The impact of lesion size on outcomes following stereotactic body radiotherapy for spinal metastases.

Abstract

Stereotactic body radiotherapy (SBRT) for metastatic spinal disease provides effective local control (LC). However, there are limited data evaluating how the pretreatment volume or number of affected vertebral levels impacts treatment effect. Therefore, this study compared the efficacy of spine SBRT treatment on LC, distant progression (DP), and overall survival (OS) between patients with limited spinal disease (≤ 2 vertebral segments or < 50 cm3 pretreatment volume) to those with more widespread spine metastases. All patients with metastatic spinal disease treated with SBRT over a 7-year interval at a single institution were retrospectively reviewed. Baseline demographic characteristics, primary tumor histology, anatomical distribution of metastases, and radiation dosing parameters and toxicity were recorded. LC (absence of recurrent disease at an SBRT-treated segment), DP (progressive disease in a new/untreated segment), and OS were measured. Cox proportional hazards models were used to identify patient and lesion characteristics associated with LC, DP, and OS. In total, 375 patients were treated with spine SBRT (465 isocenters). Median minimum and maximum target doses were 9.98 and 17.73 Gy, respectively. Overall, 10.1% of patients developed distant DP, while 11.2% of lesions recurred. New or worsening compression fracture occurred in 5.6% of lesions following SBRT, with 1.9% of radiated lesions requiring surgical management of a postradiation compression fracture. Neither radioresistant pathologies nor dose above or below 14 Gy were associated with differences in LC, DP, or OS. Multivariate regression revealed that OS was significantly worse with baseline Karnofsky Performance Status < 80 (HR 1.33, p = 0.030), lesion volume ≥ 50 cm3 (HR 1.37, p = 0.011), or treatment at ≥ 3 vertebral levels (HR 1.42, p = 0.013). DP was significantly worse with lesion volume ≥ 50 cm3 (HR 2.14, p = 0.0229) but was unaffected by number of vertebral segments treated. Notably, however, LC was not significantly associated with high lesion volume or the number of vertebral segments affected. Lesion size did not impact the recurrence rates following SBRT for spine metastases in this series, with similar excellent local disease control (> 88%) for patients with ≥ 3 vertebral segments involved or ≥ 50 cm3 pretreatment tumor volume compared to patients with less disease burden. Shorter time to DP and worse OS were both associated with high index lesion volume.