Stratifying Response to Stereotactic Radiosurgery for Spinal Metastases by Primary Site Genomic Mutations

Abstract

Spine stereotactic radiosurgery (SSRS) is effective in achieving durable local control (LC) and palliation of pain in patients with spinal metastases. Per institutional standards, SSRS prescription dose is tailored by primary site histology, with radioresistant disease receiving escalated doses. While the association between tumor histology and radioresistance is well studied, the association between specific genotypic mutations and radioresistance to SSRS is not well known. We sought to determine if a relationship exists between primary tumor mutations and clinical outcomes following SSRS. We performed a retrospective analysis of 201 patients with available primary site mutation profiles who underwent SSRS to 327 spinal metastases from 2007-2022 at a single institution. Associations with overall survival (OS) and LC were identified using univariate and multivariable Cox proportional hazards modeling adjusted for clinicopathologic and treatment-related factors. The median age was 59 years (range: 11-85) at the time of SSRS. One hundred and five patients were male (52%). The median SSRS dose to the gross tumor volume was 24 Gy (range: 12-50) given in a median of 1 fraction (range: 1-5), corresponding to a biologically effective dose (BED10) of 81.6 Gy. The most common primary tumor sites were head and neck (20%), lung (19%), and genitourinary (15%). The most common spine subsites were thoracic (57%) and lumbar (27%). Twenty-four mutated genes were identified in primary tumors with the most common being TP53 (28%), KIT (15%), PIK3CA (15%), and KRAS (10%). Prior to receipt of SSRS, 19% of patients underwent surgical resection and 23% received conventional radiation therapy (RT) to their treated sites. The median follow-up time and OS following SSRS was 97 months (95% confidence interval [CI] 86-128) and 41 months (95% CI 20-31), respectively. The median LC at 1 and 2 years following SSRS was 88% (95% CI 84-92) and 75% (95% CI 70-82), respectively. On multivariable analysis, receipt of prior RT (hazard ratio [HR] 1.86; 95% CI 1.29-2.67; P = 0.0008) and TP53 mutation (HR 1.68; 95% CI 1.20-2.35; P = 0.0024) were associated with shorter OS, whereas STK11 mutation (HR 2.14; 95% CI 0.47-4.70; P = 0.0589) trended towards shorter OS. LC was more durable with increasing BED10 (HR 0.98; 95% CI 0.96-1.00; P = 0.0166) and less durable with mutations in ATM (HR 2.40; 95% CI 1.06-5.44; P = 0.0362), STK11 (HR 4.89; 95% CI 1.40-15.01; P = 0.0119), and CTNNB1 (HR 4.49; 95% CI 1.90-10.62; P = 0.0006). While favorable outcomes among all patients receiving SSRS were seen, increasing BED10 was associated with more durable LC. Primary site ATM, STK11, and CTNNB1 mutations may be associated with poorer LC following SSRS. Further investigation into the prognostic value of individual mutations and combinations thereof is warranted.