Annals of Surgical Oncology, 9(10):10041009

DOI: 10.1245/ASO.2002.03.048

Isolated Limb Perfusion With Tumor Necrosis Factor and Melphalan for Nonresectable Stewart-Treves Lymphangiosarcoma
T. E. Lans, MD, J. H. W. de Wilt, MD, PhD, A. N. van Geel, MD, PhD, and A. M. M. Eggermont, MD, PhD

Background: Cutaneous Stewart-Treves lymphangiosarcomas represent a rare group of tumors characterized by a high grade of vascularization and by localization in an extremity with lymphedema. The multifocality and the localization makes these tumors eligible for treatment with isolated limb perfusion (ILP). ILP with tumor necrosis factor (TNF) and melphalan is a safe and highly effective procedure that can achieve limb salvage in 80% of all patients with nonresectable extremity soft tissue sarcoma or melanoma. Methods: In 10 patients with multifocal Stewart-Treves lymphangiosarcoma of the extremities, 16 ILPs with TNF plus melphalan were performed. All patients would have been candidates for exarticulation of the extremity. Results: We observed an 87% overall response rate (complete and partial responses); one patient had a mixed response, and one patient did not respond to the therapy. In nine perfusions (56%), a complete response was achieved, and five perfusions (31%) resulted in a partial response. Limb salvage was achieved in eight patients (80%), with a mean follow-up duration of 34.8 months (range, 3 to 115 months). Regional toxicity was limited and systemic toxicity minimal to moderate, with no toxic deaths. Conclusions: Multifocal Stewart-Treves lymphangiosarcomas in extremities with chronic lymphedema can be successfully treated by ILP with TNF and melphalan. Key Words: Stewart-Treves lymphangiosarcomaIsolated limb perfusionTumor necrosis factorMelphalanLimb salvage.

Sarcomas are malignant tumors arising from skeletal and extraskeletal connective tissues. The cellular elements of these anatomical structures are diverse, and, as a consequence, so is the natural history of both bone and soft tissue sarcomas (STS). STS have been classified histogenetically according to their morphological resemblance to normal congenate tissues. The best predictor of outcome for STS is the assessment of grade, a measure of
Received March 15, 2002; accepted August 21, 2002. From the Department of Surgical Oncology, Daniel den Hoed Cancer Centre, University Hospital Rotterdam, Rotterdam, The Netherlands. Address correspondence and reprint requests to: A. M. M. Eggermont, MD, PhD, University Hospital Rotterdam-Daniel den Hoed Cancer Centre, P.O. Box 5201, 3008 AE Rotterdam, The Netherlands; Fax: 31-10-439-1011; E-mail: eggermont@chih.azr.nl.
Published by Lippincott Williams & Wilkins 2002 The Society of Surgical Oncology, Inc.

differentiation and aggressivity. In Europe, the incidence of STS is approximately 5000 cases per year, and a little more than 50% of these new patients will go on to die as a result of the disease. Most STS have no clearly defined etiology, and there is a great diversity of histopathologic presentation, anatomical site, and biologic behavior. Angiosarcomas constitute one of the many categories of STS. The generic term angiosarcoma includes the histological subtypes of sarcoma originating from either lymphatic or capillary endothelium, comprising lymphangiosarcoma and hemangiosarcoma, respectively. Together these lesions account for approximately 4% of all STS. We limit ourselves in this report to one histological type of highly vascularized sarcoma, with lymphedema of the extremity as the common etiological cause. 1004

ILP WITH TNF AND MELPHALAN IN STEWART-TREVES LYMPHANGIOSARCOMA Lymphedema has long been an established factor in the multifocal development of lymphangiosarcoma. In 1906, Lwenstein was the first to report this entity in a patient with chronic posttraumatic upper extremity lymphedema. However, it was not until Stewart and Treves,1 in 1948, described this syndrome of multifocal lymphangiosarcoma (which carries their name) that this feature became more widely recognized. They described six cases of lymphangiosarcoma in chronically lymphedematous arms after radical mastectomy. Since their report, 300 cases have been reported in the literature. This patient group develops lymphangiosarcoma in edematous sites in the arm, with no correlation to areas that have been subjected to radiotherapy.2 Lymphangiosarcoma may also arise in a chronic and severely edematous extremity (elephantiasis), usually of congenital origin without a preceding tumor,35 rarely in idiopathic lymphedema of delayed onset,6,7 and even more rarely in chronic lymphedema secondary to the parasitic disease filariasis.8,9 Although the relative odds for lymphangiosarcoma of the upper extremity after breast cancer are high, the absolute risk is low and is estimated to be .001%.10 Previous reports indicate a very poor prognosis of lymphangiosarcoma in postmastectomy lymphedema. In a report by Sordillo et al.,11 overall survival, regardless of the type of therapy, 19 months after the appearance of the first lesions was 50%. The 5-year survival was as low as 13.6%. Management strategies of this multifocal disease so far have included amputation or exarticulation of the limb or limb-sparing attempts with radiotherapy, chemotherapy, or both.12 With the introduction of hyperthermic isolated limb perfusion (ILP) by Creech et al.13 in 1958, a new technique was initiated to approach extremity malignancies. In 1992, Lienard et al.14 published their first report of the efficacy of ILP with tumor necrosis factor (TNF), melphalan, and interferon gamma in patients with in-transit metastasized melanoma of the extremity. Their promising results encouraged other investigators to use this treatment regimen, sparking the modern era of clinical TNF application. After disappointing results for the local treatment of STS with various chemotherapeutics, including the failure of treating locally advanced STS with ILP with melphalan alone,15 the combination of TNF and melphalan emerged as a very promising option for the limb-saving management of locally advanced STS.16 With this multidrug approach, a very high response rate in patients with STS of the extremities was reached: an 82% limb salvage rate at a median follow-up of almost 2 years. We have demonstrated in clinical17 and experimental studies18 that TNF in combination with melphalan in the setting of ILP exerts its antitumor effects

1005

primarily on the tumor-associated neovasculature, with hardly any direct cytotoxic effect on cancer cells.19,20 This makes the group of patients with Stewart-Treves angiosarcoma especially eligible to be treated with TNFbased therapy. Here we report the data on our unique experience with this particular problem. METHODS Patients Clinical records of 10 patients with Stewart-Treves angiosarcoma were reviewed. All but two patients were postmastectomy patients with a sarcoma in the arm after a prolonged period of lymphedema. Two patients had congenital lymphedema in the legs and developed a sarcoma in this area. They were treated between April 1994 and October 2001 with melphalan and TNF via hyperthermic ILP. Patient characteristics, treatment, results, and follow-up data are listed in Table 1. All patients were women, with an age range of 45 to 82 years (mean age, 64.0 years), and 16 ILPs are described. Previous radiotherapy had failed in two patients and chemotherapy in one patient, with either a short-lived response or no response at all. When patients were referred to us, they were all considered absolute candidates for amputation by their referring specialists. Isolated Limb Perfusion The technique of hyperthermic ILP with recombinant TNF and melphalan is described in detail elsewhere.16 Briefly, recombinant human TNF (Boehringer Ingelheim GmbH, Ingelheim/Rhein, Germany) and the cytostatic drug melphalan were obtained as sterile powder (100 mg) dissolved aseptically with solvent and diluents (Burroughs Wellcome, London, UK). ILPs were performed under general anesthesia and normally took 3 to 5 hours. Isolation of the blood circuit of a limb was achieved by clamping the major artery and vein and by applying a tourniquet to compress the remaining collateral vessels. Perfusion was performed at the axillary, brachial, iliac, or femoral level. ILP consisted of a 90-minute perfusion with 1 to 4 mg of TNF and a 10 mg/L (leg) or 13 mg/L (arm) volume of melphalan (25 to 110 mg) at mild hyperthermia (39C to 40C). In two patients, .2 mg of interferon was used according to the protocol that was in effect at the moment of perfusion. The composition of the perfusate was as follows: the priming volume of 700 to 850 mL consisted of 400 to 500 mL of blood (50% RBCs, 50% plasma), 200 to 400 mL of 5% dextran 40 in glucose 5% (Isodex; Pharmacia, Uppsala, Sweden), 10 to 30 mL of 8.4% sodium bicarbonate, and .5 mL of 2500 to 5000 IU of heparin. TNF was injected as bolus into the
Ann Surg Oncol, Vol. 9, No. 10, 2002

1006

T. E. LANS ET AL.
TABLE 1. Characteristics of 16 ILPs in 10 lymphangiosarcoma patients
Duration of Type Final response Limb P/R Grade Stage of ILP outcome (mo) salvage R R R P P R P P R P R R R R P P 2 1 1 1 2 2 1 1 2 1 1 2 1 1 1 1 2 2 2 1 1 2 1 1 2 1 2 2 2 2 1 1 Brach Axil Axil Axil Iliac Iliac Axil Axil Axil Axil Axil Brach Axil Axil Axil Fem CR CR PR CR PR PR PR CR CR NC MR CR CR CR CR PR 20 13 2 115 3 4 2 18 12 NC 3 19 13 11 6 5

Patient No. 1

No. Sex/age (y) Tumors 100 100 100 100 100 100 30 20 20 50 4 30 20 20 20 20

Site Lo arm Lo arm Lo arm Total arm Up leg Up leg Up arm Lo arm/hand Up arm Up arm/elbow Up lo arm Lo arm Up lo arm Up lo arm Up arm Lo leg

Size (cm) 1 3 1 1 Max Max 3 8 2 7 6 Max Max Max 1 8 1 2 1 2 5 5 4 4 3 14 4 5 3 8 1 4 4 3 4 5 5

Dead (D) or alive (A)

Cause of death

Follow-up (mo)

2 3 4 5 6 7 8

9 10

F/78 1st 2nd 2nd ILP/80 3rd 3rd ILP/82 F/67 1st F/45 nd nd 2 2 ILP/47 F/52 1st F/67 2nd 2nd ILP/69 F/60 F/65 1st F/77 2nd 2nd ILP/78 rd rd 3 3 ILP/80 F/78 F/46

Yes Yes No Yes Yes No Yes Yes Yes Yes

D A, NED A, NED D A, NED A, NED D A, NED A, NED A

CVA PD, lung mets PD, lung mets

47 115 82 3 34 15 8 46 6 5

F, female; M, male; Lo arm, lower arm; Up arm, upper arm; Up leg, upper leg; Lo leg, lower leg; P/R, primary/recurrence; Brach, brachial; Axil, axillary; Fem, femoral; CR, complete response; PR, partial response; MR, mixed response consisting of 2 CR and 2 NC; NC, no change; NED, no evidence of disease; CVA, cerebrovascular accident; ILP, isolated limb perfusion; Max, maximum; PD, pulmonary disease; mets, metastases.

arterial line, provided that limb tissue temperature was 38C. Melphalan was administered 30 minutes later, at limb temperatures between 39C and 40C. At the end of the ILP, the limb was washed with at least 2 L of 6% dextran 70 (Macrodex; Pharmacia). Response Evaluation Complete response (CR) was defined as the disappearance of all measurable disease in the limb for 4 weeks, partial response (PR) as regression of tumor size by 50% for 4 weeks, no change as regression of 50% of the tumor in the limb or progression of 25% for 4 weeks, and progressive disease as 25% disease progression. Responses were standardized according to World Health Organization criteria. Assessment of Toxicity Regional toxicity was graded according to Wieberdink et al.21 as follows: 1, no toxicity; 2, redness and slight edema; 3, considerable edema or erythema with some blistering; 4, extensive epidermolysis or obvious damage to the deep tissues, causing definite functional disturbances, or threatening, or manifest compartmental syndrome; and 5, reaction requiring amputation. Systemic toxicity was graded according to Eastern Cooperative Oncology Group/World Health Organization criteria. RESULTS Treatment Outcome Results are listed in Table 1. Sixteen ILPs with TNF plus melphalan were performed in 10 patients. All patients
Ann Surg Oncol, Vol. 9, No. 10, 2002

would have been candidates for exarticulation of the extremity. After ILP, we observed an 87% overall response rate (CR plus PR), one patient had a mixed response, and one patient did not respond to therapy. In nine perfusions (56%) a CR was achieved, and five perfusions (31%) resulted in a PR. Limb salvage was achieved in eight patients (80%), with a mean follow-up duration of 34.8 months (range, 3 to 115 months). Four patients received a second or even third ILP on recurrent disease after a period of CR or PR achieved by the previous ILP. Figure 1AC shows clinical pictures of patient 2, who was still alive with no evidence of disease 115 months after her perfusion. Limb Salvage Limb salvage was obtained in 8 (80%) of 10 patients (Table 1). One patient with a PR showed evidence of extensive recurrent disease in her leg after her second perfusion. Another patient showed no response at all. They both underwent amputation of the perfused limb within 4 months of amputation. Systemic Metastases and Survival At a median follow-up of 34.8 months (range, 3115 months), only two patients had died of metastatic disease, and one patient died of a cerebrovascular accident (Table 1). At this moment, one patient is alive with local disease. All other six patients are alive without any evidence of local or systemic disease.

ILP WITH TNF AND MELPHALAN IN STEWART-TREVES LYMPHANGIOSARCOMA

1007

blistering and slightly disturbed mobility (Table 2). Grade 2 toxicity with redness and slight edema developed after six perfusions; in all other perfusions, no signs of local toxicity were seen. Transient motor neurapraxia of 6 months was seen after two perfusions. Overall, no severe regional toxicity was observed, despite lymphedemas being present in all extremities before the first ILP. Systemic toxicity was absent to mild and was easy manageable. There have been no cases with involvement of toxic shock, liver toxicity, hematotoxicity, or renal dysfunction. In three perfusions, a transient period of fever between 39C and 40C followed ILP, because at that time of the program, no nonsteroidal anti-inflammatory drugs were administered immediately after the ILP. Total days of admission varied between 4 and 14, with a mean of 8.1 days. DISCUSSION In a unique patient population of 10 patients with this rare, limb-threatening vascular sarcoma, we demonstrated that amputation can be avoided in most patients by treating the condition with an ILP with TNF plus melphalan. Also, survival has been remarkably good as compared with the other reports in the literature, because seven patients are still alive (six with no evidence of disease), at a mean follow-up of 35 months and a maximum follow-up of almost 10 years in a patient with 100 tumors before treatment. There is, in the literature,
TABLE 2. Leakage and regional toxicity
Patient No. 1 First a Second Third 2 3 First Second FIG. 1. (A) More than 100 small angiosarcomas in the skin throughout a lymphedematous left arm. These lesions disappeared all within 1 week after the TNF-based ILP. (B) Confluent full thickness (skinsubcutaneous fat up to the fascia) angiosarcomas in the lower arm. These lesions took 2 months to completely disappear after the ILP. (C) CR of the thick lesions, histopathologically confirmed (biopsy-scar). Recent follow-up confirms continuing CR at 121 months. 4 5 First Second 6 7 8 First Second Third 9 10
a

TNF (mg) 3 3 3 3 4 4 3 3 2 2 3 3 1 3 1 4

Melphalan (mg) 36 33 45 42 100 110 50 40 32.5 40 25 32.5 33 33 40 45

Interferon (mg) .2 .2 0 .2 0 0 0 0 0 0 0 0 0 0 0 0

% Leakage 1 1 2 1 6 1 0 10 0 0 0 1.5 0 0 0 0

Wieberdink classification 2 2 3 3 1 1 3 2 2 1 2 3 2 1 1 1

Toxicity Four perfusions were followed by regional edema (grade 3, according to Wieberdink et al.21), with some

TNF, tumor necrosis factor. Patients who underwent multiple (first, second, or third) ILPs.

Ann Surg Oncol, Vol. 9, No. 10, 2002

1008

T. E. LANS ET AL. of pre- and postperfusion angiographies.17 In sarcoma patients, we have clearly shown with magnetic resonance spectrometry studies that the metabolic shutdown of tumor is virtually complete within 16 hours after the perfusion, confirming the likelihood of TNFs mediating its most important effect on the tumor vasculature.30 Our results demonstrate that combined treatment with melphalan and TNF can eventuate in good clinical response rates in patients with extensive disease who were candidates for amputation. This implies that in ILP, we have a new treatment option in the management of StewartTreves angiosarcoma. REFERENCES
1. Stewart F, Treves N. Lymphangiosarcoma in postmastectomy lymphedema. Cancer 1948;1:64 81. 2. Ironside P, Allen PW. Test and teach number eleven. Diagnosis: post-mastectomy lymphangiosarcoma of Stewart and Treves. Pathology 1977;9:110, 178 9. 3. Dubin HV, Creehan EP, Headington JT. Lymphangiosarcoma and congenital lymphedema of the extremity. Arch Dermatol 1974;10: 608 14. 4. Laskas JJ, Shelley WB, Wood MG. Lymphangiosarcoma arising in congenital lymphedema. Arch Dermatol 1975;111:86 9. 5. Prudden JF, Wolarsky ER. Lymphangiosarcoma of the thigh. Case report. Arch Surg 1967;94:376 9. 6. Baes H. Angiosarcoma in a chronic lymphedematous leg. Dermatologica 1967;134:331 6. 7. Chen KT, Gilbert EF. Angiosarcoma complicating generalized lymphangiectasia. Arch Pathol Lab Med 1979;103:86 8. 8. Muller R, Hajdu SI, Brennan MF. Lymphangiosarcoma associated with chronic filarial lymphedema. Cancer 1987;59:179 83. 9. Sordillo EM, Sordillo PP, Hajdu SI, Good RA. Lymphangiosarcoma after filarial infection. J Dermatol Surg Oncol 1981;7:2359. 10. Cozen W, Bernstein L, Wang F, Press MF, Mack TM. The risk of angiosarcoma following primary breast cancer. Br J Cancer 1999; 81:532 6. 11. Sordillo PP, Chapman R, Hajdu SI, Magill GB, Golbey RB. Lymphangiosarcoma. Cancer 1981;48:1674 9. 12. Grobmyer SR, Daly JM, Glotzbach RE, Grobmyer AJ. Role of surgery in the management of postmastectomy extremity angiosarcoma (Stewart-Treves syndrome). J Surg Oncol 2000;73:182 8. 13. Creech O Jr, Krementz E, Ryan J, Winblad J. Chemotherapy of cancer: regional perfusion utilising an extracorporeal circuit. Ann Surg 1958;148:616 32. 14. Lienard D, Ewalenko P, Delmotte JJ, Renard N, Lejeune FJ. High-dose recombinant tumor necrosis factor alpha in combination with interferon gamma and melphalan in isolation perfusion of the limbs for melanoma and sarcoma. J Clin Oncol 1992;10:52 60. 15. Klaase JM, Kroon BB, Benckhuijsen C, Van Geel AN, AlbusLutter CE, Wieberdink J. Results of regional isolation perfusion with cytostatics in patients with soft tissue tumors of the extremities. Cancer 1989;64:616 21. 16. Eggermont AM, Schraffordt KH, Klausner JM, et al. Isolated limb perfusion with tumor necrosis factor and melphalan for limb salvage in 186 patients with locally advanced soft tissue extremity sarcomas. The cumulative multicenter European experience. Ann Surg 1996;224:756 64. 17. Eggermont AM, Schraffordt KH, Lienard D, et al. Isolated limb perfusion with high-dose tumor necrosis factor-alpha in combination with interferon-gamma and melphalan for nonresectable extremity soft tissue sarcomas: a multicenter trial. J Clin Oncol 1996;14:2653 65.

no comparable concatenation of such a patient group treated in this fashion. Various treatment regimens have been tried: radiotherapy, amputation or disarticulation, intravenous chemotherapy, or combinations of these modalities.12 Although there have been anecdotal reports of long-term survivors after systemic chemotherapy,22,23 overall there has been limited success at treating StewartTreves syndrome with chemotherapy. The 5-year survival was as low as 13.6% in a report by Sordillo et al.11 They reviewed 44 cases seen at the Memorial Hospital, New York, NY, and reported a median survival of 48 months in 11 patients treated with amputation, compared with a 20-month median survival in patients treated with radiotherapy or chemotherapy. They concluded that surgical treatment should consist of either wide excision or amputation. A 5-year survival of only 8.5% was reported by Woodward et al.24 Yap et al.23 reviewed 22 patients seen at M. D. Anderson Hospital for postmastectomy lymphangiosarcoma over a 20-year period. Thirteen patients were treated regionally, systemically, or both, with a variety of single or combination chemotherapeutic agents; the overall response rate was 42%. Five patients who did not respond to chemotherapy had a median survival of 4 months, whereas the six chemotherapy responders had a median survival of 26.5 months. Two cases described by Kaufmann et al.25 experienced longterm survival after combined-modality treatment with actinomycin D chemotherapy and high dose radiotherapy. They survived at least 13 and 19 years, respectively; nonetheless, in the first case, an amputation of the arm was performed. Overall, even in the setting of initial surgical treatment, survival was poor ( 40%) in a collective comparison of 160 cases reported in the literature.12 The length of survival in untreated patients is only 5 to 8 months.24 Planning of appropriate therapy for these patients is dependent on the presence or absence of metastatic disease at the time of presentation. Death from Stewart-Treves syndrome usually occurs as a result of pulmonary metastasis, although spread of disease may also occur to bone and liver. The rationale of the combination of TNF and melphalan delivered locally is to make use of targeting both the stromal neovascular compartment and the cancer cell compartment of the tumor.26 28 The permeability-enhancing effect of TNF has also been demonstrated in a rat ILP model to significantly increase intratumor melphalan concentrations as a key explanation for the synergistic antitumor effects observed.29 The resulting effect is a rapid and profound necrosis. The selective destructive effects of TNF-ILP on tumor-associated vessels have been illustrated in previous publications by means
Ann Surg Oncol, Vol. 9, No. 10, 2002

ILP WITH TNF AND MELPHALAN IN STEWART-TREVES LYMPHANGIOSARCOMA

18. Manusama ER, Stavast J, Durante NM, Marquet RL, Eggermont AM. Isolated limb perfusion with TNF alpha and melphalan in a rat osteosarcoma model: a new anti-tumour approach. Eur J Surg Oncol 1996;22:1527. 19. Blick M, Sherwin SA, Rosenblum M, Gutterman J. Phase I study of recombinant tumor necrosis factor in cancer patients. Cancer Res 1987;l47:2986 9. 20. Feinberg B, Kurzrock R, Talpaz M, Blick M, Saks S, Gutterman JU. A phase I trial of intravenously-administered recombinant tumor necrosis factor-alpha in cancer patients. J Clin Oncol 1988; 6:1328 34. 21. Wieberdink J, Benckhuysen C, Braat RP, van Slooten EA, Olthuis GA. Dosimetry in isolation perfusion of the limbs by assessment of perfused tissue volume and grading of toxic tissue reactions. Eur J Cancer Clin Oncol 1982;18:90510. 22. Janse AJ, van Coevorden F, Peterse H, Keus RB, van Dongen JA. Lymphedema-induced lymphangiosarcoma. Eur J Surg Oncol 1995;21:155 8. 23. Yap BS, Yap HY, McBride CM, Bodey GP. Chemotherapy for postmastectomy lymphangiosarcoma. Cancer 1981;47:853 6. 24. Woodward AH, Ivins JC, Soule EH. Lymphangiosarcoma arising in chronic lymphedematous extremities. Cancer 1972;30:56272.

1009

25. Kaufmann T, Chu F, Kaufman R. Post-mastectomy lymphangiosarcoma (Stewart-Treves syndrome): report of two long-term survivals. Br J Radiol 1991;64:857 60. 26. Folli S, Pelegrin A, Chalandon Y, et al. Tumor-necrosis factor can enhance radio-antibody uptake in human colon carcinoma xenografts by increasing vascular permeability. Int J Cancer 1993;53: 829 36. 27. Umeno H, Watanabe N, Yamauchi N, Tsuji N, Okamoto T, Niitsu Y. Enhancement of blood stasis and vascular permeability in Meth-A tumors by administration of hyperthermia in combination with tumor necrosis factor. Jpn J Cancer Res 1994;85:32530. 28. Manusama ER, Nooijen PT, Stavast J, Durante NM, Marquet RL, Eggermont AM. Synergistic antitumour effect of recombinant human tumour necrosis factor alpha with melphalan in isolated limb perfusion in the rat. Br J Surg 1996;83:5515. 29. de Wilt JH, ten Hagen TL, de Boeck G, van Tiel ST, de Bruijn EA, Eggermont AM. Tumour necrosis factor alpha increases melphalan concentration in tumour tissue after isolated limb perfusion. Br J Cancer 2000;82:1000 3. 30. Ferrero E, Zocchi MR, Magni E, et al. Roles of tumor necrosis factor p55 and p75 receptors in TNF-alpha-induced vascular permeability. Am J Physiol Cell Physiol 2001;281:C11739.

Ann Surg Oncol, Vol. 9, No. 10, 2002