Outcome of primary soft tissue sarcoma of the knee and elbow

Abstract

Purpose: Adjuvant radiotherapy (RT) has been shown to improve local control (LC) in soft tissue sarcoma (STS) of the extremity. However, concern is frequently raised regarding the tolerance of irradiation over a joint space. The purpose of this study is to evaluate the long-term outcome of patients with STS involving the knee or elbow when treated with limb-sparing surgery (LSS) with or without radiation. Materials and Methods: Review of our prospective data base between 6/82 and 12/99, identified 86 adult (> 16 year old) patients with primary STS arising in the knee (n = 65;76%) or elbow (n = 21;24%) who were treated with LSS. Tumors were deep in 79% of patients, of high-grade (HG) histology in 72%, and > 5 cm in 48%. The margin of resection was microscopically positive in 21% of patients. Arthroscopy was performed as part of an initial outside diagnostic evaluation in 10 (16%) patients with knee STS. Adjuvant RT was given to 48/86 (56%) patients due to poor prognostic features; tumor bulk (54% > 5 cm), deep-seated lesions (85%), and HG histology (77%). The type of RT was brachytherapy (BRT) in 17/48 (35%) patients with a median dose of 45 Gy or external beam radiation (EBRT) in 31/48 (65%) with a median dose of 63 Gy. Complications were assessed in terms of wound complication requiring re-operation, bone fracture, peripheral nerve damage, and joint stiffness. Results: With a median follow-up of 48 (range 4-175) months, the 5-year actuarial rates of local control (LC), distant metastasis-free survival (DMFS) and overall survival (OS) were 75% (95% CI: 64-85%), 82% (95% CI: 73-91) and 81% (95% CI: 71-91%) respectively. The 5-year LC rate for patients who received RT was 78% (95% CI: 65-91%) versus 73% (95% CI: 57-89%) for those who did not (p=0.3). The type of RT used did not influence the 5-year actuarial LC {74% (95% CI: 52-97%) for BRT versus 82% (95% CI: 68-96%) for EBRT (p= 0.7)}. In the subset of patients with STS involving the knee, the 5-year LC was only 46% in those who underwent arthroscopy compared to 79% in those who did not (p=0.4). On multivariate analysis, HG histology (p=0.02; RR: 4, 95% CI: 1-20) and size > 5cm (p=0.04; RR: 3, 95% CI: 1-6) were independent predictors of poor LC. The 5-year actuarial rates for wound re-operation, bone fracture, nerve damage, and joint stiffness were 6%, 4%, 7% (all grade£ 2), and 12% respectively. The influence of RT on 5-year complication rate was as follows: wound re-operation (RT 3% vs. 11% no RT, p=0.07), bone fracture (RT 3% vs. 5% no RT, p=0.7), nerve damage (RT 10% vs. no RT 3%, p=0.3), and joint stiffness (RT 20% vs. no RT 0%, p=0.005). The grade of joint stiffness was minor in 16%, moderate in 2%, and severe in 2%. Conclusion: Adjuvant RT did increase the joint stiffness but it was severe only in 2% of patients. The local control rate was not significantly better in the RT group compared to no RT, but the groups were imbalanced. High-grade and large tumor size emerged as independent prognostic factors for poor local control mandating further improvement in the multimodality therapy in this group of patients. The poorer local control associated with the use of arthroscopy, although not statistically significant, was intriguing but needs further confirmation.