Effectiveness of regional chemotherapy with TNF-α/Melphalan in advanced soft tissue sarcoma of the extremities

Abstract

Hyperthermic isolated limb perfusion with tumour necrosis factor alpha (TNF-α) and melphalan was repeatedly reported to achieve extraordinarily high clinical remission rates in advanced and non-resectable soft tissue sarcoma of the limbs, thus avoiding imminent mutilation or amputation for most of those patients. With the limb being isolated throughout the extracorporal perfusion, high doses of recombinant TNF-α as well as melphalan can be applied. Basically, TNF-α directly affects the vasculature of the tumour and induces a severe inflammation with consecutive deterioration of the tumour capillaries. Furthermore, TNF-α increases the tumour-selective uptake of melphalan into the tumour cells thus leading to synergy of antivascular targeted treatment and antineoplastic effects of highest dose chemotherapy supplemented by hyperthermia. Meanwhile, a lot of sarcoma centres in Europe adopted this technique and established referral programmes for patients with non-resectable soft tissue sarcomas of the limbs. Despite these programmes many patients still do not get offered hyperthermic ILP with TNF-α and melphalan as a treatment option and modality. This article summarizes multimodality in treatment of soft tissue sarcoma of the limbs and reviews the current status of melphalan-based ILP with TNF-α (TM-ILP) and its results, to enable comparison and critical consideration of other treatment options.

• Soft tissue sarcoma
• isolated limb perfusion
• TNF-α
• limb sarcoma
• radiotherapy
• limb salvage

Soft tissue sarcoma of the limbs

Excluding primary malignant bone tumours such as osteosarcoma, multiple myeloma and Ewing-sar-coma, all other malignant mesenchymal neoplasms form the heterogeneous group of soft tissue sarcomas (STS). Due to the mesenchymal origin, STS occur in every region of the body, but the vast majority of sarcomas (>60%) are localized at the extremities. The primary goal of treatment is complete and radical resection of the tumour whereas margins of roughly 20 mm are being reported to be associated with optimal local tumour control [1], [2]. Simply, there are two categories of STS with regard to their localization and resectability: one category consists of superficially localized tumours which hence are commonly diagnosed in the early stages and rarely exceed a size of more than 5 cm. Most often, these sarcomas can be sufficiently treated by surgical resection with wide margins (20 mm) or margins at least several millimetres wide. In contrast to these tumours which are associated with a good overall prognosis, the other category consists of deep seated sarcomas which are also mostly high grade sarcomas. Because sarcomas do not make any specific symptoms, deep seated tumours commonly reach a significant extent before they are detected, causing non-specific symptoms such as swelling, palpable mass, pain, impairment of function and, rarely, nerve palsy. In contrast to superficially located tumours these deep seated sarcomas are associated with more complex surgical scenarios. There are a considerable number of patients where a radical resection in the sense of clear or even wide resection margins is not feasible. Several studies have clearly demonstrated that insufficient local tumour control increases the rate of local recurrence which negatively impacts on disease free survival and overall survival [1–4]. But studies prior to those reports indicated that mutilating surgical procedures as compartmental resections and amputations–which are apparently the most radical surgical procedures in the sense of local tumour control–do not improve overall survival [5–7]. This is most likely associated with the 40–70% risk of developing distant metastases independently from what is done in order to achieve sufficient local tumour control [6], [8]. Thus, limb sparing treatment has become the key element in the management of soft tissue sarcoma of the limbs.

Multimodality: Prerequisite for limb sparing treatment

Surgery

Even considering all current treatment options involving vascular repair, endoprosthetics and plastic or reconstructive surgery, there will be a significant number of patients where surgical resection alone remains insufficient regarding local tumour control. In patients who are marginal or even incompletely resected there might be a feasibility to perform re-resection which has repetitively been reported to enhance local control and disease free survival [2–4].

Radiotherapy

The value of postoperative radiotherapy has clearly been demonstrated for sarcoma resections with clear but narrow margins according to several retrospective studies [9–13] as well as according to prospective randomized trials [14], [15]. However, beginning of postoperative radiotherapy is frequently hindered by delayed wound healing following extensive surgery. Additionally, the extent of the irradiation field is significantly greater than in preoperative radiotherapy resulting in a considerable loss of function. Davis reported that with a significantly greater rate of fibrosis, joint stiffness and oedema in patients who underwent postoperative radiotherapy [16]. These limitations and the fact that limb-sparing surgery is in general not inferior to amputations strongly enhanced the interest in neoadjuvant treatment options in order to reduce the tumour in size and vitality, thus enabling later resection without mutilation.

Preoperative radiotherapy offers the benefit of short intervals between radiotherapy and surgery, reduced irradiation fields compared to postoperative radiotherapy and a potentially reduced subsequent surgery related to downstaging. However, up to now there is no clear evidence from prospective randomized trials that the timing of radiotherapy does significantly impact on local tumour control [17]. Focusing on the incidence of local recurrences Zagars and coauthors did not detect any differences between pre- and postoperative radiotherapy in their retrospective studies [18]. Whereas late complications are more likely associated with postoperative radiotherapy [16], [18], [19], are major wound complications seriously compromising the results of preoperative radiotherapy at the limbs [16–18], [20], [21]. Thus, the value of preoperative radiotherapy remains unclear.

Chemotherapy

Preoperative chemotherapy has been shown to induce objective remissions in a limited number of patients but did not improve resectability [4], [22–24]. Furthermore, preoperative chemotherapy has failed to demonstrate any benefit in the prospective randomized trial for stage II/III sarcomas in terms of progression-free survival and overall survival [4].

By reason of the radiation-sensitizing effects of doxorubicin and ifosfamide preoperative chemotherapy plus irradiation was investigated and recently reported to be promising in a phase I study [25]. Disease-free survival and overall survival were also superior in a historical control [26] and in the just recently published prospective phase II trial Radiation Therapy Oncology Group (RTOG) [27]. But a rate of substantial systemic toxicities and a considerable number of major local complications necessitates further evaluation of this regimen [20], [23]. The results of the just recently published prospective randomized phase III study on the combination of preoperative chemotherapy with hyperthermia are indicating improved disease-free survival and overall survival in stage II and stage III sarcomas [28]. Although the majority of tumours consisted of resectable soft tissue sarcomas of the limbs, it is the only prospective neoadjuvant trial with adequate power which indicates improved surgical resectability and not significantly increased regional toxicity compared to other treatment modalities.

Besides the evidence that multimodal treatment strategies are a prerequisite to enable limb-sparing surgery, it has also clearly been demonstrated that side effects and complications of those treatment concepts are substantial in number and character, thus significantly impairing vantages of these treatment concepts [19], [20].

Isolated limb perfusion with TNF-α and melphalan: Effectiveness

Regional application of highly dosed chemotherapy seems to be reasonable to control local tumour. As a matter of principle, highest concentrations of chemotherapeutic agents can be achieved since isolation of the tumour-bearing limb is feasible. Via clamping and cannulation of the major artery and vein with placement of a tourniquet proximal to the cannula tips and via usage of an extracorporal rotor pump plus oxygenator, isolated limb perfusion (ILP) is technically established since decades [29]. However, local remission rates have been reported to be very disappointing with solely use of melphalan, doxorubicin and cisplatin [30–33]. It was the advent of recombinant tumour necrosis factor alpha (TNF-α) which made a breakthrough in local tumour control of advanced limb sarcomas. While the systemic use of TNF-α is not indicated due to its severe inflammatory characteristics [34], [35] it can be applied in high doses in the setting of sufficient isolation during limb perfusion. The role of TNF-α in the setting of melphalan-based ILP (TM-ILP) was investigated by Lejeune and co-workers who observed impressive remissions in a pilot study on 19 patients with melanoma or sarcoma of the limbs [36]. These encouraging results were underlined by experimental data [36], [37] and led to several multicentre trials which reported extraordinarily high rates of local responses in advanced limb sarcomas [36], [38–41].

TNF-α: Tumour-specific and vasoactive anti-cancer agent

Although Posner reported a decade ago on minimal or missing response in ILP with sole use of TNF-α [42] further research by the group of Lejeune and Eggermont made clear that TNF-α is the key element in ILP. Early experimental data demonstrated an immediate and severe intratumoral activation of endothelial cells by TNF-α which induces a strong up-regulation of endothelial leukocyte adhesion molecule 1 (ELAM-1) as well as vascular cell adhesion molecule 1 (VCAM-1). That is in turn followed by adhesion of polymorphonuclear cells and severe alteration of endothelium with subsequent haemorrhage and necrosis [37], [43]. Interestingly, the authors detected only a slight increase of adhesion molecules in healthy tissue and thus only a very slight inflammation at the vasculature of healthy tissue and skin. That strongly demonstrated that the tumour vasculature is the primary target for TNF-α (Figure 1). Recent laboratory studies strengthened these findings and clearly demonstrated a deterioration of the vascular structure of the tumour, extravasation of erythrocytes and severe haemorrhage within a 30-minute time frame [44]. A clinical example demonstrating the strong correlation between TM-ILP induced deterioration of the tumour vasculature and subsequent necrosis of the tumour is shown in Figure 1. The effect of TNF-α on limb sarcomas by itself can be summarized as a tumour-specific induction of severe inflammation and subsequent destruction of the vasculature which leads to an immediate and early shut-down of the metabolism within the tumour. These TNF-specific mechanisms were repeatedly demonstrated in clinical use by angiographies as well as in experimental settings using a rat model with intravital microscopy [37], [45–47].

Figure 1. Male, 54 years; repeatedly operated for popliteal entrapment, finally evolving a big lump in the popliteal fossa (high grade leiomyosarcoma). Recommended TM-ILP to achieve limb salvage. Angiogram and dynamic MRI (upper panel). Complete and tumour-specific deterioration of the vasculature eight weeks following ILP with complete corresponding necrosis of the tumour (lower panel). Subsequent marginal resection revealing a complete histopathologic necrosis. Restoration of soft tissue with a microvascular latissimus dorsi flap.

TNF-α: Catalyser for conventional anti-tumour drugs

But direct immune effects, cellular infiltrations or cytokine expressions linked to TNF-α could not be detected which explains why solely use of TNF-α does not improve response rates in ILP [42], [48], [49]. However, de Wilt and co-workers demonstrated that the induction of rapid damage to the tumour vascular endothelial lining by TNF-α leads to an augmented drug accumulation of melphalan into the tumour tissue [50]. The deterioration of the tumour vasculature was shown to induce a four- to six-fold increased uptake of the cytostatic drug into the tumour tissue compared to ILP with melphalan alone. No increased uptake of melphalan could be seen in normal tissues when melphalan-based ILP was supplemented with TNF-α [47], [49], [50] which substantiated the ILP-TNF-α/melphalan setting (TM-ILP) to be a tumour-specific targeted treatment. While response rates are not linked to specific entities of sarcoma, the vessel- and micro-vessel density was shown to influence the drug uptake. The authors found a strong correlation between the degree of tumour vascularization and the synergistic effects of TNF-α and melphalan [51]. Besides this early effect on tumour vasculature and drug uptake, TNF-α seems to provide a late therapeutic effect. Ruegg et al. reported on a reduced activation of the adhesion receptor αVβ3 in human endothelial cells when exposed to TNF-α [52]. These findings indicate that TNF-α might negatively influence tumour-associated neoangiogenesis by inhibition of the αvβ3 integrin signalling [52], [53].

Alternative cytotoxic and vasoactive drugs

Other agents than melphalan were investigated as alternative drugs. Although response rates following TM-ILP with doxorubicin were comparable to those with melphalan it was less effective regarding limb salvage and associated with an increased regional toxicity [33], [54]. Cisplatin had even lower clinical response rates and was also associated with increased regional toxicity [55]. Actinomycin D which has been used in clinical practice for many years was investigated in standardized animal models by the Rotterdam group of Eggermont, ten Hagen and co-workers. They found that the combination of Actinomycin D and TNF-α leads to idiosyncratic toxicity in normal as well as in tumour tissue resulting in severe tissue destruction with loss of the limb [56].

The same group investigated alternative vasoactive agents to diminish potential systemic side effects of TNF-α which is known to be the strongest inductor of inflammation and severe hemodynamic alterations. By replacing TNF-α with histamine, drug induced intra-tumour haemorrhage and increased drug uptake leading to a 66% overall response rate were reported [57]. Likewise Interleukin 2 was investigated and showed a 67% response rate when combined with melphalan [58]. In contrast to the synergistic effects of histamine and interleukin 2 in systemic immunotherapy, the application of both agents had a conflictive and negative effect in the melphalan-based ILP-setting [59]. Although histamine-based ILP provided excellent response rates in these experimental settings which are comparable to those of TNF, no further clinical trial or data exists. Thus, recombinant TNF-α (rh TNF-α, Beromun®) and melphalan (L-phenylalanine-mustard) still represent the standard drugs for hyperthermia isolated limb perfusion in advanced and non-resectable sarcoma of the limbs [53], [60].

Technical prerequisites and dosage

Patient conditions. Sepsis is associated with an increase of TNF-α receptors which could lead to a non-tumour-specific regional inflammation with irreversible tissue damage. Thus, patients scheduled for TM-ILP should not be in septic conditions. Patients with current chemotherapy should have recovered from leucopenia because experimental data demonstrated a complete loss of TNF-α effect as well as of synergy between TNF-α and melphalan in leucopenic rats [61].

Hyperthermia and oxygenation

To some extent hyperthermia increases the permeability of the endothelial layer and induces recruitment of poorly perfused areas of the vastly pleomorphous sarcomas by arteriovenous shunting. While temperatures below 38°C are associated with a complete loss of all anti-tumour effects, mild hyperthermia (38°–39.0°C) is proven to be most effective [62]. Hyperthermia exceeding 39°C is indeed associated with increasing rates of complete remissions but is also strictly associated with severe damage to normal tissues [42], [62]. Despite the fact that hypoxemia can improve the anti-tumour effects of ILP with either TNF-α or melphalan it does not reach the efficacy of ILP in conditions of normoxaemia or hyperoxaemia with the combined use of TNF-α and melphalan [62].

Isolation and leakage monitoring

TNF-α release in the systemic circuit can induce dose dependent hemodynamic alterations up to shock-like syndromes, therefore a reliable isolation must be warranted throughout the perfusion [35]. This is easily feasible in distal ILPs of the arm and leg where an inflatable tourniquet can be applied proximal to the cannulas and a complete disruption of all collaterals can be achieved by pneumatic pressure. To achieve sufficient isolation at the iliac or axillary level is much more sophisticated where only a tightened Esmarch bandage can be applied in the groin or in the axilla. Numerous collaterals in the pelvic region require a smart balancing between the extracorporal circuit (pressure, flow rate) and systemic hemodynamic (fluid administration, central venous pressure, middle arterial pressure, inotropes, catecholamines, etc.) to obtain sufficient isolation [63]. Although continuous leakage monitoring by radio-labelled serum (99mTc, Indium, Iodine) is in general mandatory for TM-ILP, it is of utmost importance in those proximal perfusions where the balance between the extracorpral and systemic circuit is particularly fragile. Leakage monitoring has definitively proven over the years to enable secure performance of TM-ILP.

Dosage

Dosages of 3-4 mg were established in the initial trials which have led to the approval of TNF-α for sarcoma treatment. Based on a very early single institutional experience (complete local response in all of eight patients with sarcoma of the limb) with TM-ILP and only 1 mg TNF-α [65], it remained unclear whether lower dosages of TNF-α could be as sufficient as the initial dosages which have led to approval. Confirmation of this early report came from an Italian study which consisted of a few patients where ILP using 1 mg TNF-α and doxorubicin (TD-ILP) also had an overall response rate of 90% [66], [67]. Data from further experimental models also demonstrated that dose reduction from the original 50 µg down to 10 µg did not negatively affect clinical response rates. However, the synergistic effect of TNF-α and melphalan was completely lost when further dose reduction beyond this 10 µg threshold was executed [62]. Data from the recent French randomized trial on 100 patients with four assigned dosages (0.5, 1.0, 2, 3–4 mg) did not reveal any dose dependent difference in clinical response rates [66]. Additionally, the Rotterdam experience which is based on the retrospective analysis of more than 200 patients indicated that reduced TNF-α dosages (1–2 mg) resulted in comparable response rates to those where high dose TNF-α was administered [68].

Complications, clinical response rates and clinical outcome

Complications

Complications in melphalan-based ILP with TNF-α derive either from hemodynamic alterations caused by leakage of TNF-α from the systemic circuit or from regional toxicity in the perfused limb. There are no current reports on leakage related complications, which indicates that leakage monitoring seems to be very efficient and thus leakage was not a substantial problem in TM-ILP. Earlier and sparse reports estimated leakage to be a minor problem with 2% leakage on average causing only mild and reversible hemodynamic effects as hypotension or tachycardia [35], [63], [64], [69]. All symptoms related to regional toxicity are classified according to the five grade scale implemented by Wieberdink [70]. This scale is displayed in Table I and shows that grade II reactions consisting of mild hyperaemia, erythema or mild oedema are most frequent following TM-ILP. In contrast to a very mild regional toxicity in most of the patients, the incidence of severe regional toxicity with epidermolysis, compartment syndrome (grade IV) or even tissue damage with loss of the limb (grade V) remains very low [38–40], [71]. Interestingly, the results of the Italian trials with doxorubicin-based ILP showed higher incidences of grade III and IV toxicities whereas melphalan is recommended as the standard cytotoxic agent [66], [67]. However, whenever performing ILP with TNF-α caution must be taken with the limb after revascularization to accurately judge the swelling of the muscle and to execute immediate fasciotomy in case the clinical situation is doubtful [72]. Other complications related to TM-ILP are rare and consist mainly of thromboses at the site of the arteriotomy with an incidence below 2% [73].

Table I.  Classification of the regional toxicity in TM-ILP according to Wieberdink [70] (*summarizing results from studies [36], [38], [53], [54], [66], [67], [71], [74]).

Grade	Range*	Symptoms
I	10–20%	no reaction
II	60–75%	mild oedema, hyperaemia
III	5–12%	severe oedema, blistering
IV	2–6%	epidermolysis, compartment syndrome
V	0.5–2%	tissue necrosis, amputation

Clinical response rates

Lejeune and colleagues pioneered the use of TNF-α in the ILP-setting and were among the first to describe enormous local response rates with this treatment in advanced melanoma and sarcoma [36]. Meanwhile, melphalan-based ILP with TNF-α is successfully established in currently more than 40 sarcoma centres in Europe. Based on patent and licensing rates, the registration file of TM-ILP is yet not presented to the FDA, which is why TM-ILP unfortunately cannot be offered to suitable patients in North America. In Europe TM-ILP has been an approved and registered treatment for advanced soft tissue sarcoma since 1988 [74].

The results of the multicentre trial which has led to approval of TNF-α, revealed complete and partial local responses in amounts which were not previously known in local treatment of advanced soft tissue sarcomas of the limbs [39], [74]. Furthermore, local complication rates were dramatically low compared to what was known from alternative multimodality treatment options [14], [39]. It is quite important to mention that local tumours in all 246 patients of this trial were either judged to be non-resectable (87%) or only resectable if mutilating surgery plus radiotherapy would have been performed (13%); the judgement was performed by an independent surgical review committee. Another aspect which strengthens the superiority of the results of TM-ILP is the selection bias of patients included in the multicentre trial. Only 55% of patients had primary sarcomas, while 45% had local multifocal primary or even multiple recurrences [39]. The 95% rate of patients suffering from high grade sarcomas with a considerable number of tumours exceeding 10 cm and 15% patients with overt distant metastases underline that this collective represents the most problematic subgroup of patients with soft tissue sarcoma. The five year overall survival rate was 50% and did not differ from the survival rate of patients from the Scandinavian sarcoma register (matched control study), thus indicating that TM-ILP does not negatively influence survival [39].

Over the past years reports from different institutions strengthened the results of the initial multicentre study. Focusing on those papers with patient numbers greater than 20, local response rates ranged from 63% to 91% (Table II). Those studies have in common that beyond excellent response rates regional toxicities were reported to be very low, thus resulting in an extraordinarily high limb salvage rate ranging from 58% to 87% [38], [39], [66–68],[71], [75–81].

Table II.  Local response rates and limb salvage rates for non-resectable limb sarcoma (TM-ILP).

	ILP-setting	Reference	Patients	Response rates (%)			Limb salvage
	Drugs		(n)	CR	PR	NC/PD	(%)
1	TNF- α + IFN^1 + mel	Eggermont et al., 1996	195	18*	64*	18*	84
				36^#	51^#	13^#
2	TNF-α + mel	Eggermont et al., 1999	270	28^†	48^†	24^†	76
3	TNF-α + mel	Eggermont et al., 1999	196	17^†	48^†	35^†	71
4	TNF-α + mel (+ IFN^1)	Gutman et al., 1997	35	37^#	54^#	9^#	85
5	TNF-α + mel	Olieman et al., 1999	34	35^#	59^#	6^#	85
6	TNF-α + mel	Lev-Chelouche et al., 1999	53	38*	54*	8*	85
7	TNF-α + dox	Rossi et al., 1999	20	26^‡	64^‡	10^‡	85
8	TNF-α + mel	Lejeune et al., 2000	22	18^#	64^#	18^#	77
9	TNF-α + mel (+ IFN^2)	Noorda et al., 2003	49	8^#	55^#	37^#	58
10	TNF-α + mel	Lans et al., 2005	30	20^#	50^#	30^#	65
11	TNF-α + mel	Grunhagen et al., 2005	64	42^#	45^#	13^#	82
12	TNF-α + dox^3	Rossi et al., 2005	21	55^‡	35^‡	10^‡	71
13	TNF-α + mel^4	Bonvalot et al., 2005	100	49^⁁	17^⁁	34^⁁	87
				35^‡	22^‡	43^‡
14	TNF-α + mel^3	Grunhagen et al., 2006	217	16^#	68^#	16^#	97

Notes 1 Multicentre trial for approval of ILP with TNF-α/melphalan. 2 Judged as candidates for amputation. 3 Subset of patients from trial. 10 Patients with recurrent sarcoma in pre-irradiated fields. ¹Interferon was administered to the first 55 patients of the trial. ²Interferon was additionally administered to four patients. ³Reduced dosage of TNF-α (1 mg). ⁴Prospective randomized trial on dose-reduction. *Objective clinical response criteria according to World Health Organization. ^#Response based on clinical criteria or pathology (CR: clinical complete or 100% necrosis; PR: clinical PR or 50-99% necrosis; NC/PD: equal to clinical assesment or less than 50% necrosis). ^†Necrosis of 100% is recognized as a complete response by the European Agency for Evaluation of Medicine (EMA). ^‡Assessment of response only by histopathology: (CR: 100% necrosis; PR: 50-99% necrosis; NC/PD: less than 50% necrosis). ^{^}Complete and partial response by loss of vasculature by ultrasound/MRI.

Several studies even demonstrated that TM-ILP can be successfully applied in local recurrent sarcomas with previous surgery and/or radiotherapy [74], [76], [78], [79]. Because of its low toxicity profile and low overall complication rate TM-ILP may serve as the most useful neoadjuvant induction therapy in multimodality treatment concepts. Patients with overt metastatic disease and a limb threatening tumour seem to represent a patient category with a special need for TM-ILP. Extensive surgical interventions to control local tumour and relief symptoms are not indicated prior to systemic chemotherapy because of the delay in systemic chemotherapy [20]. TM-ILP does not hamper the onset of systemic chemotherapy because the local tumour does not get touched surgically thus not resulting in a delay of systemic treatment [20], [45], [60], [82]). Figure 2 demonstrates the typical clinical course of such patients where chemotherapy was almost simultaneously conducted with TM-ILP thus resulting in sufficient local control by complete devitalization and in stabilization of systemic disease by chemotherapy. The necrotic tumours can then be easily resected when systemic treatment is finished or the treatment schedule allows a limited surgical intervention [82]. Analogous to the utilization of chemotherapy, radiotherapy may be applied subsequent to TM-ILP and resection if the margins are not secure. Thereto several authors reported the feasibility of external beam radiotherapy after ILP and its potential to increase local tumour control [81], [83], [84].

Figure 2. Overt metastatic disease (A) of a pleomorphous fibrosarcoma in the thigh in a 72-year-old male with progressive palsy of the sciatic nerve (MRI, C). Onset of chemotherapy 10 days after iliac TM-ILP. Complete response int the thigh within six weeks (MRI, D) with impetuous recovery of the nerve. Partial response of the pulmonary metastases (A, B, E, F) following four cycles of chemotherapy (adriamycin/ifosfamide). Histology of the resection subsequent to completion of chemotherapy showed a 99% necrosis.

Conclusion and key points

Neoadjuvant treatment options are required to achieve sufficient local tumour control in advanced soft tissue sarcomas of the limbs. Preoperative radiotherapy and chemotherapy are well established and accessible at any oncology centre. However, experience from the past decade reveals that these treatment modalities are associated with a significant number of substantial complications as the response rate is limited to some extent [20]. TM-ILP offers an excellent treatment option to efficiently achieve local tumour control even in catastrophic situations, thus avoiding mutilation and amputation in most of these patients. Far more, TM-ILP is associated with a low rate and severity of complications. Although the logistic efforts in executing this treatment are more demanding it is thus far the most effective concept to control advanced local tumour and does not interfere with the necessity and time schedule of inevitable systemic treatment. TM-ILP is now established in more than 50 centres in Europe with growing demand for this treatment option. Besides the challenge to resolve the patent and licensing issue in North America in order to implement and offer this effective treatment, a prospective randomized protocol would be mandatory to gain even more scientific evidence.