One hundred and fourteen patients fulfilled all criteria and were enrolled into this study. There were 67 males and 48 females. The mean age was 30.1 years. Twenty-five patients (21.9%) were aged below 18. The mean patient and doctor delays were respectively 8.5 ±10.2 (1 to 60) and 3 ±4.3 months (0 to 24). The longest patient delay was observed in chondrosarcomas (CS) (p=0.001). Pelvic pain was the main symptom reported in 89.5% of cases.However swelling was present in half of the cases only. The tumourwas located in the PI+/-IV Enneking zones in 67 patients (58%) and PII zone was involved in 39 cases (34.2%). Three histological types were diagnosed: CS,Ewing sarcoma (ES) andOsteosarcoma (OS). Eighty-one tumours (71%) were localized and 33 (29%) were metastatic. Epidemiologic data are summarized in Table 1.
Surgery was performed in 76 cases (66,7%). It concerned 86.4% of the non-metastaticpatients (70 cases) and only 18.2% of the metastatic ones (6 cases) (p<0.0001). Internal hemipelvectomy was performed in 69 cases (90.5%). In four cases, the surgery was external hemipelvectomy. In three patients an intra-lesional curettage was done as a debulking procedure.
In PI+/-IV resections (37 cases) bone defect was reconstructed in 35 cases using free fibula grafting and in 2 cases we used bone cement and plate. In PI+II (9 cases) it was done sacro-femoral arthrodesis in 7 cases and ischio-femoral arthrodesis in 2 cases. In PII+III resections (6 cases), ilio-femoral (IF) arthrodesis was performed. Total hemipelvectomy (PI/II/III) was performed in 4 cases. Reconstruction was performed in 1 case using bone cement and total hip arthroplasty. In 3 cases, there was no reconstruction. In PII resections (2 cases) reconstruction was performed only in one case by IF arthrodesis. PIII (4 cases) and partial PI resections (7 cases) were not reconstructed. Surgical margins were R0 in 60 cases (79%) R1 in 13 cases (17%) and R2 in 3 cases (4%). Post operatively, major complications were reported in 8 patients. Wound necrosis and deep infection occurred in 4 cases. They were treated by surgical debridement and antibiotics in 3 cases. In one case of reconstruction with cement and arthroplasty, septic dislocation of the hip occurred and needed secondary hindquarter amputation. Sciatic palsy had occurred in 4 cases. None of the reconstructions was revised for mechanical failure.
Chemotherapy was adminestred in 77 cases (67,5%) mostly in ES and OS. It was associated to surgery in 44 cases (57.1%), it was combined to RT as definitive treatment in 9 cases and was used in a palliative attempt in 24 cases (18 of them were metastatic). Histological response was available in 33 patients who had neoadjuvant CT (23 Ew,8 OS and 2 CS). Good response was observed in 82% of ES. Seven patients with OS (87.5%) and 2 patients with CS had poor response (p<0.001).
Radiotherapy was used in 19 patients (16.7%) among them 11 patients had ES, 11 were metastatic and only 5 patients had had surgery (detailed therapeutic protocols are reported in table 1).
According to metastatic status, there were more surgery in non-metastatic patients (p < 0,001) and more RT in metastatic ones (p = 0,002).
- Oncologic results: (Tables 2 and 3)
Patients were reviewed with a mean FU of 32 months (SD: 46,5 – range: 1 to 216). At the last FU, 68 patients (59,6%) died after a mean time from the diagnosis of 15.9 months (SD: 22.8 – range: 1 to 120). All of them but one died from disease progression. Forty-six patients (40.4%) were alive with a mean FU of 56 months (SD: 60.6 – range: 6 to 216). Among them, 42 were disease free.
Local recurrence (LR) was observed in 19 patients (25%). The mean time from surgery to onset of LR was 11.05 months (SD 17.6 - range: 1 to 60). In univariate analysis, only inadequate surgical margins (R1 or R2) and poor response to CT were significantly associated to a high risk of LR. Resection of more than one zone was also associated to higher rate of LR but was not significant. The multivariate analysis showed that only surgical margins were independently associated to tumour recurrence (HR = 8.16, p=0.01, CI=2.29 – 29.03).
Among the 81 localized tumours at presentation, 33 patients (40.7%) presented lung metastasis at the last follow up. In univariate analysis, inadequate surgical margins, poor response to CT and patients managed without surgery were significantly associated to a higher risk of metastases.
Multivariate Analysis showed that surgery was associated with a decreased risk of metastasis with an OR of: 0.03 95% CI [0.01 - 0.10].
The overall survival rate at 5 and 10 years were 38.4 % and 27.6 % respectively. The 5YOS was significantly better in non-metastatic patients (fig. 1), in patients with CS (fig. 2), when patients were treated by surgery (fig. 3), when surgical margins were safe and in good responder patients.
In multivariate analysis, significant prognostic factors for death were histological type, the metastatic status and surgery. OS and metastasis increased the risk of death with a HR of 3.64 and 3.55 respectively. On the other hand, surgery was associated with a decreased risk of death with a HR of 0.12.
- Anatomical and functional outcome:
Forty patients had biological reconstruction. At the last FU, 19 were evaluated and in all of them we obtained bone healing. Functional score was evaluated in 33 patients. The mean MSTS score was 26.27/30 (range 15 to 30). Patients with partial PI resection (7 cases) had the highest score (30/30) and those with PI/II/III resection (2cases) had the lowest score (mean 20/30). Patients with PI resection (6 cases) and those with IF arthrodesis (3 cases) had a mean score respectively of 26,67/30 and 24/30.