The palliative value of tumor necrosis factor α-based isolated limb perfusion in patients with metastatic sarcoma and melanoma


  • Dirk J. Grunhagen M.D.,

    1. Department of Surgical Oncology, Erasmus MC–Daniel den Hoed Cancer Center, Rotterdam, The Netherlands
    Search for more papers by this author
  • Johannes H. W. de Wilt M.D., Ph.D.,

    1. Department of Surgical Oncology, Erasmus MC–Daniel den Hoed Cancer Center, Rotterdam, The Netherlands
    Search for more papers by this author
  • Wilfried J. Graveland M.Sc.,

    1. Department of Statistics, Erasmus MC–Daniel den Hoed Cancer Center, Rotterdam, The Netherlands
    Search for more papers by this author
  • Albertus N. van Geel M.D., Ph.D.,

    1. Department of Surgical Oncology, Erasmus MC–Daniel den Hoed Cancer Center, Rotterdam, The Netherlands
    Search for more papers by this author
  • Alexander M. M. Eggermont M.D., Ph.D.

    Corresponding author
    1. Department of Surgical Oncology, Erasmus MC–Daniel den Hoed Cancer Center, Rotterdam, The Netherlands
    • Department of Surgical Oncology, Erasmus MC–Daniel den Hoed Cancer Center, 301 Groene Hilledijk, 3075 EA Rotterdam, The Netherlands
    Search for more papers by this author
    • Fax: (011) 31 104391011



Both patients with soft tissue sarcoma (STS) and patients with melanoma have limited treatment possibilities once the tumor has metastasized systemically. In patients with extremity STS or bulky melanoma in-transit metastases, the local tumor burden may be so problematic that, even in patients with systemically metastasized disease, an amputation may be inevitable. Isolated limb perfusion (ILP) has proven to be an excellent, local, limb-saving treatment option in patients with locally advanced extremity tumors. In this study, the authors investigated the palliative value of the ILP procedure to avoid amputation in patients who had Stage IV STS and melanoma.


From 1991 to 2003, of 339 tumor necrosis factor α (TNF)-based ILPs, 51 procedures were performed for either Stage IV STS (n = 37 patients) or Stage IV melanoma (n = 14 patients). All patients underwent an ILP with TNF and melphalan of the upper limb (n = 4 patients) or the lower limb (n = 47 patients) with 26–140 mg melphalan and 2–4 mg TNF.


The overall response in patients with Stage IV STS was 84%, and their median survival was 12 months after ILP. Limb salvage was achieved in 36 of 37 patients, with 1 patient undergoing amputation due to treatment toxicity. In the patients with Stage IV melanoma, the complete response rate was 43%. All patients with melanoma preserved their limb during a median survival of 7 months.


TNF-based ILP is an excellent procedure that provided tumor control and limb salvage for the short survival of patients with metastasized, very bulky, limb-threatening tumors of the extremity. Cancer 2006. © 2005 American Cancer Society.

Soft tissue sarcoma (STS) has a high mortality rate of up to 50% due to the propensity of the tumor to develop distant metastases.1 Patients with metastatic sarcoma (TNM Stage IV) have a median survival of only 1.5 years,1, 2 and that survival rate has not improved in the last decades.3 For patients with nonmetastasized STS, standard therapy consists of surgical resection of the tumor, often combined with radiotherapy. When large or multiple tumors make this treatment option impossible, isolated limb perfusion (ILP) with tumor necrosis factor α (TNF) and melphalan (TM-ILP) has proven to be effective as a limb-salving therapeutic option.4–6

Patients with distant melanoma metastases (Stage IV) have very poor survival expectations, with a median survival that barely exceeds 1 year. In-transit metastases in the extremities occur in 5–8% of patients with high-risk melanoma, and these patients cannot be treated surgically. It has been shown that ILP is an excellent treatment option for patients in this category.7, 8

A minority of patients with Stage IV STS or patients with Stage IV melanoma may present with combined problems of systemic metastases and locally very advanced tumor (recurrences) or in-transit metastases. The local situation may be disabling and very difficult to handle because of the poor sensitivity of these tumors to systemic therapies. Radiotherapy options also may be absent or very limited in these patients because of prior treatments, the tumor size, and the multiplicity of tumors in the extremity. In general, we offer these patients with severe local disability and no other palliative treatment options a TNF-based ILP. This study was undertaken to evaluate the value of the TM-ILP procedure in patients with Stage IV STS and melanoma who had a short life expectancy.



Of 217 patients who underwent TM-ILP for STS in the Daniel den Hoed Cancer Center between 1991 and July 2003, 37 patients were treated for systemically metastasized extremity STS. During the same period, 14 patients with American Joint Committee on Cancer/The University of Texas M. D. Anderson Cancer Center Stage IV melanoma (distant metastases present) underwent an ILP for local control of disease in the limb. There were 14 females and 23 males in the STS group with a median age of 55 years (range, 17–86 yrs), and the melanoma group consisted of 5 females and 9 males with a median age of 61 years (range, 25–78 yrs). Histologic classification of the tumors is summarized in Table 1. All patients with STS were candidates for amputation, because primary surgical resection was impossible either because of the size and location or because of the multifocality of the tumors. All patients with melanoma had large and multiple tumors in the limb (the size and number of the lesions prevented primary surgery), which caused severe local discomfort. In total, 51 patients underwent ILP for metastasized tumors. The patient and tumor characteristics are summarized in Table 2.

Table 1. Histologic Classification of 51 Patients who Underwent Isolated Limb Perfusion with Tumor Necrosis Factor α and Melphalan for Metastasized Tumors
HistologyNo. of patients
  1. MFH: malignant fibrous histiocytoma; STS: soft tissue sarcoma.

Synovial sarcoma7
MFH/high-grade pleiomorphic sarcoma5
Clear-cell sarcoma4
(Extraskeletal) Ewing sarcoma3
Other STS8
Table 2. Patient and Tumor Characteristics in 51 Patients who Underwent Isolated Limb Perfusion with Tumor Necrosis Factor α and Melphalan for Metastasized Tumors
CharacteristicNo. of patients (%)
Stage IV sarcoma (n = 37)Stage IV melanoma (n = 14)
  • XRT: external radiotherapy; CT: computed tomography; ILP: isolated limb perfusion.

  • a

    Five patients with soft tissue sarcoma and one patient with melanoma received both XRT and CT, one patient with melanoma received both XRT and prior ILP.

 Female14 (38)5 (36)
 Male23 (62)9 (64)
 ≤ 60 yrs24 (65)7 (50)
 > 60 yrs13 (35)7 (50)
Tumor site  
 Arm4 (11)0 (0)
 Leg33 (89)14 (100)
Primary/locally recurrent  
 Primary18 (49) 
 (Re)recurrent19 (51) 
Previous treatmenta  
 None16 (43)6 (43)
 XRT7 (19)3 (21)
 CT17 (46)4 (29)
 ILP2 (5)3 (21)


Patients underwent an ILP through the axillary (n = 1 patients), brachial (n = 3 patients), iliac (n = 29 patients), femoral (n = 16 patients), or popliteal (n = 2 patients) approach. The ILP technique has been described previously.5, 9 Briefly, recombinant human TNF (Boehringer Ingelheim GmbH, Ingelheim/Rhein, Germany) and the cytostatic drug melphalan (L-PAM, Alkeran; Burroughs Wellcome Ltd., London, United Kingdom), which was obtained as a sterile powder, were dissolved aseptically using solvent and diluents (Burroughs Wellcome Ltd.). ILPs were performed under general anesthesia and normally took from 2.5 hours to 4.0 hours. Isolation of the blood circuit of a limb was achieved by clamping and cannulation of the major artery and vein, connection to an oxygenated extracorporeal circuit, and application of a tourniquet to compress the remaining collateral vessels. ILP consisted of a 90-minutes perfusion with 1–3 mg TNF (arm) or 3–4 mg TNF (leg) and a 10-mg/L volume (leg) or a 13-mg/L volume (arm) of melphalan at mild hyperthermia (maximal tissue temperatures of 39.5 °C in the leg and 38.5 °C in the arm). The median dose of melphalan was 60 mg (mean, 68.7 mg; range, 26–140 mg), and the median dose of TNF was 4 mg (mean, 3.3 mg; range, 2–4 mg). In the melanoma group, all patients underwent lower extremity ILP and received a median dose 90 mg melphalan (mean, 90.4 mg; range, 50–120 mg) and a median dose of 4 mg TNF (mean, 3.7 mg; range, 2–4 mg). TNF was injected as a bolus into the arterial line provided limb tissue temperature had reached 38 °C. Melphalan was administered after 10–30 minutes at limb temperatures between 38 °C and 39.5 °C. During the procedure, leakage was monitored continuously by using a precordial scintillation probe to detect leakage of radiolabeled albumen injected into the perfusion circuit. At the end of the ILP, the limb was washed out with at least 1 L (arm) up to 4 L (iliac perfusion) of physiologic saline solution and 6% dextran 70 (Macrodex; Pharmacia, Uppsala, Sweden). In patients with melanoma who underwent iliac perfusion, an iliac lymph node dissection was performed, and patients who had palpable lymph node disease in the groin underwent an ilioinguinal lymph node dissection during the same operative session as the ILP but prior to the ILP.

Response Evaluation and Toxicity

Clinical response evaluation in patients with STS was performed 2 weeks, 4 weeks, 8 weeks, and 12 weeks after ILP and every 3 months thereafter for the first year both by clinical examination and by magnetic resonance imaging 4–6 weeks after ILP, 8–12 weeks after ILP, and every 3–6 months thereafter. Response evaluation in patients with melanoma was performed 2–4 weeks and 8 weeks after ILP by clinical examination; and, after that, at regular 3-month intervals for the first 2 years, and at longer intervals thereafter. Responses are reported according to World Health Organization (WHO) criteria.10 When histologic response could be assessed in patients with STS, the final outcome was adjusted if the pathologic response (based on the percentage of necrosis: complete response [CR], 100% necrosis; partial response [PR], 50–99% necrosis; or no change [NC], < 50% necrosis) differed from the clinical response.

Acute local toxicity of the ILP procedure was classified according to Wieberdink et al.,11 as follows: Grade 1, no reaction; Grade 2, slight erythema or edema; Grade 3, considerable erythema or edema with some blistering, slightly disturbed motility permissible; Grade 4, extensive epidermolysis or obvious damage to the deep tissues, causing definite functional disturbance, and threatening or manifest compartmental syndrome; and Grade 5, reaction that may necessitate amputation. Systemic toxicity is reported according to WHO criteria.




A clinical CR after ILP in 37 patients with metastasized extremity STS was observed in 8%. A PR occurred in 65% of patients, NC was observed in 24%, and no clinical response was determined in 1 patient (3%). Eighteen patients underwent resection of the tumor remnant after shrinkage, and patients with multiple tumors underwent a biopsy for response assessment (median time after ILP, 3 mos; range, 0.4–7.4 mos), which showed 100% necrosis in 4 patients, 50–99% necrosis in 12 patients, and 0–50% necrosis in 2 patients. In 6 patients, the histologic response differed from the clinical response (3 patients were reclassified from NC to a PR, and 3 patients were reclassified from a PR to a CR), resulting in a final 16% CR rate, a final 68% PR rate, and a final 16% NC rate. The overall response rate of TM-ILP in patients with Stage IV sarcoma (84%) did not differ significantly from our ILP experience in treating patients with nonmetastatic disease (overall response rate, 71%; P = 0.11).


Among the patients with Stage IV melanoma, a CR was observed in 6 patients (43%), a PR was observed in 7 patients (50%), and NC was observed in 1 patient (7%). The response in patients with Stage IV melanoma differed significantly from the responses obtained in patients treated in our institution with Stage IIIA disease (in-transit metastases only) or Stage IIIAB disease (in-transit metastases and lymph node metastases; 84% and 63%, respectively; P values = 0.004 and 0.184, respectively).

Outcome and Toxicity

Overall survival of the 37 patients with STS after ILP was 53% at 1 year and 15% at 2 years. The median survival after ILP was 12 months, and death was related to the disease in all patients. In patients with melanoma, the median survival was 7 months: Thirty-nine percent of the patients still were alive after 1 year, but no patients survived for 2 years. Local control of the disease for the full duration of life, which is the main objective of ILP for patients who have Stage IV disease, was obtained in 33 patients with STS and in 11 patients with melanoma. In seven patients, continued control was not achieved: Four patients with STS and three patients with melanoma experienced local progression of disease in the limb after an initial response. Because of their initial responses and the ongoing progression of their distant metastases, the local control achieved in these patients lasted long enough to avoid a situation in which an amputation was required. Overall survival and time to local progression are shown in Figure 1 (STS) and Figure 2 (melanoma). The toxicity of TM-ILP was very mild, consistent with prior reports. Local toxicity was mild (Grade 1–2) to moderate (Grade 3) in 35 patients with STS who underwent ILP and in all patients with melanoma who underwent ILP. One patient developed pain in the lower leg without the symptoms of compartment syndrome (Wieberdink Grade 4). In one other patient, there was extensive necrosis of the tumor after ILP that developed into a secondary infection. Eventually, an amputation was inevitable in the postoperative period, which is considered a Grade 5 local toxicity of the ILP procedure. This was the only amputation that had to be performed in the described patient population. Limb function of the other 36 patients with STS was excellent in 27 patients; mildly disturbed in 5 patients, and severely affected (leading to the use of crutches) in 4 patients. All patients with melanoma had excellent limb function after perfusion. Systemic toxicity was absent or mild, with only 1 patient who had STS developing a fever > 40 °C for > 24 hours and with 1 patient who had melanoma developing Grade 4 leucopenia, which lasted only for 1 day and did not require any type of intervention (Table 3). The median hospital stay for the 51 patients was 10 days. Outcomes of the patients are summarized in Table 4.

Figure 1.

Overall survival (OS) and time to local progression (TTLP) are illustrated after isolated limb perfusion (ILP) in patients with Stage IV sarcoma. X-axis: the time from ILP in months; Y-axis: cumulative percentage.

Figure 2.

Overall survival (OS) and time to local progression (TTLP) are illustrated after isolated limb perfusion (ILP) in patients with Stage IV melanoma. X-axis: the time from ILP in months; Y-axis: cumulative percentage.

Table 3. Local and Systemic Toxicity of 51 Patients who Underwent Isolated Limb Perfusion with Tumor Necrosis Factor α and Melphalan for Metastasized Tumors
ToxicityNo. of patients
Grade 0Grade 1Grade 2Grade 3Grade 4Grade 5
  • a

    Wieberdink grade for toxicity.

  • b

    World Health Organization criteria.

  • c

    Support of vasopressors was needed.

Local toxicitya 1031811
Systemic toxicityb      
Temperature24 (< 38 °C)13 (38–39 °C)11 (19–40 °C)2 (> 40 °C, < 24 hrs)1 (> 40 °C, > 24 hrs)
Shockc51 (absent)0 (present)    
Table 4. Outcome of 51 Patients after Isolated Limb Perfusion for Metastasized Tumors
OutcomePercentage of patients
STS (n = 37 patients)Melanoma (n = 14 patients)
  • STS: soft tissue sarcoma; OS: overall survival; LPFS: local progression-free survival; LS: limb salvage; NA: not applicable.

  • a


1-Yr OS5338
2-Yr OS15NA
2-Yr LPFS8678a


The current data show that ILP can provide excellent local control and salvage of the limb in patients with Stage IV STS or melanoma in combination with limb-threatening local tumors.


The overall response rate (84%) in the current study was relatively high compared both with data in literature (63–91%5, 6, 12) and with our own experience in patients with nonmetastatic disease (71%). However, this high overall response rate largely was attributable to the proportion of patients who had a PR (68%). Especially in patients with metastatic disease, any response that leads to local control and prevents amputation is worth the effort in light of their limited life expectancy.

The median survival of patients after ILP was 12 months. This corresponds well with data from other studies, which have reported a survival of 11–26 months in patients with metastatic disease,12–14 taking into account that there is a lead-time bias effect in the current study, because survival was calculated from the date of ILP.


The overall response rate in the 14 patients with Stage IV melanoma was 93% with a CR rate of 43%. This is remarkably good for a Stage IV melanoma population, because the results reported from ILP in patients with melanoma usually deal only with Stage IIIA or IIIAB disease. The responses with TM-ILP achieved in our patients with Stage IV melanoma who had bulky disease were as good as the results reported from melphalan-only ILPs in patients who had Stage IIIA and IIIAB disease,15 but they were markedly lower than the results reported from TM-ILP in patients who had Stage IIIA and IIIAB disease.8, 16, 17 Reasons for this reduced response rate are unknown, but we speculated previously that, in these patients with Stage IV melanoma, all of whom had very bulky disease in the limbs, a reduced immunocompetence in Stage IV melanoma may play a significant role, along with an increased aggressiveness of the melanoma itself.17 However, further research will be needed to explain properly this decreased response rate, which has been found universally in this and other (melphalan-based) studies.8, 18, 19 Survival in patients with Stage IV melanoma is extremely poor and reportedly is only 41–59% at 1 year, depending on the site of the metastases,20 which is in line with the 39% 1-year survival rate in the current study, again, taking into account the lead-time bias effect.


Systemic toxicity from TM-ILP virtually was absent in the procedures performed. This is a very important observation; because, in general, a palliative procedure should be relatively free of side effects. The dreaded, shock-like symptoms of (high-leakage) ILP with TNF could be avoided easily in the postoperative phase, because we know from the extensive experience in our institution that generous fluid management and the use of indomethacin in the immediate postoperative period prevents the symptoms of fever and the transient drop in blood pressure.21 Local toxicity was severe (Grade IV or V) in 2 patients in the study population and led to amputation of the limb in 1 of those patients. This is a severe, adverse, therapy-related effect, despite the fact that an amputation of the limb in this patient was inevitable if no treatment was offered. In our experience with TM-ILP in nearly 350 procedures for patients with melanoma and nonmelanoma tumors, only 2 treatment-related amputations (Wieberdink Grade 5) had to be performed.

The rationale for palliative treatment of patients with Stage IV disease by a locoregional treatment procedure deserves further discussion. For both systemically metastasized STS and melanoma, the reported effectiveness of systemic chemotherapy is very poor. In patients with STS, no effect of systemic chemotherapy on survival could be demonstrated either in the adjuvant setting after resection of the primary tumor22 or in patients who already had metastatic disease.23 In fact, a recent study from the Memorial Sloan-Kettering Cancer Center has shown that, despite improved insight into the tumor biology of sarcomas, the prognosis for patients with STS has not improved in the last 20 years.3 For patients with melanoma, the same problem is encountered: Although some treatment regimens seem to increase response rates, none of the treatment schedules have been able to prolong overall survival.24 These data indicate that in both patients with STS and patients with melanoma who have distant metastases, the objective of treatment should be improving quality of life during the terminal phase. Previous studies have demonstrated that ILP may offer palliation in both patients with melanoma25 and patients with STS26 who have advanced disease. The 51 patients presented in the current study all had locally advanced tumor activity for which amputation of the limb was inevitable. Their pre-ILP limb function was very poor; and, especially in the patients with melanoma, necrosis of the tumors resulted in socially disabling conditions. TM-ILP preserved the limb in 50 of 51 patients with excellent limb function in 31 patients.

The currently series of 51 patients showed that TM-ILP can offer important benefits to patients with metastasized melanoma and sarcoma with low treatment-related morbidity. When surgical resection of the tumor is impossible or is possible only at the cost of severe disability, a TNF-based ILP can offer limb salvage in 98% of patients, which means an important improvement in patient mobility during the last, often short, period of life. Therefore, we recommend considering TM-ILP for patients in this category as soon as the local situation of the limb prompts for action.