Regional chemotherapy by isolated limb perfusion prior to surgery compared with surgery and post-operative radiotherapy for primary, locally advanced extremity sarcoma: a comparison of matched cohorts

Extremity soft tissue sarcomas (ESTS) are rare tumours comprising over 50 different histological subtypes [1, 2]. Optimal management of locally advanced ESTS requires multimodal therapy and the precise role of isolated limb perfusion (ILP) in the overall treatment strategy remains to be fully defined. For small or superficial ESTS, treatment usually involves surgery alone, gaining wide surgical margins while preserving function. Adjuvant radiotherapy is reserved for high-grade tumours greater than 5 cm or small tumours that focally involve margins adjacent to a critical structure [3]. When sarcomas attain even greater dimensions, such as over 10 cm, it becomes increasingly difficult to achieve negative margins without recourse to amputation or function-limiting surgery. While amputation for ESTS does not improve survival for sarcoma over limb-conserving surgery, it may still be considered for large sarcomas when the risk for local recurrence is viewed as significant [4‐6].

Induction chemotherapy by ILP using melphalan with recombinant human tumour necrosis factor alpha (TNFα) prior to a limb-conserving surgical resection was introduced as a strategy for locally advanced sarcomas considered irresectable other than by an amputation [7]. TNFα is a multifunctional cytokine, which causes increased vascular permeability, associated with increased extravasation of cytotoxic agents and selective destruction of tumour-associated vessels by endothelial apoptosis and inflammation [8‐10]. A multi-institutional case series of patients with locally advanced sarcomas, considered to be irresectable with limb-conserving surgery and adjuvant radiotherapy, which were treated using ILP with TNFα reported a limb salvage rate of 84% [7]. In light of these results, TNFα has been licensed for this indication since 2006.

ILP is now used more widely in large, high-grade tumours that are compatible with limb-conserving surgery but risk positive resection margins. Such patients would undergo an induction ILP prior to a wide resection of the primary tumour, with the aim of gaining as wide a margin as is compatible with preserving limb function. Post-operative radiotherapy is then typically only offered if the pathology specimen demonstrates viable tumour at a compromised margin. The alternative, which might be considered the standard management, would be a function-preserving wide resection with adjuvant radiotherapy to compromised margins. It is known that with adjuvant radiotherapy, planned microscopic positive margins over a critical structure are fully compatible with long-term local control [11].

In a retrospective cohort analysis, we sought to determine whether the peri-operative morbidity and oncological outcomes differed between these alternative multi-modality approaches to large high-grade sarcoma of the lower limb.

Standard treatment for locally advanced ESTS consists of limb-conserving surgery with adjuvant radiotherapy that may be delivered pre or post-operatively [15]. The role of (neo)adjuvant chemotherapy outside of specific chemo-sensitive subtypes in ESTS is controversial. In the EORTC 62931 study, Woll et al. randomised patients with localised, grade II or III extremity sarcoma to receive adjuvant cheomotherapy in the form of doxorubicin, ifosfamide and lenogastrim in addition to surgery, radiotherapy and, if appropriate, ILP [16]. No benefit in terms of relapse-free or overall survival with this adjuvant regime compared to the control cohort were identified (5-year OS 66.5% vs 67.8%). More recently, in the ISG-STS 1001 study, Gronchi et al. randomised patients with localised, high-grade extremity sarcoma of 5 specified subtypes to receive neoadjuvant standard chemotherapy, in the form of epirubicin and ifosfamide, or histotype-tailored regimes [17]. This trial closed early after an interim analysis demonstrated no benefit in the histotype-tailored regimes. However, at 48 months, overall suvrival in the standard cohort was 89%, which suggests a potential benefit to neoadjuvant standard chemotherapy in these histological subtypes. As of yet, there is consensus regarding the role of (neo)adjuvant chemotherapy and its use is not considered standard in the most recent guidelines [18]. In contrast, induction chemotherapy with ILP is widely recognised to produce markedly better response rates than systemic chemotherapy and has a well-established role in facilitating function-preserving resections in locally advanced ESTS that would otherwise require amputation [7, 19‐24].

With induction ILP increasingly used for ESTS that may be amenable to standard surgical management, the question as to which strategy, if either, is superior has arisen. Previous case series have clearly shown that induction ILP prior to surgery is an effective approach to deal with compromised surgical margins, as wide surgical resections were not possible. Similarly adjuvant radiotherapy has been shown to be effective in preventing local relapse after positive surgical margins [25]. The ability to directly compare these treatment strategies is hampered by the rarity of locally advanced non-metastatic primary ESTS and the scarcity of specialist centres performing ILP. As such, a randomized study comparing these strategies is not feasible [26]. Although subject to the limitations of any retrospective study, this comparison of matched cohorts provides valuable evidence to compare these two approaches.

The present study has found that the peri-operative and oncological outcomes of patients undergoing induction ILP prior to surgical resection are very similar to those undergoing standard surgical management. No significant difference was found in the rate of peri-operative complications between cohorts. Although two patients in the ILP cohort required amputations, only one was due to procedure-related complications. Similarly, no significant difference was found in the rates of local recurrence between these treatment arms, despite the use of adjuvant radiotherapy being significantly less frequent in the ILP cohort. There are no clear guidelines regarding the use of radiotherapy following ILP, although it is typically considered following an inadequate response on histological assessment of the specimen. However, following a microscopically complete resection with over 50% necrosis in the specimen, adjuvant radiotherapy is unlikely to be of additional benefit [27].

As a retrospective study, this series is subject to the bias inherent with this methodology. However, analysis of the tumour characteristics in terms of grade, size and histological subtype showed that the cohorts were well matched and this is reflected in the identical rate of metastatic spread. Similarly the assessment of the tumours “resectability” by independent experts indicated that the technical difficulty of surgery appeared to be similar in both cohorts. Although the cohorts were reasonably well matched, they were not randomized and we identified some important differences. There was a higher proportion of popliteal fossa and leg tumours in the ILP cohort. It is generally accepted that achieving local control for sarcomas is more challenging in the distal than the proximal extremity. Therefore, this may represent a bias towards patients with more challenging tumours being referred to specialist centres providing ILP. A further limitation to this study is the inability to comment on the post-operative limb-function in each cohort, an additional important factor that may influence the choice of treatment strategy in these patients.

In the absence of any significant difference in outcomes between these treatment strategies, the morbidity associated with their use becomes increasingly important. Induction ILP carries the risks associated with an additional operation and short term toxicity associated with regional chemotherapy. That being said, the peri-operative morbidity associated with ILP is generally very low. Severe regional toxicity following ILP is rare occurring in 2–15% and the need for amputation even more so with rates of 0–2% [28]. The long-term complications from external beam radiotherapy, the technique used in the majority of patients in this study, are well documented [29]. However, during the period of this study, advances have been made in radiotherapy techniques. When given pre-operatively, the dose of radiotherapy may be reduced and the use of intensity-modulated radiotherapy, which is associated with lower toxicity rates, is becoming more widespread [15, 30]. Even so, the long-term sequelae of radiotherapy, including the risk of second malignancies, remain significant and are of particular importance in young patients [31].

Induction ILP followed by a wide resection provides an alternative treatment to standard therapy in the management of locally advanced ESTS. Consideration should be given to induction ILP in young patients with large, high-grade extremity sarcomas who would benefit most from avoiding radiotherapy.