Purpose: Compared to paediatric cancer patients adolescents and young adults may have disadvantaged access to care. Therefore we investigated the correlation of patient, tumour and institutional characteristics with the outcome of osteosarcoma in this age group.

Method: Analysis of consecutive patients aged 15–24 years with newly diagnosed high-grade osteosarcoma entered into the Cooperative Osteosarcoma Study Group(COSS) registry 1980–2004 and treated in pediatric (PO) or medical oncology institutions (MO). Standardised multimodal therapy according to a COSS-protocol. Event-free survival rates (EFS) evaluated in relation to patient demographics and registering institution (MO vs PO and treatment volume as: < 1, 1–3 or > 3 osteosarcoma/year).

Results: 944 patients identified (median age: 17.35 years; range: 15.01–24.99; 79% aged < 20 years). Patients > 20 years were more likely than younger patients to be treated in centers with low treatment volume (p< .0001) and MO (p< .0001) but otherwise comparable. After a median follow-up of 5.59 years (range: 0.12 – 27.92) for all patients and 8.08 years (range: 0.19 – 27.92) for 617 survivors, actuarial 5/10 year event-free survival probability (EFS) was 58%/54%. Upon univariate analysis of the total cohort neither of the institutional variables correlated significantly with EFS. There was a correlation between treatment in PO and improved EFS for patients > 20 years (p=.001) and for those with primary metastases (p=.009). Upon multivariate testing type of center (odds ratio: 1.26; p=.022) but not treatment volume were significant.

Conclusion: Within a framework of standardised regimens and consultation supportby our group’s infrastructure, similar EFS-probabilites were obtained regardless of institutional treatment volumes. Observed variations in outcome between PO and MO may be partly due to different distributions of presenting factors but deserve further investigation.