Dosimetric Analysis of Local Failure Outcomes and Vertebral Compression Fracture Risk in Single-Fraction Spine Stereotactic Radiosurgery for Metastatic Sarcoma

Abstract

Sarcoma spinal metastases (SSM) are particularly difficult to manage given their poor response rates to chemotherapy and their inherent radioresistance. We sought to analyze dosimetric parameters impacting local failure and vertebral compression fracture outcomes in a homogenously treated cohort of patients with SSM treated with single-fraction spine stereotactic radiosurgery (SSRS). A retrospective review was conducted on a cohort of patients with SSM treated with definitive SSRS at a single tertiary institution. 16-24 Gy was delivered to the GTV and 16 Gy uniformly to the CTV. Kaplan-Meier analysis was conducted to assess time to local failure (LF). The log-rank test was utilized to examine group differences. Patients were censored at time of last follow-up or death. Cox proportionate hazards modeling was used to determine hazard ratios (HR) and their respective 95% confidence intervals (CI). A total of 66 patients with 96 lesions underwent SSRS for SSM. Median follow-up was 17 months (IQR 8-28). Median age was 55 years (IQR 41-63). The most common histological subtype was leiomyosarcoma (41%) followed by liposarcoma (9%). 81 lesions received 24 Gy to the GTV, 12 received 18 Gy and 3 received 16 Gy. Median GTV and CTV volume was 13.6cc (IQR 5-27) and 51.6cc (IQR 30-80) respectively. 27% of patients had Bilsky 1b or greater disease. 16 of 96 lesions demonstrated progression representing a crude local failure rate of 17% with median time to failure of 8 months (IQR 5-18). The 1-year actuarial progression free survival (PFS) was 89% with a median PFS of 13 months (IQR 16-63). Median overall survival (OS) was 15 months (IQR 8-28) from SSRS. 8% of patients developed vertebral compression fractures at a median of 13 months post SSRS (IQR 7-25). Every 1 Gy increase in GTV minimum dose (DMin) across the range (5.8-25cc) was associated with a reduced risk of local failure (HR = 0.875 [95% CI 0.787-0.974], p = 0.01). Stratifying thresholds for GTV DMin, a local control benefit was seen as low as 12 Gy and higher (HR = 0.329 [95% CI 0.11-0.97, p = 0.044) with a significantly greater magnitude benefit seen at 14 Gy (HR = 0.267 [95% CI 0.09-0.77, p = 0.014) and above 15 Gy (HR = 0.091 [95% CI 0.03-0.41], p = 0.0018). There were no other queried variables besides GTV Dmin associated with local control including: GTV: volume, mean, Dmax, D90, CTV: volume, Dmin, Dmean, Dmax, or D90. There was an increased risk of VCF with increasing CTV DMean (HR = 2.4 [95% CI 1.4-4.1], p = 0.002) and CTV D90 (HR = 2.2 [95% CI 1.2-4.0], p = 0.01); however, no association with GTV parameters. This study represents one of the most homogenously treated and the largest cohorts of patients with sarcoma spinal metastases treated with single-fraction SSRS. Despite inherent radioresistance, SSRS confers durable and high rates of local control in SSM without unexpected long-term toxicity rates. Increasing GTV minimum dose is significantly associated with superior local control with no corresponding increased risk of VCF.