Biopsy specimen of skin with Kaposi sarcoma showing numerous tumor cells occupying the entire subcutaneous layer (hematoxylin-eosin, ×200).
Cross section of a Kaposi sarcoma lesion 6 weeks after isolated limb perfusion with tumor necrosis factor and melphalan showing necrosis and dense inflammatory infiltration with no evidence of tumor cells (hematoxylin-eosin, ×200).
Lev-Chelouche D, Abu-Abeid S, Merimsky O, Isakov J, Kollander Y, Meller I, Klausner JM, Gutman M. Isolated Limb Perfusion With High-Dose Tumor Necrosis Factor α and Melphalan for Kaposi Sarcoma. Arch Surg. 1999;134(2):177-180. doi:10.1001/archsurg.134.2.177
Copyright 1999 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.1999
Although the classic form of Kaposi sarcoma is considered indolent and benign, at times its evolution is more severe, with an acute onset and debilitating complications necessitating aggressive treatment and even amputation.
To evaluate the efficacy of hyperthermic isolated limb perfusion (ILP) with tumor necrosis factor α and melphalan as a limb-sparing modality for extensive regional Kaposi sarcoma.
University hospital and national referral center.
Five patients, aged 60 to 82 years, with extensive, symptomatic, classic Kaposi sarcoma of the lower limb were operated on. All were candidates for amputation owing to debilitating symptoms.
Patients underwent ILP through the iliac (n=2), femoral (n=2), and popliteal (n=1) vessels. Tumor necrosis factor α, 4 mg, and melphalan, 1.5 mg/kg body weight, were perfused for an overall time of 90 minutes. The limb was heated to 40°C. Clinical and pathological responses were recorded for all patients after 6 to 8 weeks.
The overall response rate was 100%: 1 of 5 patients had complete response and 4 of 5 had partial response. Two patients had progression of disease 2 months after ILP but one of them was asymptomatic and did not require any further treatment. The second patient underwent amputation. Thus, limb preservation was achieved in 80% (4 of 5 patients). Median follow-up was 24 months. There were no deaths associated with treatment or major system complications. Local complications were all reversible.
These findings suggest that hyperthermic ILP with tumor necrosis factor α and melphalan can be considered an effective palliative and limb-sparing treatment modality for extensive Kaposi sarcoma.
THE INTEREST in Kaposi sarcoma has escalated during the last decade because it has become the most frequent malignancy associated with the acquired immunodeficiency syndrome. However, the classic form of this disease was first described more than 100 years ago by Kaposi as a rare multifocal neoplasm. This form is an infrequent cutaneous angiolymphoproliferative disease that usually involves the lower limbs, and mainly affects people of Mediterranean, Eastern European, or Jewish heritage.1 It occurs most frequently in men after the sixth decade of life. Unlike other forms, classic Kaposi sarcoma usually has a chronic course following an indolent and benign clinical course. It is not generally life-threatening, and thus may not require specific therapy early in its course. However, at times it can result in local aggressive lesions, causing pain, edema, ulceration, and bleeding, which seriously impair the quality of life.2,3
Currently, treatment options are confined to old modalities, mainly radiotherapy and various chemotherapeutic agents, and some new experimental therapies based on biomodulators such as interferons.4- 7 In numerous studies, some of these treatment options achieved palliation but this was mostly temporary at the expense of major toxic effects. Because the population most commonly afflicted with this form of Kaposi sarcoma is elderly and often affected with concomitant disease (mainly of the cardiovascular and respiratory systems), and because toxicity must be kept to a minimum, it seems an ideal setting for using local treatments.8 To date, different local therapies have been used for the treatment of Kaposi sarcoma, including simple excision, cryotherapy, laser treatment, and intralesional therapy, and, although these therapies are easy to perform, relatively safe, and often sufficient for a very localized small tumor burden, they are not a feasible treatment for an extensive disease with dozens of lesions. Kaposi sarcoma is a highly radiosensitive neoplasm and thus local radiation has been widely used for control of individual problem-causing lesions with complete regression rates of up to 80% to 90%.9,10 However, recurrences are frequent, and this modality is limited and cannot be used repeatedly. At times, failure of all other treatment options makes amputation inevitable.
Isolated limb perfusion (ILP) is a form of regional therapy that enables the administration of chemotherapeutic agents at a dose of up to 20 times the maximally tolerated systemic dose while avoiding exposing the entire body to major side effects of these agents. High-dose recombinant tumor necrosis factor α (TNF-α) administered in combination with the cytotoxic drug melphalan via ILP resulted in remarkable response rates for melanoma and nonresectable soft tissue sarcoma confined to the limb, with a limb-sparing rate of more than 90%.11- 13
Because Kaposi sarcoma can evolve into a major local problem in the inflicted limb, this study evaluated the role of ILP with TNF-α and melphalan in the palliation and limb preservation for such patients.
Between June 1994 and November 1997, 5 patients with advanced Kaposi sarcoma were treated with TNF-α and melphalan via ILP (Table 1). All were men, with an age range of 60 to 82 years (mean age, 73 years). All had very extensive recurrent disease confirmed histologically (Figure 1). Except for 1 patient with 1 very large lesion located on the ankle that was not amenable for resection, the remaining 4 patients all had more than 20 lesions. All patients had symptoms due to their Kaposi sarcoma: pain (5 patients), bleeding (2 patients), and infected necrotic lesions (2 patients). Previous therapy had failed in all patients, with either a short-lived response or no response at all: radiotherapy (4 patients), systemic chemotherapy (2 patients), and photodynamic therapy (1 patients) At the stage when they were referred to us, they were all considered absolute candidates for some form of amputation.
The study was conducted within the framework of an "investigator's own responsibility study," which was approved by the National Helsinki (Finland) committee. Informed consent was obtained from all patients.
The technique of hyperthermic ILP with high-dose recombinant TNF-α and melphalan is described elsewhere.13 Briefly, to isolate the limb, the main artery and vein are dissected and all collaterals ligated. The vessels are cannulated and connected to a pump oxygenator similar to that used in cardiopulmonary bypass. A tourniquet (200-400 mm Hg) or Esmarch band is applied to the root of the limb to ascertain complete vascular isolation and to ensure that all tumors lie within the perfused area and receive adequate drug dosing. The temperature of the perfused limb is maintained at 39° to 40°C during the entire procedure by using both external heating and warming of the perfusate.
Systemic leakage from the perfused limb is monitored continuously by using technetium Tc 99–radiolabeled human serum albumin injected into the perfusate. Radioactivity above the precordial area is recorded using a Geiger counter.
Recombinant TNF-α was administered at a dose of 4 mg as a bolus into the arterial line of the perfusate. Melphalan (Alkeran, Glaxco Wellcome, Research Triangle Park, NC) was administered at 1 mg/kg body weight 30 minutes after recombinant TNF-α administration and perfusion was continued for a further 60 minutes.
Evaluation of response according to Union Internationale Contre le Cancer criteria was performed 6 to 8 weeks after ILP and included clinical response assessed by counting and measuring the lesions and a pathological estimation assessed by multiple incisional biopsy specimens from the remaining tumors or tumor bed (Figure 2).
Patients underwent ILP via the iliac (n=2), femoral (n=2), and popliteal (n=1) vessels (Table 2).
There were no deaths associated with the treatment. None of the patients experienced any form of systemic complication. Local limb complications assessed according to Wieberdink et al14 manifested as redness in 3 patients (grade 2) and blisters in 4 (grade 3). No patient experienced neuropathy and none necessitated amputation.
The overall response rate was 100% (5/5). One patient (20%) achieved complete response and 4 patients (80%) achieved partial response. Within the limitations of a short follow-up (3-42 months; median, 24 months), local recurrence occurred in 2 patients within 2 months. One patient required amputation due to recurrent lesions that became necrotic and infected. In the second patient, the lesions reappeared but because they were asymptomatic, to date, he has not required any further treatment. Thus, limb preservation was achieved in 4 of 5 patients who were previously all candidates for amputation.
One patient died during the follow-up (36 months) of a myocardial infarction.
Tumor necrosis factor α has long been recognized for its antineoplastic properties, which have been demonstrated repeatedly in various in vitro and in vivo studies.15,16 The effect of TNF-α on tumor vasculature, and its role as an important mediator in the immune system generating many different kinds of cytokines and mediators, makes Kaposi sarcoma an ideal target for its antitumoral effect. Kaposi sarcoma is assumed to be a neoplasm of vascular origin and thus contains numerous amounts of endothelial cells and blood vessels. Using its procoagulant properties,17- 19 TNF-α activates specific adhesion molecules on the surface of endothelial cells within the tumor and membranal receptors on macrophages and leukocytes. Rapid migration of leukocytes to the tumor capillary bed leading to adhesion to the endothelium, release of oxygen free radicals and other cytokines, and extensive local inflammatory reaction may result in obstruction of the tumor's capillary bed. This TNF-α–induced endothelial damage is exclusive to tumor vasculature, whereas normal vasculature is spared.
A link between the development of Kaposi sarcoma and a suppressed immune system was suspected for a long time since it occurs more commonly in conditions of iatrogenic immunosuppression,20 in transplant recipients,21 in patients with hematological malignant neoplasms,22 and in the elderly who have a more deficient immune system. This speculation has received major support with the emergence of human immunodeficiency virus (HIV)–associated Kaposi sarcoma. Tumor necrosis factor α has the ability to modulate the immune functions and induce the generation of many cytokines,23,24 such as interleukin 1, interleukin 6, interleukin 8, interferon, oxygen free radicals, and arachidonic acid products, some of which have a definite anticancer effect.
However, systemic administration of TNF-α cannot be clinically implemented for patients with Kaposi sarcoma in the same way as for patients with other malignant disease since the dose required to achieve its anticancer effect is 10 times higher than the maximal tolerated systemic dose in humans. Immediately on its systemic administration, TNF-α leads to marked hypotension and septic shocklike syndrome and eventually to multiorgan failure.25 This effect provided the rationale for a trial evaluating direct intratumoral injection of TNF-α,26 and, although regression of lesions was observed in 94% of cases, systemic toxicity was common and limited the clinical usefulness of TNF-α.
Our growing experience of ILP with TNF-α and melphalan for treatment of patients with melanoma and soft tissue sarcoma,13 and the understanding that TNF-α in high doses might achieve a major response and thus limb preservation in patients with Kaposi sarcoma who were candidates for amputation, led us to implement this method.
All 5 patients responded to the treatment, with more than 50% regression of the tumor mass. Although this is only a small group of patients, it is essential to remember that all had very extensive and mutilating disease, and that these results were obtained despite the fact that previous treatment modalities had failed, such as chemotherapy and irradiation. Four of the 5 patients, including 1 in whom the tumor recurred a few months after treatment, achieved limb salvage and are asymptomatic. There was no evidence of leakage of TNF-α to the systemic circulation due to meticulous isolation of the limb and thus no systemic side effects of TNF-α administration. These results are encouraging and suggest that hyperthermic ILP with TNF-α and melphalan can be considered as an effective treatment for limb salvage in patients with extensive Kaposi sarcoma not limited by major side effects.
Considering the entire procedure, it cannot be postulated that the TNF-α is the only effective factor since it is given in combination with the alkylating agent melphalan. However, from past experience with other malignant tumors such as soft tissue sarcoma, ILP with chemotherapeutic agents alone resulted in a low response rate, and the addition of TNF-α showed marked improvement. Moreover, there is a clear synergism between TNF-α and various cytotoxic drugs, among them melphalan.27
The other factor is hyperthermia. Throughout the procedure, the limb is heated to about 40°C. Several studies in patients with HIV demonstrated regression of disseminated lesions of Kaposi sarcoma when treated with low-flow extracorporeal perfusion hyperthermia.28,29 Hyperthermia by itself is also known to potentiate the activity of melphalan and TNF-α and may be partly responsible for the observed efficiency of the treatment.27,30
The fact that 2 of our patients had relapse of disease early after the procedure might indicate that at least in some patients control of disease is only temporary, and an additional treatment modality should be considered to achieve a sustained response.
One possibility is to exchange the chemotherapeutic drug melphalan with one of the vinka alkaloids or antibiotics, which have a known effect on Kaposi sarcoma.31
Recently, biomodulators such as interferons are being studied for their effect on Kaposi sarcoma.7,32,33 The interferons are a group of naturally occurring proteins that inhibit the growth of tumors in vivo and of many transformed cells in vitro. Limited information is available to assess the efficacy of interferons of different types, mainly interferon alfa, at different doses, in the treatment of Kaposi sarcoma. However, these studies do demonstrate therapeutic benefits of such treatment.
Combining TNF-α with interferon appears logical, since TNF-α itself potentiates the action of different interferons. These and other biological agents in combination with TNF-α via ILP should be studied in the near future.
Liposomal daunorubicin hydrochloride, a newly described agent, shows promising primary results in the treatment of classic and HIV-associated Kaposi sarcoma.34,35 Further studies are needed to compare its efficacy with that of ILP and TNF-α or even to combine both treatment modalities for a long-standing response.
The long-term effect of ILP and TNF-α in patients with Kaposi sarcoma is as yet unknown; however, it seems to be a safe and effective treatment for classic Kaposi sarcoma, and might even be considered a palliative method for HIV-related Kaposi sarcoma.
Reprints: Mordechai Gutman, MD, Department of Surgery, Tel Aviv Sourasky Medical Centre, 6 Weizmann St, Tel Aviv 64239 Israel.