Outcomes Specific to Spinal Metastases with Paraspinal Disease Extension Following Spine Stereotactic Body Radiotherapy

Abstract

Spinal metastases with paraspinal disease (PD) extension are known to have worse outcomes following stereotactic body radiotherapy (SBRT). Characteristics of the PD itself have not been investigated to determine the impact on outcomes such as local control, which is the purpose of this study. We retrospectively reviewed those patients who had SBRT for spinal metastases with PD disease, identified from a prospectively maintained database. Spinal metastases previously irradiated or surgical resected were excluded. The extent of PD was classified as involving the rib, neuroforamina, and muscle invasion. The gross tumor volume of PD (GTV_PD) and the clinical target volume of PD (CTV_PD) were segregated from the bony compartments based on the treatment plan contours. The outcomes of interest included the cumulative risk of local failure (LF), re-irradiation rates (ReRT), and overall survival (OS). LF and ReRT were estimated for each treated sites using the competing risk model (death as the competing risk), while OS was evaluated per patient using the Kaplan Meier method. A total of 86 patients with 96 spinal metastases sites with PD were included. Of the 96 treated sites, 65% (62/96), 29% (28/96) and 6% (6/96) of PD spanned 1, 2, and 3 vertebral levels respectively. The median follow-up was 12.4months (range: 0.6-100months). The 6- and 12-month OS for the cohort was 81% and 51%, respectively. 33/86 (38%) patients had radioresistant cancer (gastrointestinal, renal cell carcinoma, thyroid, sarcoma, or melanoma). Involvement of rib, neuroforamina and muscle invasion were observed in 39% (37/96), 65% (62/96) and 21% (20/96) of the treated sites, respectively. Epidural disease was present in 57% (55/96) of treated sites. The median GTV_PD volume was 7cc (range: 0.3-114cc), and the median CTV_PD volume was 24cc (range: 0.4-248cc). The prescribed doses were 24 Gy/2 fractions (fx) (80%), 28 Gy/2 fx (10%) and 30 Gy/4 fx (10%). There were 84 treated sites with at least one post-treatment MRI available for LF assessment. The crude LF risk was 32% (27/84), and the 6- and 12-month cumulative LF rates were 12% and 28%, respectively. There was a trend towards an increased risk of LF when PD involved the rib (35% vs 24% at 1 year respectively, P = 0.07) and muscle (67% vs 20% at 1 year respectively; P = 0.06), but no difference in LF for neuroforamina involvement (26% vs 34% at 1 year respectively, P = 0.5). There were no differences in LF based on cancer radioresistance (P = 0.6), GTV_PD volume (P = 0.3) or CTV_PD volume (P = 0.4). Of the 96 treated sites, 14% (14/96) were re-irradiated (9 with repeat SBRT and 5 with conventional EBRT) at a median of 15 months (range: 4.7-59 months) post initial SBRT. The cumulative incidence of ReRT at 6- and 12-months were 1.2% and 7.3%, respectively. PD involving adjacent rib and muscle may be associated with worse LF following SBRT. Further expansion of the cohort and dosimetric analyses are ongoing.