The idea of perfusing a limb with one or more cytotoxic agents for treatment of cancer of an extremity was advanced first in the 1950s. The procedure is performed under general anesthesia. The axillary, external iliac, or femoral vessels are cannulated, a tourniquet is applied to the limb affected, and the extremity is perfused with an agent or agents designed to destroy cells of the cancer. Circulation in the isolated limb is maintained by connecting the arterial and venous cannulas to an external pump, oxygenator, and exchanger of heat.

Franklin, et al.,¹⁰² in 1988, undertook an analysis retrospectively of 465 patients with melanoma of an extremity but with no sign of metastasis, of whom 227 had had isolated limb perfusion with melphalan and 238 had not. All patients had a wide excision of the melanoma and about 20% of patients from each group also had END. Analysis of sets failed to reveal any increase in the time between treatment and the appearance of metastasis clinically in the limb itself, in regional nodes, or at distant sites, or an increase in survival overall. The authors acknowledged flaws inherent in an analysis undertaken retrospectively and they did that as follows: " . . . a trial [of adjuvant isolated limb perfusion] . . . will have to include a large number of patients in order to show differences in prognosis, which from these results cannot be expected to be very great."

Koops, et al.,¹⁰³ in 1998, studied 832 patients with melanoma of an extremity. Although the melanoma in every patient was thicker than 1.5 mm, there was no evidence clinically of metastasis. All patients had undergone wide local excision of the melanoma and some of them had had an END. Some patients were selected randomly for treatment by isolated limb perfusion with melphalan and others received no other treatment at all. The median period for follow up was 6.4 years. An analysis of sets revealed that the patients who had received melphalan were less apt to develop a metastasis between the site of the melanoma and the regional nodes (metastasis in transit) than patients who did not receive that treatment. The finding pertained only to patients whose melanoma measured between 1.5 to 2.99 mm in thickness. There was no evidence that the entire study group or any set of it benefitted from the procedure in terms of survival overall, which motivated the authors to make this statement: " . . . prophylactic ILP [isolated limb perfusion] with melphalan cannot be recommended as an adjunct to standard surgery in high risk primary limb melanoma."

The results compiled by Koops and coworkers, in 1988, confirmed the impression of Franklin, et al., namely, that isolated limb perfusion was of no benefit to patients who probably had metastasis already. Not only did the therapy not help, but patients who received it suffered severe side effects, among them erythema, edema, pain, nerve injury, and arterial or venous thrombosis, the latter sometimes necessitating amputation of a limb.

The lack of utility of isolated limb perfusion was stated unwaveringly in 2000 by Eggermont⁴¹ as follows: "Prophylactic ILP [isolated limb perfusion] should no longer be performed. It is a harmful procedure with significant morbidity and costs and without any impact on survival." Newton Bishop, et al.,¹⁰⁴ in 2002, concurred with that conclusion tersely thus: "There is no role for adjuvant isolated limb perfusion . . . "

Authors of 17 separate reports devoted to adjuvant therapy with decarbazine and/or BCG, C. Parvum, or levamisole have written that it is useless. Five communications concerning a sizable trial of "active specific immunotherapy" with either vaccines of melanoma cells or viral lysates of melanoma cells told of total lack of influence of any of those regimens.⁴¹ As of the time of writing, there is no compelling argument in support of a contrary point of view. The interferons, however, continue to be advocated in some quarters and for that reason they will now be discussed in extenso.

The interferons are a family of drugs now being marketed as adjuvant therapy effective for metastasis of melanoma. At first, only low doses of interferon were given and when no benefit was found to derive from them, higher doses were administered with nearly invariable debilitating side effects that, almost always, required that the dose be reduced. Side effects include toxicity to bone marrow, a flu-like illness, fever, chills, lethargy, headache, myalgia, nausea, weight loss, and, rarely, even death.

In 1994, Cascinelli, et al.¹⁰⁵ published preliminary results of a study performed by them as part of a WHO program, that study consisting of 426 patients with melanoma, all of whom had undergone complete dissection of the regional nodal basin because of proven metastasis there. Two hundred and eight of those patients also received interferon in low dose (recombinant interferon a-2a, 3 million units injected subcutaneously three times a week for three years). After a median period of follow up of 19 months, patients who received interferon enjoyed a longer interval free of disease, that is, devoid of any sign of a "new" metastasis, than those who did not (p=0.01) receive interferon. Women younger than 51 years and men older than 50 years seemed to benefit most from interferon. In both sets, survival overall and length of time without a "new" metastasis increased. Although the report portrayed interferon as having promise, in 2001 Cascinelli and coworkers,¹⁰⁶ having studied the patients for a longer period, informed that no benefit for survival could be attributed to interferon for any patient who had received it.

In 1995, Creagan, et al.¹⁰⁷ told of their study of 264 patients with a melanoma thicker than 1.69 mm or a melanoma of any thickness that had metastasized to nodes suspected of harboring cells of melanoma clinically and actually proven to do that histopathologically. None of the metastases was discovered as a result of END. One hundred and thirty one patients were treated with interferon a-2a administered in high dose (20 million units/m2 injected into the muscle three times a week for 12 weeks) and 131 patients did not receive interferon. Two patients were excluded from the study. Patients who received interferon did not have a better survival overall than those who were not given it and they did not have any increase in the length of time during which no manifestation of metastasis of melanoma was detectable clinically. The set of patients who had metastasis of melanoma to nodes and who had received interferon, however, did seem to have a very slight increase in the time it took for another metastasis to become apparent clinically (p=0.04). This negligible increase was recognized only after what was termed "adjusted" analysis. This prompted the following statement issued tentatively by the authors as they attempted to integrate the results of their study: " . . . [treatments with interferon] suggest a possible benefit for selected patients . . . " The authors warned additionally that the results which appear to indicate a delay in the development of a "new" metastasis following administration of interferon may be skewed by virtue of problems inherent to the analysis of sets.

In 1996, Kirkwood, et al.¹⁰⁸ showed that interferon a-2b, given in high doses, can be advantageous to patients with presumed or proven metastasis of melanoma. They studied patients who had a melanoma thicker than 4.0mm and no evidence of nodal metastasis after END or a melanoma of any thickness that had metastasized to regional nodes as demonstrated in sections of tissue stained only by hematoxylin and eosin. This latter group was represented by patients who had 1) an END, 2) nodes that were palpable clinically at the time of wide local excision, and 3) nodes that became positive some time after wide local excision had been performed. None of the patients were found to have metastasis beyond the nodes, at least as that phenomenon was assessed clinically. In each of the patients, the nodal basin was removed in its entirety. Median survival for patients who had no apparent metastasis beyond the nodes clinically and who received interferon in high doses increased from 1.0 to 1.7 years and survival overall increased from 2.8 to 3.8 years. The effect was more dramatic for those patients who, at the outset, had nodes that contained a metastasis of melanoma, but not those whose melanoma was thicker than 4.0 mm. This led the authors to this inference about the group of patients who had undergone END: "The apparent difference of outcome of IFN [interferon] and observation in this relatively small subgroup suggests that the effect of IFN treatment on microscopic metastatic disease may be worthy of further evaluation." In every patient, interferon caused side effects sufficient to mandate reduction in the dosage of it in all patients at some point in the study.

The study of Kirkwood and coworkers was the first randomized and prospective one that appeared to demonstrate benefit for patients who received interferon in high doses. On the basis of evidence presented by Kirkwood, et al., the Food and Drug administration (FDA) of the United States granted approval for marketing and use of Interferon a-2b as accepted adjunctive therapy in patients with metastasis of melanoma detectable clinically in nodes only or with a melanoma greater than 4.0mm in thickness. Taking the lead from the FDA, interferon a-2b was licensed in the United Kingdom in 1977 as adjuvant therapy for melanoma.

De Takats, et al.,¹⁰⁹ in 1998, discussed possible benefits of interferon in an article titled, "Adjuvant therapy for melanoma: how should we respond to high-dose interferon?" In it they expressed this view, guardedly, of the matter: "From current data available from these studies, the value of adjuvant IFN-a remains uncertain." The studies to which the authors referred included those of Cascinelli, et al.,¹⁰⁶ Creagan, et al.,¹⁰⁷ and Kirkwood, et al.¹⁰⁸

In 1998, Grob, et al.¹¹⁰ studied 489 patients with a melanoma thicker than 1.5 mm and regional nodes that were negative for melanoma clinically. None of the patients underwent END or SNB. All of them had had wide local excision of the melanoma. Two hundred patients received interferon a-2a in low dose (3 million units three times a week for 18 months). After a median period of follow up of five years, patients who received interferon were found to be less likely to have a "new" metastasis of melanoma for up to three years from the onset of treatment (p= 0.035). After three years, however, metastasis of melanoma was discovered as often in patients who had been treated with interferon as in patients who had not received the drug. Although survival overall in the two groups of patients was similar, patients who received interferon did seem to live longer (p= 0.059). Eleven of 248 patients who took interferon developed serious adverse affects from it and, as a consequence, withdrew from the trial; 24 patients withdrew because of adverse effects that did not seem to be terribly severe to the investigators. The authors adopted this curious stance: "Our findings show that a low-dose regimen can favourably impact the course of primary melanoma when patients are treated before the development of clinical node metastases."

In 2000, Kirkwood, et al.,¹¹¹ in an effort to confirm the results of their study published in 1996, conducted another trial with interferon a-2b that was given in high doses. The criteria by which patients became eligible for this trial were similar to those employed by them in 1996. Unlike the trial in 1996, however, patients with a melanoma thicker than 4.0 mm were not required to undergo END. Patients were randomized into three groups, namely, ones that received high dose interferon, low dose interferon, or no interferon. The study was designed to detect differences in the length of time that elapsed before the development of a "new" metastasis as determined clinically and to gauge survival overall. The results of the study were assessed prematurely, the reason for that being stated by the collaborators as follows: "This decision was reached based on estimates of the protracted time that would be required to reach the original targets and the low likelihood of demonstrating a significant improvement in OS [overall survival] for either treatment arm [no interferon or interferon in low or high doses]." All of this brought Kirkwood, et al. to the following conclusion: "The results of the intergroup E1690 trial demonstrate an RFS [recurrence free survival, that is no additional metastasis clinically] benefit of IFN a-2b that is dose-dependent and significant for HDI [high doses of interferon] . . . " In other words, patients treated with interferon in high dose lived for longer periods of time during which no "new" metastasis became apparent clinically. Those patients, however, did not live longer than patients who merely were observed or who had received interferon in low dose. An interesting observation made by the authors, after comparing the results of their study in 1996 to the current one, was this: there was a significant increase statistically in terms of survival for patients who just were observed in the more recent study than for those in the earlier one. The authors sought to explain the reasons why the results of their present study differed from those they had obtained in 1996. According to them, patients in the second study who did not receive interferon lived longer than did patients in the first study because of significant differences statistically in the two studies. The authors were uncertain about which factors made for the differences and about the absence of any benefit in regard to survival for those who had received interferon in high dose. They postulated that the differences may have resulted from factors that could not be controlled, such as the selection of patients. Just as was the case in the earlier study, severe toxic effects of interferon were encountered in the more recent study. In sum, the results of the study indicated overwhelmingly that benefit for patients who received interferon was negligible at best and in those patients who received high doses of the drug unassociated with any lengthening of time of survival.

In 2000, Eggermont⁴¹ made a revealing comment about studies concerning adjuvant therapy, including interferon, to this effect: "No standard systemic adjuvant therapy with confirmed impact on overall survival has been identified for clinically node negative . . . patients after excision of the primary [melanoma], nor for clinically node positive . . . patients after lymph node dissection for metastatic regional node involvement."

In May 2001, Kirkwood, et al.¹¹² presented the results of a study in which they compared the effect of a combination of interferon a-2b in high doses and GM2-KLH/QS-21 (GMK), a ganglioside*** vaccine versus GMK vaccine alone. Of 774 patients eligible for the study, 77% had a metastasis of melanoma in nodes detectable clinically. The other 23% had a melanoma thicker than 4 mm, and 28% of those patients had a negative SNB, and the rest of those patients (72%) appearing to have a negative nodal basin as judged clinically. After a median period of follow up of 16 months, analysis of the results indicated that patients who received a combination of interferon and GMK vaccine had a better outcome, overall, than those who received the vaccine only. The results at 16 months were extrapolated to two years and the revised results showed once again that patients who received both interferon and the vaccine had a better outcome than those who received vaccine alone. During the period of two years there was no clinical evidence of a "new" metastasis in 62% of the patients who received combination therapy. Only 49% of patients who received just vaccine seemed to remain free of a "new" metastasis clinically. The survival rate for two years was 78% in patients who received the combination of drugs, whereas only 73% of those patients who had vaccine alone survived for that period of time. Analysis of sets revealed that patients without nodal metastasis derived benefit from the combination of interferon and vaccine tor survival overall statistically (P=0.46). This is how the authors expressed the results of the study: " . . . [this trial] demonstrated the greatest benefit and the only independently statistical significant benefit, in node negative subset," that is, patients who had no metastasis of melanoma to nodes survived longer if they received both the vaccine and interferon, however for those with metastasis the interferon had no effect on the length of survival.

A serious limitation of the study of Kirkwood, et al. published in 2001 was that it did not include patients who did not receive interferon nor vaccine, that is, a cohort that was just observed. The coworkers defended that criticism by claiming that patients who received GMK vaccine only were equivalent to patients who were observed only. The reason for that assertion was that those patients seemed to have an outcome similar to patients in equivalent trials who did not receive adjuvant treatment. In brief, this study showed that there was a limited benefit overall, measured over two years, for patients who received interferon. That aid was restricted to patients with a melanoma thicker than 4.0 mm and with no nodal involvement, only 23% of the patients, but it did not apply to patients whose nodes were found to be involved at the outset of the study, that is 77% of the patients.

Spitler¹¹³ criticized the study of Kirkwood, et al. in these lines: "(1) the median follow-up was short for this patient population (16 months); (2) there was no control group given observation alone; (3) the difference in overall survival in the two arms was not great (78% for IFN alfa-2b and 73% for GMK); and (4) the vaccine regimen was changed between the positive phase II trial and phase III trial . . . " In other words, for Spitler the study of Kirkwood, et al. lacked substance because the period of follow-up was too short and no proper observation group was available with which to compare the results. Moreover, the data was not deemed to be compelling in favor of or against either regimen.

In July 2001, Eggermont,¹¹⁴ in a section in the European Journal of Cancer captioned "Current perspective," wrote this: "The role of interferon-alpha [as adjuvant therapy] in malignant melanoma remains to be defined." Furthermore, what had emerged from all of the studies, in his view, was "a confusing picture." He pointed out problems in assessing and comparing results of clinical trials that employed interferon, among those problems being that there were not enough patients studied and cohorts of patients that were studied could not be compared because their melanomas were at different stages of development and/or the treatment administered differed both in the frequency and dose of interferon.

All of this brought Eggermont to the following position: "IFNa has only modest activity in melanoma . . . In the adjuvant setting, results are heterogeneous and inconsistent, reflecting its modest activity, which overall translates into an impact on relapse-free survival, which can be seen with various doses of IFNa, but no proof of a significant impact on survival."

Retsas,¹¹⁵ in November 2001, analyzed the design of studies of adjuvant therapy referred to inter alia here, the interpretation of the data, and the conclusions reached by those who employed a particular therapy. His analysis did not include the "final" results of the study of Kirkwood, et al. concerning the effect of interferon in high doses in combination with vaccination of GMK, but it did address the preliminary results of that study as they were presented at the 25th ESMO Congress in Hamburg.

These analyses prompted Retsas to make the following observations: "First, that there is currently no standard adjuvant treatment that unequivocally prolongs overall survival. Second, that prolongation of freedom from disease seems possible with currently available treatments . . . Third, any demonstrable benefit is apparent in patients with higher, rather than a lower risk of relapse." He ended his article with these unvarnished words: "There is no evidence that patients with clinically occult metastatic regional lymph nodes found at elective dissections or with scintigraphic detection of the sentinel node derive significant benefit from adjuvant or any other systemic treatment. The same applies to patients with high risk primary lesions without demonstrable regional lymph node metastases." In brief, Retsas declared that no credible evidence warrants the use of interferon in patients whose melanoma is thicker than 4.0 mm or in those with nodal metastasis confirmed by END or SNB.

Spitler¹¹³ offered this opinion, straightforwardly, about the matter of interferon: "Alpha-interferon is the only agent approved for adjuvant therapy of melanoma in the United States, but as the benefits are not great and the side effects are substantial, it is hard to justify recommending this therapy to patients."

Last, Newton-Bishop, et al.,¹⁰⁴ in 2002, presented guidelines drawn, by consensus, from a multidisciplinary working group and coordinated by both the Melanoma Study Group and the British Association of Dermatologists. This is the statement, just as it was made by them: "Although interferon is now licensed for adjuvant use in the UK, a second confirmatory trial with mature data is considered necessary before this regimen can be recommended as standard treatment, not least because of the side effects commonly experienced."

In sum, despite assertions of investigators, such as Kirkwood and co-workers, to the contrary, all the data available in regard to adjuvant therapy for patients with primary or metastatic melanoma testifies, overwhelmingly, to the fact that not a single adjuvant therapy is effective, consistently and reliably.