Disease Site Group (DSG) Consensus
The draft guideline and systematic review were approved by the Melanoma Disease Site Group in December 2006. One member of the group suggested reversing the presentation of the data in one of the meta-analysis figures to provide a consistent direction of results, i.e., biochemotherapy benefit to the left and chemotherapy benefit to the right. However, since the data summarized in the figures represents a positive outcome in one analysis (tumour response) and a negative outcome in the other analyses (death or progression of disease), the current format of the analysis figures is consistent with the data, and the Evidence Synthesis section of the original report was revised to clarify this.
Report Approval Panel
Prior to submission of this evidence-based series report for external review, the report was reviewed and approved by the Program in Evidence-based Care (PEBC) Report Approval Panel in January 2007. The Panel consists of two members, including an oncologist, with expertise in clinical and methodology issues. Key issues raised by the Panel included whether the biochemotherapy regimens tested in the phase II trials that led to the reported phase III trials were consistent; if variation in response rates across the reported trials may be related to different tumour response evaluation criteria; the appropriateness of using 12-month mortality data for the meta-analysis when the report Introduction indicates that median survival for this patient group is six to eight months; and the inconsistency in the Results section of commenting on a survival trend in the Rosenberg trial, which suggests some benefit with chemotherapy, but not the Eton trial, which suggests some benefit with biochemotherapy. In addition, one member of the panel suggested it would be helpful to have explicit statements about "policy determining" outcomes. Since, in this case, the relevant outcomes appeared to be overall survival, coupled with treatment toxicity, a defining statement was suggested for the recommendation to indicate that.
The authors discussed the Report Approval Feedback and revised the report where appropriate. With regard to the treatment regimens used, an optimum regimen has not been identified for biochemotherapy and the regimens used in both phase II and phase III trials has varied, generally corresponding with specific institution or organization preferences. The criteria used to define a tumour response in most trials were that of the World Health Organization (or a similar definition) and a brief statement was added to the Trial Descriptions section of the Systematic Review to summarize that data. In considering the conduct of the meta-analysis, since the median survival for most of the reported trials was around 11-12 months; it was agreed that this was a reasonable time-point for data pooling. In addition, the six-month survival data were pooled with a similar result to that obtained at 12 months, and this was indicated in the Results section of the report. The need for consistency in presenting data was acknowledged and, following discussion of the results of the Rosenberg trial, a comment on the contrasting results of the Eton trial was added to the text of the Results section of the report. Finally, although the Program in Evidence-based Care guidelines are considered in policy determination, the authors consider their main purpose as providing guidance for clinicians and, therefore, do not wish to comment on policy-determining outcomes. In developing the recommendations, all relevant outcomes were considered, and it was felt that the current wording of the recommendation accurately reflects that fact; therefore, the recommendation was not revised.