Purpose of the study: The rate of failure can be high for massive reconstruction prostheses after tumor resection. We studied the causes and possible factors of failure.

Material and methods: The series included 91 patients who underwent surgery from 1972 to 1994 for resection of a bone tumor involving the knee joint. A GUEPAR prosthesis was implanted in all cases for reconstruction (megaprosthesis in 58 cases and composite prosthesis in 33). The extensor system had to be reconstructed in 37 patients. A GUEPAR II implant was used in 73 patients; 48 of these implants had an antirotation system. The analysis was retrospective. Outcome was studied in terms of survival and independent factors predictive of failure unrelated to the tumor.

Results: Mean follow-up was 72 months. At last follow-up, 68 patients were living disease free. There were nine cases of rupture of the extensor system. Preservation of a continuous extensor system at the time of bone resection reduced the risk of rupture (p=0.036). Seven allografts fractured, two loosened, and six became infected. Use of an allograft did not reduce the risk of loosening (p=0.17). Intraxial laxity was observed in 17 patients. Use of an antirotation system was a factor of risk of intraxial laxity (p=0.0023) but not of aseptic loosening. Aseptic loosening was observed in 18 patients: 10 femur reconstruction and 8 tibia reconstruction. The difference was not significant (p=0.6). In all, 104 revisions were required in 53 patients; 36 revisions of the prosthesis, 23 of them for mechanical causes. Overall median survival, excepting tumor-related causes, was 130 months. It was 130 months for femur reconstructions and 117 for tibia reconstructions (p=0.57). Age, length of resection, tumor location, use of an allograft, and use of an antirotation system were not found to be significant prognostic factors for implant survival.

Discussion: As reported by many others, we found that the rate of failure of massive prostheses for infectious and mechanical causes remained high in patients treated for bone tumors involving the knee joint. Survival of massive implants is much lower than that of gliding prostheses.

Conclusion: Technical progress is required to improve the survival of massive implants used for the treatment of bone tumors involving the knee joint.

Purpose: The purpose of this work was to analyse and compare survival in patients with osteosarcoma (OS) or Ewing sarcoma (EW) of the pelvis as a function of treatment.

Material and methods: This retrospective series included 31 patients with OS (n=15) or EW (n=16) of the pelvis who were given a homogeneous therapeutic sequence associating chemotherapy, surgery and/or radiotherapy. Mean follow-up was 37 months (2–144). Mean age was 20 years for EW and 28 years for OS. Localisations in the pelvis were: zone I (n=12), zone I and II (n=4), zone II (n=1), zone II and III (n=7), zone III (n=1), and zone I, II and III (n=6). All patients were given chemotherapy, 15 underwent surgery, and 16 were given radiotherapy alone. Five patients were given complementary radiotherapy after surgery. Actuarial survival curves were compared with the logrank test. Comparison factors were presence of surgical resection, presence of initial or secondary metastasis, tumour response (radiographic measure), and pathology (good or poor responder) after chemotherapy.

Results: Five-year survival rate for patients with EW was 53%, 31% for OS. There was no significant difference in survival rates between tumour type. The only factor significantly correlated with lower survival rate was presence of initial metastasis.

Discussion and conclusion : The pelvic localisation of osteosarcoma and Ewing sarcoma is a factor of poor prognosis. Unlike data reported in the literature, surgery did not appear to influence outcome, not being found to be a factor of better prognosis. Surgery does however appear to improve short-term survival. In the pelvic localisation, osteosarcoma appears to have a poorer prognosis in terms of survival than Ewing tumour.