Outcomes for Spine Stereotactic Body Radiation Therapy and an Analysis of Predictors of Local Recurrence

Abstract

To investigate local control, survival outcomes, and predictors of local relapse for patients treated with spine stereotactic body radiation therapy. We reviewed the records of 332 spinal metastases consecutively treated with stereotactic body radiation therapy between 2002 and 2012. The median follow-up for all living patients was 33months (range, 0-111months). Endpoints were overall survival and local control (LC); recurrences were classified as either in-field or marginal. The 1-year actuarial LC and overall survival rates were 88% and 64%, respectively. Patients with local relapses had poorer dosimetric coverage of the gross tumor volume (GTV) compared with patients without recurrence (minimum dose [Dmin] biologically equivalent dose [BED] 23.9 vs 35.1Gy, P<.001; D98 BED 41.8 vs 48.1Gy, P=.001; D95 BED 47.2 vs 50.5Gy, P=.004). Furthermore, patients with marginal recurrences had poorer prescription coverage of the GTV (86% vs 93%, P=.01) compared with those with in-field recurrences, potentially because of more upfront spinal canal disease (78% vs 24%, P=.001). Using a Cox regression univariate analysis, patients with a GTV BED Dmin ≥33.4Gy (median dose) (equivalent to 14Gy in 1 fraction) had a significantly higher 1-year LC rate (94% vs 80%, P=.001) compared with patients with a lower GTV BED Dmin; this factor was the only significant variable on multivariate Cox analysis associated with LC (P=.001, hazard ratio 0.29, 95% confidence interval 0.14-0.60) and also was the only variable significant in a separate competing risk multivariate model (P=.001, hazard ratio 0.30, 95% confidence interval 0.15-0.62). Stereotactic body radiation therapy offers durable control for spinal metastases, but there is a subset of patients that recur locally. Patients with local relapse had significantly poorer tumor coverage, which was likely attributable to treatment planning directives that prioritized the spinal cord constraints over tumor coverage. When possible, we recommend maintaining a GTV Dmin above 14Gy in 1 fraction and 21Gy in 3 fractions.