The present work used the French nationwide prospective database NETSARC to assess the survival of young patients 15–30 years with sarcoma diagnosed between 2010 and 2017. To the best of our knowledge, this study is the 1st to report the survival of young (15–30 years) patients treated for sarcoma at a national level. The present study reported 3y-OS rates in AYA patients with sarcoma of almost 80% which is consistent with the 2y-OS of 80% reported by Raze et al. in 2016 . The study did not identify significant differences in OS between AYAs treated in RSC and in non-RSC, but being treated in RSC is associated with improved 3 y-LRFS and PFS. Of note, the present study considered in the AYA population all patients with age ranges from 15 to 30 years i.e. extended range compared to the usual 15–24 years as defined by French National Cancer Institute (INCa). Indeed, epidemiology based on advanced biological and histopathological characterisation of AYA neoplasms [21, 22], and sarcoma incidence related to pediatric histology (which stays significant up to the age of 30 years) both support the rationale for considering young adults 25–30 years in the group of AYAs [4, 23]. In addition, sarcoma patients aged from 25 to 30 years are still often managed in non-expert centers at the present time, and careful review by MDTB involving both pediatric and adult oncologists would also be advisable in these young adult population [12, 16, 24,25,26].
The initial management of patients with sarcoma is a highly sensitive issue. Reference centers showed better compliance to international guidelines and notably performed pathological review to confirm the absence of microscopically residual disease (R0) [5,6,7, 27, 28]. According to our results, the initial management in non-RSC showed less compliance with clinical practice guidelines with less pre-treatment procedures, including biopsy and imaging reported in only 48.1% and 56.5% of the patients, respectively. In addition, whereas the quality of first surgery in RSC is consistent with previous results (60% R0, 20% R1 and 5% R2) , the quality of surgery in non-RSC revealed only 20% R0 and up to 20% R2 although patients presented less negative prognostic criteria. In addition, the substantial reoperation rate may result from inadequate and/or inappropriate initial surgeries, and potential correlation with the lack of pre-treatment biopsy and imaging can be raised. Indeed, second resection after first macroscopic residual resection (R2) occurred in 44% of the AYAs first operated in RSC, and in up to 64% of the AYAs first operated in non-RSC. Among second resections performed after incomplete initial resection, re-operations were most often performed in expert centers (AYAs in RSC: 70%; AYA in non-RSC: 48%). These results support early referral to expert centers for initial surgery, and confirm results previously reported . Whereas reoperations are mostly performed in RSC after first resection in non-RSC, few AYAs (n = 5) with first surgery in RSC were reoperated for incomplete resection in a non-expert center in this series. Nevertheless, such situations remain marginal and limited, and voluntary patient transition to non-RSC treatment centers for personal reasons cannot be excluded.
The quality of surgery is known as a major prognostic factor for relapse-free survival and overall survival in bone and soft tissue sarcomas [29,30,31,32,33,34]. However, in France as in the majority of European countries so far, the initial management of any sarcoma patient may be carried out in non-oncologic-specialized clinic, regardless of sarcoma expertise or number of sarcoma patients treated. Conversely, in the Scandinavian countries and the United Kingdom, a patient with sarcoma must be managed, upon suspicion, in an expert center [28, 35]. Resection in one of the 26 national expert centers reduced the risk of relapse by almost 35% in 35,784 sarcoma patients compared to those in non-expert centers . Even if the difference is less clear than previously reported, our series showed that 3y-LRFS rates were 82% in AYAs managed in RSC compared to 71% in non-RSC. In addition, expert centers provide high-quality medical management including surgery complementary treatments. The highest level of expertise of the treatment center has to be required in young patients with sarcoma, and same requirements should also apply to the total young patient population < 30 years  to ensure accurate management and facilitate access to clinical trials . This issue of particular concern prompted the French National Cancer institute INCa to create in 2011 the French academic society for AYA Adolescent and Young Adult Oncology and Hematology Group (GO-AJA) gathering pediatric and adult oncologists and hematologists at the national level . However, at local level, patient management remains heterogenous, and young patients aged from 15 to 24 years may still be treated in any unit with expertise in oncology, with specific accreditation requirements for pediatric oncology to treat patients aged 15–18 years .
PFS and LRFS after adjustment for negative prognostic criteria were better in AYA patients treated in RSC compared to AYA managed in non-RSC (HR 0.83 and HR 0.58), respectively. Differences in PFS, and especially in LRFS, may result from a lower compliance with clinical practice guidelines at initial surgery, putting patients at risk for more frequent local and distant relapses. In addition, patients with higher grade tumors had worse PFS (HR 1.88) and LRFS (HR 1.82) compared to patients with grade 1–2 tumors which may reflect the increased aggressiveness of the disease both locally and at distant sites.
The absence of differences in OS may result from the current relatively short-term follow-up; a median follow-up of 39 months may be still insufficient to detect a significant difference in OS considering the 2y-OS of 80% and 5y-OS of 60% in AYA patients with sarcoma, aged 15 to 24 years in France . Moreover, the influence of other non-observed potential negative prognosis factors in NETSARC centers cannot be excluded. Differences were evidenced in AYA characteristics in RSC and in non-RSC; we reported more bone sarcomas, lower limb tumors, large tumors, and grade 3 tumors in AYAs in RSC. Conversely, AYAs in non-RSC had more STS, non-grade 3 tumors, and smaller tumor sizes. This heterogeneity in patient characteristics may lead to the observed patient referral in France: patients presenting with negative prognostic criteria (grade 3, larger size, and metastatic status at diagnosis) are easily diagnosed and therefore referred to an expert center. As a consequence, a majority of bone sarcomas and lower limb tumors are managed in expert centers. The first actor in the patient management is most often an orthopaedic surgeon, which may be reluctant for resection of bone tumor presenting negative clinical or paraclinical prognostic criteria. Thus, expert centers are more likely to be consulted for diagnosis and therapeutic advice. For patients presenting with small /superficial STS is more likely to be directed to digestive or plastic surgeon in non-expert sarcoma centers, more often confronted with benign STS, and decision for resection adopted without resorting to further expertise, neither pre-surgery review for more accurate diagnostics.
The AYA population in this study showed more STS (60%) than bone sarcomas (40%), which contrasts with the rates previously reported by Raze and colleagues in 2018 in the 15–24 years old patients (bone sarcoma: 53%; STS: 47%) . This shift to increased STS rate in our overall AYA population may result from the STS/bone sarcoma ratio of 9/10 reported in adults, and therefore supported in our series by the 25–30 year old population .
Limits of our study
The NETSARC database does not collect all medical treatments chemotherapy and targeted therapy administered before and after surgery, and do not allow to explore the impact of medical treatments administered in RSC and non-RSC so far. However, from 2016, additional data collection showed that neoadjuvant treatment administration was mostly observed in AYA in RSC (56.9%) and rarely in AYA patients in non-RSC (14%), which may have contributed to better local control and to increase 3y-LRFS in AYA in RSC. Detailed information regarding neoadjuvant treatment (systemic anticancer drugs, radiotherapy or combination thereof) are not specified and impact on 3y-PFS in AYA in RSC are not accessible.
Comparisons in sarcoma management in AYAs are a critical issue not only regarding age ranges but also sarcoma histotype representations. If embryonal tumors are obviously more common in children than in AYAs, representation of sarcoma histotypes in AYA is closer to pediatric than to adult sarcomas. Whereas pediatric sarcoma treatment is exclusively performed in reference centers in France, and showed a quality of management recognized as a robust reference, all the AYAs up to 30 years may also derive benefit from a similar high-quality management. Subgroup analyses according to histotype would be highly informative, but have not been achieved so far, still limited by the high heterogeneity of sarcomas.