Clinical question 2: Does surgery at a specialised sarcoma centre improve outcomes?

  • Population: Adult and Paediatric patients with bone and soft tissue sarcoma 
  • Intervention: Multidisciplinary team, surgery
  • Comparison: Treatment at non-specialised centre
  • Outcomes: Local control, limb salvage rate, 30-day and 90-day surgical mortality, overall survival

Authors: Abay Sundaram, Ganaps Perianayagam, Angela Hong, and Sarcoma Guidelines Working Party

As it is not possible to conduct a randomised trial for this clinical question, available evidences are from analysis of series of patients treated within and outside specialist centres. This has potential bias as those patients with a good prognosis (for example a patient with a small, superficial tumour that is readily widely excised) are often not referred to specialised sarcoma centres whereas those with more advanced disease with complex treatment requirements usually are referred. Literature search and screening identified 66 studies for this clinical question, which are all retrospective in nature. There is a large body of publication using cancer registry data (the United States National Cancer Database, French National Sarcoma network, Netherlands Cancer Registry). Heterogeneity exists in these publications with respect to the definition of specialised sarcoma centres. Multiple studies compare high volume centres with low volume centres, but even amongst these studies, the volume of cases required to reach the threshold of high volume varies. Some studies represent all possible surgical centres longitudinally and divide them into terciles or percentiles before creating arbitrary cut offs for high and low volume.

Local control

Forty five studies draw conclusions with respect to local control when high and low volume centres or specialised and non-specialised sarcoma centres are compared (6, 7, 10, 11, 15, 19, 22, 25-61). There are 19 studies that identify differences in rates of local recurrence; 29 identify difference in microscopic (R1) and macroscopic (R2) surgical margins; 14 identify differences in locoregional recurrence-free survival while some studies drew conclusions for more than one of those parameters.

No randomised data is available due to the nature of the question. The evidence available for review was all retrospective analyses despite authors suggesting that patients were prospectively enrolled in registries  (31). A large proportion of studies utilised data gleaned from national or regional sarcoma registries or databases. These are exposed to errors in data entry or confirmation bias as specialised sarcoma services generally host these registries and draw attention to the disparities in standard treatment between their facility and a non-specialist facility. The addition of surgical margins to registries eliminates this bias as they are reported on by third parties and results are objective i.e. malignant cells are present or not. Surgical margins were included in the definition of “local control” due to correlation with local recurrence and locoregional recurrence-free survival and to objectivity  (34, 39, 47).

One Australian study was available, authored by Tan and colleagues in 2018 (22) of a prospectively-enrolled multicentre database study across Victoria (Peter MacCallum Cancer Centre) and New South Wales (Prince of Wales Sarcoma unit). The findings were consistent with international studies and showed that patients treated initially in non-sarcoma centres more commonly had initial positive margins and were significantly more likely to have a local recurrence than if treated initially in a specialised centre (24.1% at non-sarcoma centres vs 6.5% at specialised centres, p=0.038).

One study is of particular interest due to its applicability and scalability to our health care system and geographic challenges in Australia. Schmitz and colleagues performed a retrospective analysis of the National Cancer Database in America for retroperitoneal sarcomas treated between 1998 and 2012 (61). The distance patients travelled to treatment centres and the annual case volumes for these centres were divided into quartiles. They reported that patients who travelled shorter distances to low volume hospitals (4 miles to local medical centres with a median annual volume of 1 case per year) compared with those patients who travelled longer distances to high volume centres (56 miles to institutions with median annual case volumes of 10 cases per year), had significantly greater risk of R2 resections (macroscopic residual disease) (4.4% vs 2.6%, p=0.003).

The evidence available included a variety of histological subtypes, representing primary retroperitoneal sarcomas, soft tissue sarcomas of the extremity, bone tumours, and specific analyses of myxofibrosarcoma  (39, 44), chondrosarcoma of the rib (60), synovial sarcoma (52), malignant fibrous histiocytoma/unclassified pleomorphic sarcoma (47), liposarcoma (36), and giant cell tumour of bone (55). While this list of subtypes is not exhaustive, analyses and inclusion of greater than 150 potential soft tissue sarcoma subtypes is simply not feasible. There are technical nuances when performing resections of different subtypes with respect to margins and potential for local recurrence, however this only strengthens advocacy for initial resection at a specialised centre with a multidisciplinary team. Treatment at a specialised centre was synonymous with improved local recurrence-free survival and is likely to be multifactorial and related to surgical technique, experience, and tumour spillage during initial resection.

There is a paucity of evidence comparing local recurrence when treated at specialised or non-specialised centres for paediatric sarcoma. This may be explained by the well-established pathway of referring to specialist paediatric hospitals for management of complex paediatric issues including all oncological diagnoses. The systematic review identified two studies that included paediatric patients, and both had findings that were consistent with the adult population  (34, 52).

Only one study found no statistically significant difference in positive margin rates between high and low volume centres, however this study specifically assessed the outcomes of primary malignant bone tumours of the vertebral column (45). In this study patients treated at high volume centres were more likely to receive surgical treatment (91% vs 80%, p<0.0001) and, if surgery was performed, patients were more likely to undergo an en bloc resection (48% vs 30%, p<.0001). However, there were no differences in margin status between high volume centres and low-volume centres. This may be explained by the compromise between margin-negative surgery in the vertebral column and preserving neural elements and function. This is a similar finding to Hoekstra and colleagues, who noted lower rates of R2 resections (macroscopic residual tumour) in high volume centres but higher rates of R1 resections, with the rationale that microscopic positive margins may be justifiable to preserve functionality of a limb affected by locally advanced soft tissue sarcoma (7). Berger and colleague, found that overall survival following a margin positive (both R1 and R2) resection was improved for patients at specialist centre, potentially due to the use of adjuvant RT (29).

Irrespective of the definition of a specialised sarcoma service, the evidence from this systematic review remains consistent that definitive surgical management at specialised sarcoma centres improves local control as defined by local recurrence, positive surgical margin, or local recurrence free survival.

Limb salvage rate

The systemic review identified eleven retrospective studies reporting the limb salvage rate by treatment centres. Nine of these studies were on soft tissue sarcoma and two were on both bone and soft tissue sarcoma. Six studies including one from Australia clearly showed that the limb salvage rates were higher at specialised sarcoma centres (11, 50, 62-65). In a large series of 14,038 patients with primary bone sarcoma and the definition of 20 cases per year as high volume centre,  Malik et al. reported patients treated at high volume centres were more likely to undergo resection with limb salvage surgery than amputation (odds ratio 1.34, 95% CI 1.14 to 1.59,  p = 0.001) after adjusting for stage, size, grade and other patient and treatment factors. Gutierrez et al. reported the outcomes of 1,937 patients with extremity soft tissue sarcoma treated in Florida, patients treated at high volume centres were younger and were more likely to undergo limb sparing operation (90.6% vs 86.2%, p=0.048). These studies concluded that planned initial excision, systematic re-excision in a referral centre, depth of surgical experience and immediate availability of multidisciplinary and multimodality therapy in high volume centre, early presentation (<3 months) to a cancer centre were the key reasons for better limb salvage rates and oncological outcomes. 

For the five studies that did not show a significant difference in limb salvage rates between specialised sarcoma centres and non-specialised centres, three studies did not perform multivariate analysis adjusting for factors such as stage and grade (6, 35, 40).

30-day and 90-day surgical morality

30-day and 90-day mortality is widely used as a surgical outcome parameter to identify cases that are a “failure to rescue” following perioperative complications (43). Differences in 30-day and 90-day mortality between low- and high-volume hospitals are unlikely to be an immediate consequence of the sarcoma itself, rather, differences in mortality may be caused by inadequacies in perioperative optimisation, surgical technique, and post-operative access to critical care facilities, leading to perioperative complications. This postulation is supported by the study by Schmitz and colleagues 2019, which showed a difference in 30-day and 90-day mortality between patients that travelled a long distance to a high-volume centre compared with patients who travelled a short distance to a low-volume centre (66).

Systematic review identified 12 studies reporting the 30-day and 90-day mortality rates between high-volume and low-volume centres, all of which were retrospective reviews with variations in the definition of a high-volume or low-volume centres (10, 26-29, 42, 43, 58, 61, 62, 67, 68). Two studies commented directly of no differences in these parameters between high-volume and low-volume centres (27) or before and after centralisation of sarcoma services (42). Three other studies failed to show statistical significance despite a trend towards improved 30-day and 90-day mortality for high-volume centres, one of which was a study of 125 patients (15 patients in high volume centres and 110 patients in low volume centres) with desmoplastic small round cell tumours, a rare entity and thus was likely underpowered (67); and the other two had larger sample sizes but the difference in mortality was small and unlikely to be clinically significant (30-day 0.7% vs 1.5% and 90-day 2.3% vs 3.7% for high-volume and low-volume centres respectively) (68) & (90-day mortality 6.2% vs 6.4% for high-volume and low-volume centres respectively) (29).

The remaining seven studies showed a statistically significant lower 30-day or 90-day mortality or both. Five of these pertained to retroperitoneal sarcoma (10, 62). Lazarides et al. reported only on 30-day mortality, with high-volume 0.3% and low-volume 0.4%, a result that is unlikely to be clinically significant (10). Interestingly, the study by Gutierrez et al. (62), had a smaller sample size than the study by Villano et al. (4205 vs 7541) despite achieving statistical significance with the same reported difference (0.7% high-volume centre vs 1.5% low-volume centre, p = 0.028). This difference may be explained by Villano et al. employing a multivariate analysis with logistic regression for 30-day and 90-day mortality, whereas Gutierrez et al. performed a univariate analysis utilising a chi-square test to determine statistical significance (62, 68).

Overall survival

Thirty-four publications reported overall survival endpoint according to surgery at a specialised sarcoma centre or not. Of these, 21 publications included surgery at a specialised sarcoma centre as a variable in multivariable analysis. There is no Australian data addressing this endpoint. Sixteen studies showed that surgery at a specialised sarcoma centre was a factor predicting an overall survival benefit in multivariate analysis.

Ten of these studies used the NCDB with large number of patients (range 1266-25,406) including soft tissue sarcoma of extremity, retroperitoneal sarcoma and chordoma (10, 18, 28, 43, 48, 57, 58, 61, 69, 70). These studies examined overall survival by surgical volume (minimum number of case per year or top one percentile), and all showed a reduction in the risk of death when surgery was performed at high volume centre. Villano et al. reported that overall mortality risk was reduced by 4% per additional case of retroperitoneal sarcoma (HR 0.96, 95%CI 0.95 to 0.98) up to a threshold of 13 cases per year (58). In extremity soft tissue sarcoma, surgery at a high volume centre with more than 20 cases per year was associated with reduced mortality (HR 0.81, 0.75-0.88, p<0.001) (10).

There are two prospective cohort studies using data from the French National Cancer Institute funded clinical network for sarcoma (NETSARC) (31, 32). In retroperitoneal sarcoma, surgery in a NETSARC centre was an independent predictor for better overall survival, with a two-fold lower odds ratio of death compared with surgery outside (OR: 0.496, p=0.001)(32).  In extremity soft tissue sarcoma, surgery in a NETSARC centre was found to be consistently associated with a reduction in the risk of death (HR 0.681, 95% CI 0.618–0.749, p<0.001) (31). The favourable prognostic value of surgery in a NETSARC centre was retained in addition to the presentation to a NETSARC multidisciplinary tumour board. In the Netherlands, surgery for soft tissue sarcoma in a high volume hospital had a significant beneficial effect on survival compared with surgery in a low volume hospital (RR 1.3, 95% CI 1.02-1.6,p=0.03) (59).

There are only two studies on primary bone tumour.  Hu et al. reported the outcomes for 182 patients with osteosarcoma around the knee (71). Overall survival was significantly better when both the initial biopsy and resection were performed at the sarcoma centre (HR 2.8, 95%CI 1.503-5.207, p=0.001). For chordoma, treatment at an academic centre (overall cancer case load of >500 per year and with postgraduate training in 4 or more specialities) was associated with improved 5-year overall survival  compared with treatment at community cancer centre (76.08% vs 52.7%, p<0.0001) (18).

There are five studies that did not identify surgery at specialised centre/or by specialised sarcoma surgeon to be a factor for better overall survival in multivariate analysis  (46, 49, 63). Toulmonde et al. compared surgical outcomes of 586 patients with retroperitoneal sarcoma by specialised surgeons vs non-specialised surgeons (56). Surgery performed by a non-specialised surgeon was a factor associated with higher rate of R2 resection and development of abdominal sarcomatosis (HR 0.5, 95% CI 0.3–0.96], P = 0.04) but not a factor associated with better overall survival. The study by Traub et al. was a comparison between unplanned excision elsewhere with planned excision at sarcoma centres but all patients (n=500) had further treatments at one of two sarcoma centres (65). In the multivariate analysis, no parameter was associated with increased  risk of overall, metastasis free, or local recurrence free survival rates. The authors concluded that as a result of aggressive re-excision and multidisciplinary treatment, a negative effect of unplanned excision outside a sarcoma centre on oncologic outcomes could not be confirmed.

In summary, the systematic review shows that the literature is consistent in supporting a better survival when surgery is performed in specialised sarcoma centre.

Evidence summary



There have been on randomised trials comparing the outcomes between treatment at specialised sarcoma centres with non-specialised centres.

  N/A  -

Definitive surgery at specialised sarcoma centres improves local control as defined by lower rates of local relapse, higher rates of margin negative resections, and improved local recurrence free survival.


  (6,7,10, 11, 15, 19, 22, 25-61)

The limb salvage rate for patients with extremity bone and soft tissue sarcoma treated at specialised sarcoma centres is generally higher.



  (6, 10, 11, 35, 40, 50, 62-65, 69)

Evidences suggest a favourable pattern of lower 30-day and 90-day mortality when sarcoma surgery is performed in specialist sarcoma centres compared with non-specialised centres.


  (10, 26-29, 42, 43, 58, 61, 62, 67, 68)

Overall survival is greater in adult patients with sarcoma undergoing surgery in high volume, specialised sarcoma centres.

  III-2,      III-3

  (10, 18, 28, 43, 48, 57, 58, 61, 66, 67)

Evidence-based recommendation


Patients with suspected sarcoma to be referred to a specialised sarcoma centre for surgical management to reduce the risk of local recurrence, surgical complication, and to improve limb salvage and survival.


Practice Point

Patients with suspected sarcoma should be referred to specialized sarcoma centre early for management including planned biopsy.

*National Health and Medical Research Council. NHMRC levels of evidence and grades for recommendations for developers of guidelines. Canberra: NHMRC; 2009. (

To view the administrative report click here

To view the technical report click here