Purpose: To further elaborate on prognostic factors for Ewing's sarcoma of bone and to document improvements in relapse-free survival (RFS) and trends in local therapy over the study period (1977 to 1993).

Patients and methods: A retrospective analysis was performed on a combined Gesellschaft Für Pädiatrische Onkologie und Hämatologie/Cooperative Ewing Sarcoma Study and United Kingdom Children's Cancer Study Group/Medical Research Council data set of 975 patients registered with the respective trial offices before the current collaborative European Intergroup Cooperative Ewing's Sarcoma Study trial. Both groups independently undertook studies with similar chemotherapy during the period.

Results: The key adverse prognostic factor is metastases at diagnosis (5-year RFS, 22% of patients with metastases at diagnosis v 55% of patients without metastases at diagnosis; P: <.0001). For the group with metastases, there was a trend for better survival for those with lung involvement compared with those with bone metastases or a combination of lung and bone metastases (P: <.0001). In the group of patients with no metastases at diagnosis, multivariate analysis demonstrated that site (axial v other), age-group (< 15 v > or = 15 years), and period of diagnosis had significant influence on RFS (all P: <.005). RFS was superior in the period after 1985 compared with the period before 1985 for nonmetastatic patients (45% v 60%, respectively; P: <.0001) and for metastatic patients (16% v 30%, respectively; P: =.016). Patients who relapsed within 2 years of diagnosis had a less favorable prognosis than patients who relapsed later (5-year survival after relapse, 4% v 23%, respectively; P: <. 0001). There were other changes over the period; in particular, radiotherapy or amputation were more common in the period before 1986, whereas endoprosthetic surgery was widely used in the later period.

Conclusion: Survival and RFS improved over the period. Prognostic factors are metastases at diagnosis, primary site, and age.