Predictors of survival in patients with prostate cancer and spinal metastasis

Abstract

ObjectProstate cancer is the second most common malignancy to cause death in men, with metastases to the spine being the most common site of metastatic burden. A retrospective observational study was performed to determine survival of patients in whom spinal metastasis from prostate cancer had been diagnosed.MethodsThe patient population was obtained from the Prostate Clinical Research Information System (CRIS) at the Dana-Farber Cancer Institute. Patients were observed over a period of 19 years, between June 1990 and April 2009. Clinical covariates were studied in their relationship to overall survival, the primary outcome, by using the Kaplan-Meier method and Cox regression.ResultsOf a total of 9010 patients in the Prostate CRIS database, 333 were identified as having developed spinal metastases. The median overall survival after diagnosis of spinal metastasis was 24 months (95% CI 21–28 months). The estimated 1-year overall survival was 73% (95% CI 67%–77%). In 85% of patients, at least 1 additional site of metastasis was documented. Among 28 patients who had no additional sites of metastases, the median survival was 55.9 months, whereas an increasing burden of disease was associated with shorter survival (p = 0.0001). The association was observed regardless of whether the metastatic burden was characterized as the presence of additional (nonspinal) bone metastasis, the presence of additional nonbone metastasis, or as the number of concomitant metastatic sites (all p = 0.0001). In multivariate analysis, a higher prostate-specific antigen level at the diagnosis of spinal metastasis, a longer duration between the diagnosis of prostate cancer and spinal metastasis, and the presence of additional metastasis at the time of diagnosis of spinal metastasis (all p = 0.0001) were independently associated with a shorter overall survival.ConclusionsThe results of this study are important for oncologists, neurosurgeons, and primary care physicians who have patients with prostate cancer that metastasizes to the spine, because these results can be used to form a prognosis and guide the physician in making appropriate decisions regarding the patient's treatment. Future work should include building a predictive model that accurately determines survival in patients with metastatic disease, because this would guide the physician in devising the most appropriate treatment plan for each patient.