Defining resectability criteria for primary retroperitoneal sarcoma: a challenging imperative

Retroperitoneal sarcoma (RPS) represents about 15% of all soft tissue sarcomas and carries the most unfavourable prognosis. Resection continues to be the cornerstone of curative-intent treatment for primary RPS, but the frequently intimate relationship with central compartment anatomical structures typically makes it challenging to safely achieve wide margins. Preoperative radiotherapy has been invoked as a potential solution to the local control problem, but the current evidence is widely debated. Given the central role of resection, calls to standardize criteria for resectability abound. Repeatedly, the aspirational model cited is that of pancreatic cancer, where anatomic criteria define categories of resectable, borderline resectable, and unresectable disease. Although there are natural parallels between RPS and pancreatic cancer, in the face of a rare disease with immense histologic diversity and a dearth of data, such an approach oversimplifies the nuances relevant to surgical decision-making given the unique features of RPS that continue to challenge even the experienced surgeon. Abutment of key organs and vascular structures by a large (>20 cm) mass often necessitates multivisceral resection as well as vascular resection and reconstruction. Each tumour is akin to a fingerprint, with no two identical operations. Although attempts to provide a standardized sequential approach to resection1 may be conceptually attractive, such fixed perspectives do not fully capture the distinctive challenges of assessing resectability, both pre- and intraoperatively. Criteria for non-resectability that have been used for clinical trial exclusion comprise involvement of the superior mesenteric artery, aorta, or bone2,3, giving the impression of agreed-upon standards. In reality, perspectives vary, even among highly specialized centres. In a survey of Transatlantic Australasian Retroperitoneal Sarcoma Working Group (TARPSWG) members, there was no unanimous agreement on absolute contraindications to resection other than involvement of the first three jejunal branches of the superior mesenteric artery (SMA) such as to require enterectomy. It is recognized by most experienced RPS surgeons that combining multiple complex major visceral resections, such as pancreaticoduodenectomy with extended right hepatectomy, in addition to the standard nephrectomy and colectomy, leads to a notably increased risk of serious complications (Fairweather et al., British Journal of Surgery 2025, in press), without the clear promise of better disease control. Nevertheless, such combinations are sometimes proposed and executed, particularly in the young, fit patient. Similarly, most, but not all, would agree that resectability is precluded by extension of disease into the contralateral retroperitoneal space such as to necessitate bilateral nephrectomy. Reflecting these varied perspectives, retrospectively collected data on all patients with primary RPS evaluated at 10 high-volume centres reveals marked heterogeneity in the proportion of patients deemed resectable according to the individual centre4,5. In a study of RPS-dedicated multidisciplinary team conferences in the UK, the conclusions reached about resectability differed between centres for over half of cases presented6. Judgement about resectability remains opinion-based, varying with surgeon experience in, comfort with, and fundamental attitudes to the complex resections required in advanced cases of primary RPS. One resource that may help harmonize and elevate local decision-making is the monthly TARPSWG-sponsored RPS tumour board that is freely available to all members of the international community. Naturally, embarking on a major multivisceral resection requires consideration not only of technical factors but also of patient factors. Nearly half of patients with primary non-metastatic RPS who do not undergo resection exhibit prohibitive performance status/advanced co-morbidities rather than technical unresectability7. In addition, half of patients presenting for resection are malnourished, a modifiable factor known to be associated with increased postoperative complications8. Similar considerations exist for pancreatic cancer and are now incorporated into updated consensus definitions of borderline resectable disease (for example Eastern Cooperative Oncology Group performance status ≥2)9. However, patient factors relevant to RPS resection extend beyond those that can fit into an algorithm or preoperative score to guide resectability. What is often not discussed and infrequently studied is the patient perspective on undertaking such major and life-altering operations. While 30-day postoperative morbidity and mortality rates are concordant with those of other major operations (16.4% and 1.8% respectively10), the field is plagued by a paucity of data, as little is known about long-term quality of life and functional impairment11. A prospective study of patient-reported outcomes following primary RPS resection demonstrated significant deteriorations in physical functioning after surgery and high rates of neuropathic pain; however, global quality of life, as measured by the EORTC QLQ-C30, was comparable to the general population, possibly indicating the insensitivity of generic tools in this population12. No RPS-specific quality of life tools currently exist, limiting our understanding of the impact of such operations on long-term non-oncological outcomes. Ultimately, resection requires a motivated patient and therefore consideration of how the individual weighs disease extirpation against preservation of organ function, potential functional debility, and consequent impacts on long-term quality of life (that said, quality of life with a large RPS in situ is usually not normal). A definition of ‘resectability’ should inherently consider the patient perspective—criteria based on anatomic factors and patient performance status alone are unhelpful if the consequences of resection are not acceptable to the patient. Taken together, the lack of quality of life data, the complex nature of resection, and the potential for serious complications create a major challenge in providing patients with adequate information to make an ‘informed’ decision. In this regard, experienced surgeons may come to question whether the consent obtained preoperatively can ever be truly informed, even if the surgeon makes a sincere effort to elicit patient values and preferences13. Consensus criteria for resectability of various GI cancers now incorporate tumour factors that reflect biology and predict treatment response as well as long-term prognosis. Here too, RPS tumour biology is remarkably heterogeneous, with somewhat limited data to guide understanding of tumour behaviour. The term ‘soft tissue sarcoma’ refers to an aggregate of >100 histologically diverse tumours; generally, lymphatic spread is rare and proclivity for distant metastasis varies by tumour histologic type14. Even if we consider only the most common histologies seen in RPS, relevant biological criteria differ considerably among subtypes and a one-size-fits-all approach is not likely to be helpful. Unlike for cancers such as hepatoma, liver metastasis from colorectal adenocarcinoma, and pancreatic cancer, where biomarkers such as alpha fetoprotein (AFP), carcinoembryonic antigen (CEA) and carbohydrate antigen (CA19-9) are associated with disease extent and prognosis, no reliable biomarkers currently exist for prognostication in RPS. The Complexity Index in Sarcomas (CINSARC)15 transcriptional signature has not proven valuable for stratifying risk in RPS. Without the ability to predict response to neoadjuvant therapy, the relevance of defining a borderline resectable group is questionable. Analysis of biobanked specimens collected as part of the STRASS2 trial2 will provide an opportunity to systematically correlate genome-level properties with response to neoadjuvant treatment in high-grade RPS. Despite the caveats, the establishment of objective criteria for unresectability would be of benefit. Currently, in advanced cases, the same patient may be treated with curative intent at one sarcoma centre and with palliative intent at another6, a situation that is not acceptable as we seek to achieve equitable access to high-quality cancer care. Additionally, generating useful data to guide the curative treatment of patients with RPS requires criteria for trial enrolment, including factors governing surgical resectability. An obvious dilemma then exists—establishing resectability criteria requires more data, but the generation of data is itself predicated on the existence of agreed-upon criteria for resectability. The challenge inherent in developing evidence-based treatment algorithms for RPS is clear from the high variability in tumour behaviour, combined with limited good-quality prospective data. In this regard, we anticipate a major advance as data accumulate in RESAR, the international collaborative prospective RPS registry. Even then, decision-making for the individual patient will consider their performance status, motivation, and perspective, along with the potential morbidity rate of the required procedure as judged by an expert and experienced surgeon, and weighed against the individual patient’s specific tumour biology (Fig. 1). Even as we continue to strive together to develop agreed criteria that define resectability in RPS, we must embrace a holistic approach that tailors treatment decisions to individual circumstances. Intersection of technical, tumour, and patient factors in determination of candidacy for surgical resection Although for straightforward/mildly challenging cases, tumour and patient factors may not weigh heavily, with increasing technical difficulty, surgical candidacy requires increasingly favourable tumour biology, patient motivation, and performance status. None. The authors declare no conflicts of interest. Available upon request to corresponding author. Fahima Dossa (Conceptualization, Writing—original draft, Writing—review & editing), and Carol J. Swallow (Conceptualization, Supervision, Writing—original draft, Writing—review & editing)

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.