What Is the 5-year Survival of Patients With or Without Surgery in Spinal Metastatic Disease?

Background

Despite advances in imaging, nonoperative management, immunotherapy, and surgery over the last 15 years, sizable improvements in survival for patients with spinal metastases have not been observed. Older studies may no longer be accurate in light of rapid technological improvements. The performance of scoring utilities, such as the Skeletal Oncology Research Group (SORG) nomogram and the New England Spinal Metastasis Score (NESMS), has also not been investigated beyond the first year of treatment. QUESTIONS/PURPOSES (1) What was the 5-year survival of patients who presented with spinal metastases and were treated with or without surgery according to contemporary indications, stratified by tumor type? (2) How effective were the SORG and NESMS at prognosticating survival up to 5 years after presentation?

Methods

We used our healthcare system's registry to identify patients treated operatively and nonoperatively for spinal metastatic disease (full surveillance inclusive of follow-up from January 1, 2017 to December 31, 2025). Data included in this study were drawn from two large, tertiary academic hospitals in a single major city, as well as from two local affiliate community hospitals. This study environment allowed for capturing a broad sample of tumor and treatment types, which is representative of the complex care associated with spinal metastatic disease. Query of the registry identified 997 patients; 71 patients were excluded because their initial presentation was < 5 years from the final data date; they were therefore ineligible for 5-year evaluation. Those whose initial presentation was < 5 years from study initiation, but had died, were maintained. One percent of patients (13 of 926) were eligible for 5-year follow-up but had an unknown vital status. Missing data were uncommon, with no variable having > 1% missingness. Sixty percent (551) of patients were treated with surgery. Overall, the surgical and nonoperative cohorts were comparable regarding demographic and clinical characteristics. Patients treated with surgery were slightly older than those treated nonoperatively (62 years versus 59 years; p < 0.01). The nonoperative cohort had a female predominance, whereas the surgical cohort did not (56% versus 43%; p < 0.001). Both cohorts had a median CCI score of 8 (p = 0.7). There were no differences in the use of immunotherapy between those treated with surgery (29%) and those treated with nonoperative care (27%; p = 0.6). We evaluated survival over the course of 5 years after presentation. In this registry, death is captured through a combination of medical record documentation as well as linkage to external state and federal sources. We performed multivariable Cox proportional hazards regression analysis for 5-year survival using each spinal metastasis prognostic score. The SORG is a prognostic score from 0 to 100 that represents the calculated probability of survival. NESMS ranges from 0 to 3, with higher scores representing a more favorable clinical state. Separate models were generated using SORG and NESMS as the primary explanatory variable, respectively. Both models adjusted for age, sex, race, comorbidity burden, immunotherapy, and surgical treatment.

Results

At the 5-year time point, the Kaplan-Meier survival estimate was 8.9% (95% confidence interval [CI] 7.2% to 10.9%). Patients with lung cancer demonstrated the lowest survival rate among the discrete tumor types at every time point. When accounting for age, sex, race, comorbidities, immunotherapy, and treatment type, the SORG score was associated with survival (HR 0.98 [95% CI 0.98 to 0.99]; p < 0.001), meaning that each single-point increase in SORG score was associated with a 2% decreased likelihood of mortality. When controlling for the same variables, the NESMS was also associated with 5-year survival (HR 0.61 [95% CI 0.56 to 0.66]; p < 0.001), meaning that each 1-point increase in SORG score was associated with 40% reduction in mortality.

Conclusion

We present long-term survival for patients with spinal metastases in the modern period of spine oncology care, stratified by primary cancer type. Although mortality rates over time remain high, our study is the first to show that approximately 10% of patients will survive 5 years or longer. These data can aid in patient counseling as well as surgical decision-making. Understanding the 5-year survival forecast for patients with spinal metastases will help physicians tailor treatment decisions and can help surgeons prioritize minimizing surgical morbidity with shorter, simpler operations, when possible. Both the SORG and NESMS can be used to support long-term estimations regarding survival. LEVEL OF EVIDENCE Level III, therapeutic study.