Abstract

Optimal fractionation of spinal stereotactic radiosurgery (SRS) for spine metastases remains unknown. Retrospective data suggest decreased local failure (LF) with fractionated SRS of brain metastases. We evaluated our institutional outcomes of spinal SRS with the hypothesis that fractionation improves the rate of local failure compared to single-fraction treatment.This IRB-approved, retrospective analysis included patients with spine metastases treated with spinal SRS between October 2002 and November 2014 with evaluable follow-up imaging and no prior irradiation to the given segment. The exposure of interest was single- or multi-fraction SRS with a primary endpoint of the cumulative incidence of LF with death as a competing risk. We assessed bivariate associations between fractionation and single-fraction equivalent dose (SFED in Gy10) as well as high-risk features, defined as epidural extension (Bilsky Scale), paraspinous extension, and gastrointestinal (GI) vs non-GI primary. We calculated the rates of LF and vertebral body compression fracture (VCF) at 1-year, and assessed LF by fractionation when limited only to courses receiving SFED>18 Gy. We analyzed the association between fractionation and LF using subdistribution hazard ratios (SHR) estimated from competing risks regression with death as a competing risk and adjusting for lesion-specific characteristics as well as SFED to determine contribution of these variables to the estimated effect of fraction number on LF. We calculated relative attenuation for the contribution of SFED to this association, defined as [SHRfractions-SHRfractions+SFED] ÷ [SHRSFED-1].In 293 patients with 516 spinal segments, lesions treated with single fraction compared to multi-fraction SRS had less epidural (19% vs 36%, p<0.001) and paraspinous (20% vs 35%, p<0.001) extension, more GI histology (17% vs 10%, p?=?0.039), received a higher mean SFED (18.3 Gy vs. 16.6 Gy, p<0.001), and had a lower 1-year LF (8% vs 14%, p?=?0.02), with no difference in VCF (7% vs. 5%, p?=?0.38). After adjusting for high-risk features, single fraction SRS was associated with lower LF (SHR?=?0.45, 95% CI 0.24-0.84, p?=?0.02). After adjustment for SFED, this association of fractionation was attenuated by 53% and became insignificant (SHR?=?0.78, 95% CI 0.44-1.37, p?=?0.38). Overall, 1-year LF for SFED>18 Gy was 6% compared to 15% for <18 Gy (p<0.001). When limited to courses with SFED>18 Gy (n?=?261), single fraction SRS had no improvement in 1-year LF compared to multi-fraction (6.6% vs 4.6%, p?=?0.77).Single fraction SRS was associated with better local control compared to multi-fraction; however, much of this association was attenuated by SFED but not by high-risk features of treated lesions. To clarify the role of fractionation, we have initiated a prospective, randomized trial of single vs. multi-fraction SRS utilizing the same SFED.

View details for DOI 10.1016/j.ijrobp.2023.06.903

View details for PubMedID 37784661