• © 2007 by American Society of Clinical Oncology

In Reply

We appreciate the comments by Drs Lejeune and Eggermont regarding our reporting of the results from the American College of Surgeons Oncology Group (ACOSOG) trial Z0020 in the Journal of Clinical Oncology.¹ They raise three major concerns regarding the way the trial results were reported. They questioned the use of 3 months as an end point for defining response rates. They imply that “this is an uncommon end point for a treatment where most CRs are only reached between month 3 and month 6 and where reports in the literature have not reported CR rates at 3 months.” Unfortunately, there are no uniform ways to present data from regional therapy trials that convey both a sense of efficacy as well as durability. Interestingly, in their own published work, and in contrast to their comments quoted herein, Drs Lejeune and Eggermont have demonstrated that the addition of tumor necrosis factor (TNF) to a melphalan-based perfusion markedly decreases the time to partial and complete responses (14 and 61 days, respectively).² As such they monitor response at 4 weeks, 8 weeks, and then at 3-month intervals.³ In their studies with TNF, they have categorized three groups of complete responders defined as achieving a complete response within 6 weeks as fast, achieving a complete response within 12 weeks as intermediate, and achieving a complete response after 4 months as slow responders.⁴ This latter group makes up the minority, not majority of patients. Nevertheless, it was this observation that led us to include the 6-month results in our article, even though it was not an original end point of the study. The clinical end points utilized in the ACOSOG study are very similar to the WHO criteria of response frequently used by Lejeune and Eggermont in their published reports,^2,4 except that the response duration in the ACOSOG trial needs to have a durability of at least 8 weeks not 4 weeks.

The authors wanted a clarification of patient stratification related to tumor burden in each arm of the trial and why patients dropped out of the study relative to each treatment arm. High tumor burden was defined as any tumor larger than 3 cm in diameter or more than 10 tumor lesions within the region of perfusion. No patients with stage IV disease were eligible for the ACOSOG trial. While patients in the melphalan plus TNF arm were older and more likely to have high tumor burden, the distribution of sex and race was evenly distributed. Table 1 shows the tumor burden stratification and previous regional treatment in each arm. In comparing arms, no statistically significant difference in disease progression was observed at 3 months based on tumor burden, prior perfusion, regional nodal status, sex, optimal hyperthermia by 30 minutes, hospital volume, or type of extremity (upper v lower). Several patients did drop out of the analysis before the 3- and 6-month reviews as summarized in Table 2. Roughly equal numbers of patients were excluded from both arms.

Finally as experienced clinical trialists, Drs Lejeune and Eggermont certainly recognize that it is not at all unusual for the results of a randomized multi-institutional trial to have response rates that are lower than large single institutional studies. This can occur for a multitude of reasons but usually is related to differing levels of experience in performing the treatments among the various institutions. In light of this, the ACOSOG Z0020 trial could not convincingly validate that the addition of TNF to a melphalan-based perfusion justified the toxicity associated with its administration. However, as we clearly state in the last paragraph of our article, which we hope everyone will read, there may be subgroups of patients (such as those with bulky disease or those who do not respond to an initial melphalan alone perfusion) who may benefit from the addition of TNF to regional perfusion treatments. Unfortunately, this trial was not powered to address these issues within these specific subpopulations of patients. Certainly for groups that remain enthusiastic about TNF in the regional setting for advanced extremity melanoma there is always the opportunity to address these issues in a more focused randomized prospective trial in the future.