analysis was published. Later on, the same authors published an abstract describing an individual patient data meta-analysis on the same subject (Wheatley 2007), considering 13 RCTs and 6067 participants; these investigators showed a benefit in terms of both DFS (OR = 0.87; 95% CI 0.81 to 0.93) and OS (OR = 0.9; 95% CI 0.84 to 0.97).
Finally, in a meta-analysis published in 2010, Mocellin and colleagues considered 14 RCTs (including 2 trials published after the last work of Wheatley and colleagues (Eggermont 2008; Garbe 2008) and including 8122 participants: They also found that interferon significantly improves both DFS (risk reduction = 18%) and OS (risk reduction = 11%) of participants with high-risk cutaneous melanoma (Mocellin 2010).
Implications for practice
Our findings support the efficacy of adjuvant interferon alpha (interferon) for the treatment of participants with high-risk (AJCC TNM stage II-III) cutaneous melanoma in terms of both disease-free survival and, to a lesser extent, overall survival. High-grade toxicity was observed in a minority of participants, who reported impairment in their quality of life. Nevertheless, participants reported relief from symptoms and improvement in quality of life after treatment discontinuation.
Until better selection methods or more effective therapies are available, the findings of the present meta-analysis lend support to the use of interferon in the routine clinical setting to provide patients with the best chance of survival.
Implications for research
The results reported above cannot be fully satisfying in terms of antimelanoma efficacy. In fact, considering a 5-year OS rate of 60% for participants with TNM stage II–III cutaneous melanoma (Balch 2009), the number needed to treat (NNT) is approximately 35 participants (95% CI 21 to 108 participants). Therefore, much more effort is clearly warranted to improve these results. This might be achieved in three main ways.
A) By identifying participants who actually need adjuvant treatment:
B) By elucidating the molecular mechanisms underlying melanoma responsiveness to interferon:
C) By developing new treatments that are more effective than interferon against minimal residual disease.
Until better selection methods or more effective therapies are available, the findings of the present meta-analysis lend support to the use of interferon in the routine clinical setting to provide patients with the best chance of survival. Moreover, we must remember that other well-established adjuvant treatments, such as those routinely administered to people with breast, colorectal, and ovarian carcinomas, are associated with risk reductions very similar to those found in this meta-analysis for those with high-risk melanoma treated with interferon (Ascierto 2008). Therefore, the need for better therapeutic strategies is an urgent issue for virtually all tumour types.
Finally, the findings of this meta-analysis support interferon alpha as the reference treatment in RCTs investigating new therapeutic agents for the adjuvant treatment of high-risk melanoma participants. One such trial, comparing interferon and ipilimumab in AJCC TNM stage III-IV melanoma participants after surgery (an immunostimulating agent acting by inhibiting CTLA-4), has been designed by ECOG and is currently ongoing (ECOG E1609, clinicaltrial.gov identifier: NCT01274338).
We truly thank our data manager (Dr Marta Briarava) for her help in retrieving scientific articles and for setting up the database we used to collect and analyse the data from clinical trials. We also thank the Cochrane Skin Group, and in particular Prof Hywel Williams, Miss Laura Prescott, and Dr Finola Delamere for their assistance.
The Cochrane Skin Group editorial base wishes to thank Dedee Murrell who was the Key Editor for this review; Jo Leonardi-Bee and Philippa Middleton who were the Statistical and Methods Editors, respectively; the clinical referees, Eleni Linos and Rubeta Matin; and the consumer referee, Kathie Godfrey.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Appendix 1. CENTRAL (Cochrane Library) search strategy
#2 (interferon*) or (ifn alpha) or (ifn alfa)
#3 MeSH descriptor Interferons explode all trees
#4 MeSH descriptor Interferon-alpha explode all trees
#5 MeSH descriptor Melanoma explode all trees
#6 (#1 OR #5)
#7 (#2 OR #3 OR #4)
#8 (#6 AND #7)
Appendix 2. MEDLINE (OVID) search strategy
1. randomized controlled trial.pt.
2. controlled clinical trial.pt.
5. clinical trials as topic.sh.
8. 1 or 2 or 3 or 4 or 5 or 6 or 7
9. (animals not (human and animals)).sh.
10. 8 not 9
11. exp Melanoma/
13. 11 or 12
15. exp Interferons/ or exp Interferon-alpha/
16. (interferon alpha or interferon alfa).mp.
17. 14 or 15 or 16
18. 10 and 13 and 17
Appendix 3. EMBASE (OVID) search strategy
3. (crossover$ or cross-over$).mp.
4. placebo$.mp. or PLACEBO/
5. (doubl$ adj blind$).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer]
6. (singl$ adj blind$).mp. [mp=title, abstract, subject headings, heading word, drug trade name, original title, device manufacturer, drug manufacturer]
7. (assign$ or allocat$).mp.
8. volunteer$.mp. or VOLUNTEER/
9. Crossover Procedure/
10. Double Blind Procedure/
11. Randomized Controlled Trial/
12. Single Blind Procedure/
13. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12
14. exp MELANOMA/
16. 14 or 15
18. (interferon alpha or interferon alfa).mp.
19. exp INTERFERON/
20. exp alpha interferon/
21. 17 or 18 or 19 or 20
22. 13 and 16 and 21
Appendix 4. AMED (OVID) search strategy
1. randomized controlled trial$/
2. random allocation/
3. double blind method/
4. single blind method.mp.
5. exp Clinical trials/
6. (clin$ adj25 trial$).mp. [mp=abstract, heading words, title]
7. ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$ or mask$ or dummy)).mp. [mp=abstract, heading words, title]
8. (placebo$ or random$).mp. [mp=abstract, heading words, title]
9. research design/ or clinical trials/ or comparative study/ or double blind method/ or random allocation/
10. prospective studies.mp.
11. cross over studies.mp.
12. Follow up studies/
14. (multicent$ or multi-cent$).mp. [mp=abstract, heading words, title]
15. ((stud or design$) adj25 (factorial or prospective or intervention or crossver or cross-over or quasi-experiment$)).mp. [mp=abstract, heading words, title]
16. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15
17. exp Melanoma/
18. melanoma$.mp. [mp=abstract, heading words, title]
19. 17 or 18
20. exp Interferons/
21. interferon$.mp. [mp=abstract, heading words, title]
22. 20 or 21
23. 16 and 19 and 22
Appendix 5. LILACS search strategy
melanoma$ and (interferon$ or ifn$)
Last assessed as up-to-date: 22 August 2012.
Contributions of authors
SM was the contact person for the editorial base, co-ordinated the contributions from the co-authors, analysed the data, and wrote the final draft of the review.
ML and SP contributed to the methods section of the protocol and inputted the data into RevMan.
PP checked the protocol for readability, clarity, and for relevance to consumers.
SM and VCS are the guarantors of the final version of the review.
Declarations of interest
Sandro Pasquali, Marko B Lens, Pierluigi Pilati, and Simone Mocellin have all declared they have no conflicts of interest.
Vanna Chiarion-Sileni was reimbursed for consulting activity as a melanoma expert on advisory boards for GSK, Roche, BMS, and SMD. This was always done with the approval of her institution and outside her work time. She received compensation for advisory board participation from Merck, Roche Genentech, Bristol-Meyers Squibb, GlaxoSmithKline MSD for the drugs ipilimumab, dabrafenib, vemurafenib, trametinib, and anti-Pd1. She was the medical leader of the following excluded study:
Chiarion-Sileni V, Guida M, Romanini A, Ridolfi R, Mandala M, Del Bianco P, et al. Intensified high-dose intravenous interferon alpha 2b (IFNa2b) for adjuvant treatment of stage III melanoma: A randomized phase III Italian Melanoma Intergroup(IMI) trial [ISRCTN75125874]. [Abstract 8506] ASCO Annual Meeting 2011. Journal of Clinical Oncology 2011;29(15 Suppl 1):8506.
Sources of support
- University of Padova, Padova, Italy.
- No sources of support supplied
Differences between protocol and review
The final version of this review differs from the original protocol for a number of reasons:
- First of all, the critical revision of two authors (VCS (oncologist) and ML (dermatologist)) suggested a number of changes in order to make the text clearer to non-melanoma experts. This required the rephrasing of many sentences throughout the text. Moreover, a new author (VCS), who was not involved at the stage of writing the protocol, has joined the team of this work at the review stage and has brought her viewpoint (she is an oncologist) on several aspects regarding this subject.
- Secondly, all authors have added new references and replaced others in order to provide readers with the best and most updated information on this subject.
Although there were 3 ongoing studies listed in the last published review, a search of MEDLINE and Embase in 2014 found some relevant results, and this is a rapidly moving field, this review has currently been deemed not in need of an update because the review team assessed the aforementioned search results and found that only 1 trial (Eggermont AM, Suciu S, Testori A, Santinami M, Kruit WH, Marsden J, et al. Long-term results of the randomized phase III trial EORTC 18991 of adjuvant therapy with pegylated interferon alfa-2b versus observation in resected stage III melanoma. Journal of Clinical Oncology 2012;30(31):3810-8. (DOI: 10.1200/JCO.2011.41.3799)) would be eligible for the update. However, the review authors deemed that the data from this article were just an update of a trial that they had already included in the review (Eggermont 2008), with the findings being virtually identical to the original findings. Thus, an update has not been considered necessary at this time. Our Trials Search Co-ordinator will run a new search towards the end of 2015 to re-assess whether an update is needed.
Medical Subject Headings (MeSH)
Antineoplastic Agents [*therapeutic use]; Chemotherapy, Adjuvant [methods; mortality]; Disease-Free Survival; Interferon-alpha [*therapeutic use]; Melanoma [*drug therapy; mortality; surgery]; Randomized Controlled Trials as Topic; Skin Neoplasms [*drug therapy; mortality; surgery]
MeSH check words
* Indicates the major publication for the study