| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
for Advanced Melanoma and Soft-Tissue SarcomaDepartment of Surgery, Sarcoma and Melanoma Unit, The Royal Marsden Hospital, Fulham Road, London SW3 6JJ, United Kingdom
Correspondence: Address correspondence and reprint requests to: Andrew J. Hayes, MD, PhD; E-mail: andrew.hayes{at}rmh.nhs.uk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
may increase response rates for bulky melanoma and for sarcoma, but the potential for major systemic toxicity has limited its use.
Methods: Between October 2000 and April 2004, 49 ILPs were performed with melphalan and TNF- . All procedures were performed with continuous leakage monitoring and regional hyperthermia.
Results: Forty-nine ILPs were performed for melanoma (n = 30), sarcoma (n = 16), or other tumors (n = 3). The most common indications were widespread in-transit disease for melanoma (n = 29) and irresectable primary disease for sarcoma (n = 9). Complete and partial responses for melanoma were 40% and 37%, and for sarcoma they were 20% and 33%. At a median follow-up of 14 months, 66% of melanoma patients who responded had not experienced local progression, compared with only 37% of sarcoma patients. Progression-free survival was significantly less for patients with sarcoma than melanoma (P = .0476). Four of 16 patients with sarcoma subsequently required amputation for progressive disease.
Conclusions: ILP with melphalan and TNF- is a valuable treatment for advanced in-transit melanoma. Significant response rates were also seen in irresectable sarcoma, although the duration of response was limited.
Key Words: Isolated limb perfusion • Sarcoma • Melanoma • Tumor necrosis factor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
ILP with melphalan used prophylactically in high-risk extremity melanoma has not been shown to have any effect on the development of distant disease or overall survival and has shown only a minimal decrease in the rate of development of in-transit disease.4 In contrast, when ILP with melphalan is used therapeutically for advanced melanoma, response rates ranging from 48% to 95% have been reported.5–10
Tumor necrosis factor (TNF- ) is a cytokine with direct and indirect antitumoral effects. Its effects may be mediated by a specific destructive effect against the tumor vasculature that synergizes with the cytotoxic effect of melphalan.11,12 Because TNF- is a key physiological mediator of the systemic inflammatory response, systemic administration at doses with an antitumor effect has profound, and potentially lethal, side effects. Hence, it can only be used clinically in ILPs, and there is an absolute requirement for continuous monitoring of leakage from the perfusion circuit.13
It has been suggested on the basis of several case series that melphalan plus TNF- ILP has better complete and overall response rates that melphalan-only ILP performed at equivalent temperatures,14–19 although recent early data from randomized studies suggest that there may be no benefit in the addition of TNF- over melphalan-only perfusion.20 The response rates for extremity sarcoma treated by ILP with a variety of cytotoxic agents (including melphalan and doxorubicin) are low.21–24 The series of perfusions for sarcoma with TNF- in conjunction with cytotoxic agents (principally melphalan) have reported considerably improved partial (PR) and complete response (CR) rates over cytotoxic agents alone.25–29 As a consequence of these encouraging case series in which limb salvage rates in the order of 80% to 90% were reported for tumors viewed as irresectable other than by amputation, melphalan/TNF- ILP has been adopted with some enthusiasm on mainland Europe in the management of extremity sarcoma, although this practice has not been reflected either in the United Kingdom or North America.
We report a single-institution experience of 49 consecutive melphalan/TNF- ILPs for advanced irresectable extremity soft tissue malignancy.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Operative Approach
All patients were weighed before surgery. Perfusions were performed with patients under general anesthesia, and patients were prewarmed in the anesthetic room under anesthesia by using heated blankets (Bair Hugger; Arizant Healthcare, Prairie, MN) for approximately 45 minutes. Limb isolation was achieved by a pneumatic tourniquet high in the upper thigh for femoral perfusions or an Esmarch bandage secured about a Steinman pin placed into either the anterior superior iliac spine or humeral head for iliac or brachial perfusions, respectively. Before tourniquet inflation, all patients received anticoagulation therapy with heparin (200–300 U/kg). A calibration dose of 1 MBq of isotope (99mTc human serum albumin) was injected into the systemic circulation, and baseline isotope counts were obtained by using a precordial scintillation probe. After vascular cannulation and tourniquet inflation, the isolated limb was connected to an extracorporeal bypass circuit primed with 5000 U of heparin, compound sodium lactate solution, and one unit of cross-matched red blood cells.
Drug Infusion
Typical flow rates were calculated at .18% of cardiac output per minute for femoral perfusions and .09% of cardiac output per minute for brachial perfusions. The hematocrit during perfusion was maintained at >25% to maintain optimal tissue perfusion. Regional hyperthermia of 38.5° C to 39.5° C was confirmed by subcutaneous temperature probes placed in the perfused limb. Once satisfactory flow rates were established, a 10-MBq dose of isotope was injected into the perfusate of the isolated circuit. When a stable limb circulation with adequate flow rate, regional hyperthermia, and an acceptable leak rate had been confirmed, TNF- (tasonermin, Beromun; Boehringer Ingelheim, Ingelheim am Rhein, Germany) was injected into the perfusion circuit. Fifteen minutes after injection of TNF- , melphalan (Alkeran; Glaxo-Smithkline, Middlesex, UK) was infused into the circuit (total dose over 20 minutes), and then the perfusion was continued for a further 40 minutes before the cytotoxic agents were washed out. The dose of melphalan used was 1.0 mg/kg body weight for lower limb perfusions and .5 mg/kg for upper limb perfusions. The dose of TNF- was 1 mg for upper limb perfusions and 2 mg for lower limb perfusions. PaO2 in the oxygenated circuit was maintained >400 mm Hg (50 kPa) throughout the perfusion to enhance tumor cell cytotoxicity. At the end of the perfusion period, the drugs were washed out of the perfusion circuit with approximately 3 L of crystalloid solution, until the venous line was clear. The tourniquet was then released, and protamine was administered to reverse the anticoagulation.
Isotope Leakage Monitoring
All perfusions were performed with continuous leakage monitoring by the use of a radioisotope administered to the isolated limb; systemic radioactivity was constantly monitored by a gamma probe over the precordium. Leak rates were calculated by using the activity ratio of both isotope doses and the ratio of blood volumes in the systemic and isolated circulations.13 Leakage monitoring was continued for 20 minutes after cessation of bypass. Leakage rates of <5% were considered acceptable, and perfusion was continued, whereas leakage of >10% was an indication for cessation of the procedure.
Postoperative Protocol
After ILP, patients were admitted to the high-dependency unit overnight. Neurovascular observations were performed on a regular basis and decreased in frequency when stability was confirmed. When stable, patients were transferred to the ward (usually the following morning). Patients were encouraged to mobilize on day 2 after the procedure and were discharged when mobilizing independently. All patients were followed up in the outpatient clinic 6 weeks after the procedure for an assessment of response and at 12-week intervals thereafter. Standard World Health Organization criteria were used for classifying response rates, with CR defined as complete clinical or radiological resolution of disease, PR defined as clinical or radiological evidence of disease reduction by 
50%, and an objective response defined as that not amounting to a PR. The remaining patients were categorized as having either stable disease or progressive disease.
Statistical Analysis
Response rates for patients with sarcoma and melanoma were compared by using the Mann-Whitney U-test. Curves for progression-free survival were constructed by using the Kaplan-Meier method, and the log-rank test was used to assess the statistical significance of differences in progression-free survival.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The sex distribution was approximately equal. In patients undergoing treatment for melanoma, all patients but one underwent the procedure for in-transit disease, at a median of 47 months (range, 0–310 months) after primary treatment. More than 50% of patients had undergone previous treatments of in-transit disease by surgical excision (with or without skin grafting) or laser ablation. In contrast, 50% of patients with sarcoma underwent ILP as primary treatment for irresectable disease, with most of the remainder (n = 6) undergoing ILP for locally recurrent disease at the primary site after an initial resection at another institution. Of the 16 patients who underwent ILP for sarcoma, 11 patients had grade 3 tumors, 3 had grade 2 tumors, and 2 had grade 1 tumors.
|
was 2 mg for lower limb perfusions. (In the first few cases when experience with the drug was limited and in two patients in whom the leak rate was approaching the upper limit of the acceptable range, a reduced dose of 1 mg was used.) For upper limb perfusions, the dose of TNF- was 1 mg. The dose of melphalan was 1 mg/kg for lower limb perfusions and .5 mg/kg for upper limb perfusions. The goal perfusion temperature was between 38.5° C and 39.5° C, which was achieved in 90% of procedures. There were 39 superficial femoral perfusions, 3 popliteal perfusions, and 5 brachial perfusions. Three patients with metastatic melanoma underwent inguinal node dissection performed 7 days after the femoral perfusion for synchronous disease in the inguinal node basin. Two iliac perfusions were attempted, but both were unsuccessful. In one patient, encasement of the iliac vessels with melanoma prevented successful vascular cannulation, and in the other, bulky tumor in the proximal thigh caused venous compression and prevented adequate flow rates.
Complications
The complications of the perfusions are listed in Table 2. Local tissue reactions were classified from grade 1 to grade 5.30 Only one patient had a grade 5 acute tissue reaction (tissue destruction mandating amputation). This 83-year-old patient with mild peripheral vascular disease presented with a massive irresectable primary sarcoma occupying the entire posterior compartment of the calf. The limb was rendered nonviable after the procedure because of necrosis of the entire posterior compartment, which left an unreconstructable deficit. (The foot and toes remained viable and well perfused, thus suggesting that the tissue injury was a direct antitumor effect rather than a cannulation-associated vascular injury.) No patient had grade 4 reactions (deep tissue damage), but three patients had grade 3 reactions (severe erythema, edema, and skin blistering); grade 2 reactions (mild erythema and edema) were common.
|
. The response rates for sarcoma and melanoma have been calculated from 42 patients out of 49 consecutive perfusions in the entire series. Complications were recorded for the entire series of 49 perfusions. The seven perfusions not included in the survival analyses were excluded for the following reasons. Three patients underwent perfusions for different tumor types: one patient who underwent two successful perfusions for a malignant eccrine poroma initially had a CR and on the second occasion had a PR, and one patient with an advanced squamous cell carcinoma underwent a successful perfusion resulting in a PR. Four patients were excluded from the analysis on technical grounds. Two patients with melanoma had technically unsuccessful perfusions such that adequate TNF could not be administered. One patient with a massive calf sarcoma after a successful perfusion had a dramatic response causing such extensive necrosis to the calf that an amputation was mandated in the immediate postoperative period. One patient with melanoma underwent a successful perfusion but defaulted from follow-up after the procedure, so no outcome data are available.
|
Figure 1 shows clinical photographs of a patient with in-transit melanoma before and after perfusion, demonstrating a CR that has been ongoing. Figure 2 shows a patient with an isolated fungating metastatic deposit of sarcoma before and after perfusion that resulted in an excellent PR. Of the 26 patients who underwent technically successful perfusions for melanoma without distant metastases at the time of perfusion, 9 have subsequently developed distant metastases, and 5 have died of distant disease within the follow-up period. None of these patients underwent a palliative amputation after perfusion. Of the 15 patients who underwent technically successful perfusions for primary or locally recurrent sarcoma with no evidence of distant disease at the time of perfusion, only 1 has subsequently developed distant metastases. Four of these patients have required amputations for progressive local disease.
|
|
shows Kaplan-Meier curves for progression-free (local disease) survival after ILP for patients with sarcoma and melanoma. With the log-rank test, there was a significant difference between the two curves (P = .0476; hazard ratio, 2.47; 95% confidence interval, .89–6.84). The CR, PR, stable disease, and progressive disease rates within the two groups were, however, not significantly different (P = .143; Mann-Whitney U-test).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
has been less clearly defined, and there is considerable variation in the pattern of use between Europe and North America. ILP has been widely used for extremity melanoma and sarcoma for many years, both in the adjuvant and therapeutic settings.31 Initial reports in the literature were usually single-institution case series.5–9 However, definitive randomized studies confirmed that there was no benefit in terms of distant disease–free survival or overall survival for adjuvant ILP after resection of primary cutaneous melanoma, even though there was a small, but demonstrable, reduction in the development of local disease within the limb.4 Consequently, because the procedure carries a chance of significant morbidity, it is now accepted that ILP should be used only for therapeutic treatment of locally advanced disease.
Reported series of the efficacy of therapeutic ILP for advanced irresectable extremity melanoma indicate CR and PR rates that are considerably greater than those seen either with systemic chemotherapy [with DTIC (dacarbazine)] or radiotherapy.13 Randomized studies of this technique compared with other techniques are therefore essentially unethical, and data on efficacy in the therapeutic setting rely principally on nonrandomized studies.
The most common indication for therapeutic ILP is in the management of widespread in-transit metastases from melanoma that are unable to be controlled by simpler techniques. Reported response rates for melphalan-only ILP vary widely between series and range from 26% to 82% for CRs and 65% to 100% for overall response rates. Other cytotoxic agents have been used, but none has been demonstrated either to be more effective or to have less toxicity.2,3 The second major indication for therapeutic ILP is limb salvage for extremity soft tissue sarcomas that are considered irresectable other than by amputation.
TNF- is a cytokine that has been shown in pre-clinical models to have a direct antivascular tumorigenic effect and to act synergistically with melphalan. As a physiological mediator of septic shock,32 its major dose-limiting systemic effect is hypotension, which precludes its use as a traditional systemic agent. Theoretically, it is ideally suited to combination therapy with melphalan in ILP, where the dose-limiting systemic side effects can be avoided by washing out circulating drugs from the isolated limb circulation before tourniquet release. Some authors have heralded TNF- as a major advance for melanoma and particularly for advanced sarcoma.33 For melanoma, the putative advantage for TNF- inferred from noncontrolled studies was for bulky "sarcoma-like" in-transit metastases.33 However, results at 3 months from a randomized controlled study of patients with advanced extremity melanoma undergoing either melphalan-only perfusion or TNF- and melphalan perfusion indicate that response rates are equivalent in the two arms.20 What is not clear from these data (published in abstract form) is whether the patients had principally low-volume cutaneous disease, where it is generally accepted that the addition of TNF- has no benefit over melphalan alone, or whether a significant proportion of patients had bulky disease; it has been postulated that TNF- has a potential synergistic effect on the tumor vasculature, as is seen with sarcoma.
The response rates in this series mirror previous studies, but there are important differences. Our data show overall response rates for melanoma of 77%, with 40% complete clinical response rates; some other series have reported overall response rates as high as 95%.10 Response rates for TNF- melphalan perfusions for advanced soft tissue sarcoma are generally lower than for melanoma, and this is reflected by our own data. However, as with melanoma, there is great variability in the reported response rates, with some small series reporting a CR rate of 70%.27 The largest series is of 186 patients with extremity sarcoma considered irresectable by defined criteria.25 All of these patients subsequently underwent surgical resection of the tumor mass weeks after perfusion. CR rates on clinical criteria were cited as 18%—almost identical to our data. However, when histopathologic response (defined by the extent of necrosis in the tumor remnant) was included, this allowed the authors to upgrade some PRs to CRs on the basis of 100% tumor necrosis. The final cited figures for overall CRs and PRs were 29% and 53%, respectively.25
Why is there such variability in reported response rates for both sarcoma and melanoma in what is essentially a standardized operative technique? Some of the theoretical reasons for disparity are the dose of drugs administered, the classification of tumor response, the preoperative tumor burden, and, finally, the definition of irresectable disease.
We used a dose of 1 mg/kg of melphalan, and most perfusions were performed via the superficial femoral vessels with the tourniquet applied to the upper thigh. The original dose range of melphalan (1–1.5 mg/kg) was for perfusions performed at the iliac level. Hence, the actual concentration of melphalan used in our patients was relatively high because the drug is perfused into a smaller limb volume (our tourniquet was placed at the mid thigh) when compared with patients undergoing iliac perfusions. A variety of doses have been suggested for TNF- , ranging from 1 to 4 mg, but most units now use 
2 mg of TNF- for lower limb perfusions, with a good tumor response being reported with low-dose TNF- (1 mg) in some series.34,35 In >85% of the lower limb perfusions in this series, 2 mg of TNF- was used. Hence, it is likely that the tissue dose of TNF- and melphalan in this series was at least equivalent to that used in series with higher response rates.
Tumor response rates were assessed clinically, according to standard criteria. Very few of the patients in this series went on to have subsequent resections, so pathologic response rates are not included in this series. This is in contrast to some series of patients with sarcoma treated by ILP, in which ILP was used as an induction treatment to downstage the primary tumor.25
The most likely reason for differing results between series is probably different patient populations in terms of preoperative tumor burden. The response rates will be strongly influenced by the extent of disease at the time of perfusion, and this may be influenced by several factors that cannot be controlled between series. For example, for in-transit melanoma, we advocate ILP only when disease can no longer be managed by multiple resections or carbon dioxide laser therapy. We routinely offer carbon dioxide laser ablation even when there are in excess of 100 cutaneous lesions and perform repeat procedures at relatively short intervals because this procedure is so well tolerated. ILP is offered only when the volume or frequency of disease precludes these simpler techniques. This contrasts with some units where ILP is offered at an earlier stage in the disease process.36 The likelihood of inducing a CR is much smaller in a limb affected by a large number of bulky in-transit metastases than in one affected by less advanced disease.
For sarcoma, the potential variation in the patient population in which ILP is performed is even greater. By far the most important variable is whether the patient undergoes a resection of the tumor remnant after ILP or whether the only treatment is the ILP, because resection of the tumor remnant would still require an amputation. In the former case, the ILP is serving as induction chemotherapy before a marginal resection, whereas in the latter case, the response rates are for ILP alone. It is very likely that the limb-salvage rates will be completely different in each scenario.
The classification of a tumor as irresectable (either because of proximity to neurovascular structures or because of likely functional deficit after resection) is open to considerable debate between operating surgeons. A series of 55 patients from 4 European institutions undergoing ILP for irresectable sarcoma included 34 with isolated primary sarcomas, half of which were in the thigh. In contrast, no perfusions were performed for thigh sarcomas in our series, even though the thigh is the most common site for limb sarcoma and our institution receives >400 new sarcoma referrals per year. In our unit, all thigh tumors are resected radically (with plastic surgical reconstruction if necessary) coupled with adjuvant radiotherapy; using this approach, we have reported a 5-year local recurrence rate of rate of 15% and an amputation rate of 4.8%.37 All those who underwent ILP for sarcoma in our series had distal limb tumors and did not undergo a planned enucleation of the tumor remnant. It is therefore difficult to directly compare results between our series and the European series where ILP is used usually as induction chemotherapy before surgical resection.
Our data confirm that ILP with TNF and melphalan is a valuable palliative treatment modality for multiple in-transit metastases from melanoma, a disease that carries a poor prognosis. The median duration of response, although still short, is extremely valuable in a palliative context because the likelihood of developing distant metastases in the same time frame is relatively high.
The role for ILP in the management of soft tissue sarcoma is less clearly defined. Although CR rates of 20% and overall response rates of 53% undoubtedly make this procedure a valuable treatment for truly irresectable disease, some points are worth noting. For sarcoma, ILP is used for treatment of primary disease rather than regionally metastatic disease, and, as a consequence, the likelihood of distant dissemination in the short term is less than with in-transit melanoma. Therefore, because the patients with sarcoma are more likely to live longer than those with melanoma and the duration of effect is in fact shorter, further palliative measures to the primary disease may be required. This is evidenced by the fact that 27% of sarcoma patients still required a palliative amputation even after ILP. In extremity sarcoma, the real challenge is to define the value of ILP in terms of local recurrence rates when it used as induction chemotherapy before a planned surgical resection compared with a more standard approach of primary radical resection and adjuvant radiotherapy, a question that can be definitively answered only in the context of a multicenter randomized study, which presents huge logistical challenges.
Received for publication October 19, 2005. Accepted for publication April 3, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
