Impact of Dosimetric Factors on Local Failure in Patients with Spine Metastasis after Stereotactic Body Radiotherapy: A Multi-Institutional Study

Abstract

<h3>Purpose/Objective(s)</h3> Stereotactic body radiotherapy (SBRT) is increasingly utilized in the treatment of spine metastasis. Preliminary data suggest that higher doses may improve local control, but may also increase risk of toxicity including vertebral compression fracture (VCF). Our single institution study recently found that a minimum dose of at least 21 Gy to 80% of planning target volume (PTV D80%) in 2 fractions was associated with higher risk of VCF. The purpose of this study is to assess this variable's relationship with local failure (LF) and VCF in an international, multi-institutional cohort. <h3>Materials/Methods</h3> Patients with radiation naïve solid tumor spine metastases treated with SBRT at 5 international institutions from 2010 - 2020 were included. Patients with surgical stabilization were excluded. LF was defined per spine response assessment in neuro-oncology criteria. Variables examined included spinal instability neoplastic score (SINS) factors and dosimetric/volumetric characteristics of radiation planning. All fractionation scheme doses were converted to 2-fraction equivalent doses (2fxED) using the linear quadratic model with an alpha/beta of 3 and used in analyses of PTV D80%. Univariate and multivariate Cox proportional hazard models for LF and VCF were constructed using the Fine and Gray competing risk method. <h3>Results</h3> 357 vertebral segments from 234 patients were treated with SBRT. Common primary tumor types were prostate (n = 50, 22%), non-small cell lung cancer (n = 42, 18%), breast (n = 35, 15%), and kidney (n = 25, 11%). Most (n = 199 pts, 242 segments) were treated with 2 or 3 fractions with a median prescription dose of 24 Gy or 27 Gy, respectively (range: 14-45 Gy in 1-5 fractions). Median follow was 21.1 months (0.6-88.4 mo). LF was 12.6% and 18.1% at 1- and 2-years, respectively. Variables associated with increased LF on univariate analysis were lower PTV D80% at 2fxED ≤21 Gy (HR 0.36, 95% CI 0.20-0.65, p = 0.001) and lower prescription dose BED3 (HR 0.98 as continuous variable, CI 0.97-0.99, p = 0.03). On multivariate analysis, only PTV D80% at 2fxED ≤21 Gy remained associated with increased risk of LF (HR 0.46, CI 0.24-0.85, p = 0.01). VCF incidence was low, at 4.2% and 6.7% at 1- and 2-years, respectively. Higher SINS was predictive of VCF (HR 1.22, CI 1.1-1.36, p <0.001). Patients with higher PTVD80% at 2fxED>21 Gy had a higher, albeit non-significant, risk of VCF (HR 3.30, CI 0.80-13.6, p = 0.10), consistent with our previous single institution study. <h3>Conclusion</h3> This multi-institutional, international study suggests that a PTV D80% at 2fxED ≤21 Gy may increase the risk of LF, while >21 Gy may increase the likelihood of VCF. Prospective studies are needed to further explore this complex relationship and identify the dose/fractionation schedule that best balances local control with normal tissue toxicity including VCF.