J Neurosurg Spine 33:870–876, 2020
Stereotactic body radiation therapy (SBRT) is utilized to deliver highly conformal, dose-escalated radiation to a target while sparing surrounding normal structures. Spinal SBRT can allow for durable local control and palliation of disease while minimizing the risk of damage to the spinal cord; however, spinal SBRT has been associated with an increased risk of vertebral body fractures. This study sought to compare the fracture rates between SBRT and conventionally fractionated external-beam radiation therapy (EBRT) in patients with metastatic spine tumors.
METHODS Records from patients treated at the University of California, San Francisco, with radiation therapy for metastatic spine tumors were retrospectively reviewed. Vertebral body fracture and local control rates were compared between SBRT and EBRT. Ninety-six and 213 patients were identified in the SBRT and EBRT groups, respectively. Multivariate analysis identified the need to control for primary tumor histology (p = 0.003 for prostate cancer, p = 0.0496 for renal cell carcinoma). The patient-matched EBRT comparison group was created by matching SBRT cases using propensity scores for potential confounders, including the Spinal Instability Neoplastic Score (SINS), the number and location of spine levels treated, sex, age at treatment, duration of follow-up (in months) after treatment, and primary tumor histology. Covariate balance following group matching was confirmed using the Student t-test for unequal variance. Statistical analysis, including propensity score matching and multivariate analysis, was performed using R software and related packages.
RESULTS A total of 90 patients met inclusion criteria, with 45 SBRT and 45 EBRT matched cases. Balance of the covariates, SINS, age, follow-up time, and primary tumor histology after the matching process was confirmed between groups (p = 0.062, p = 0.174, and 0.991, respectively, along with matched tumor histology). The SBRT group had a higher 5-year rate of vertebral body fracture at 22.22% (n = 10) compared with 6.67% (n = 3) in the EBRT group (p = 0.044). Survival analysis was used to adjust for uneven follow-up time and showed a significant difference in fracture rates between the two groups (p = 0.044). SBRT also was associated with a higher rate of local control (86.67% vs 77.78%).
CONCLUSIONS Patients with metastatic cancer undergoing SBRT had higher rates of vertebral body fractures compared with patients undergoing EBRT, and this difference held up after survival analysis. SBRT also had higher rates of initial local control than EBRT but this difference did not hold up after survival analysis, most likely because of a high percentage of radiosensitive tumors in the EBRT cohort.