RESUMO
PURPOSE: Estimating the survival time of patients with spinal metastases based on pre-treatment parameters is important for the best choice of therapy. Following two previous studies, this sequel analyzes possible changes in the impact of various parameters and scoring systems and includes a comparison to the previous dataset for the purpose to find the most predictive parameters and scores for this patient group. METHODS: Included were 196 patients retrospectively with confirmed spinal metastases treated between 2005 and 2010 (35% surgery, 65% conservative). Possible prognostic factors [primary tumor, Karnofsky Performance Scale (KPS), visceral metastases, number of bone metastases, pathological fracture and neurologic status] and six scoring systems (Tokuhashi original/revised, Tomita, van der Linden, Bauer original and modified) were analyzed using Kaplan-Meier curves and Cox-regression models. RESULTS: Median overall survival was 7 months with 9% of all patients alive at the time of analysis. Stepwise multivariate analysis showed significant influence on survival for visceral metastases (p<0.0001), primary tumor (p<0.0001), KPS (p<0.0001) and number of spinal metastases (p=0.0271). All scoring systems significantly predicted longer survival at a better score (absolute scores, p<0.001) in this dataset. Significant differentiation between the prognostic groups was seen only for the Tokuhashi original, the Bauer original and modified scores (p<0.001). In comparison to the previous dataset with varying age, gender and primary tumor distribution, the Bauer original and modified scores were the least influenced by the different patient collectives. CONCLUSIONS: The Bauer modified score has shown consistent impact on predicting the remaining survival in patients with spinal metastases and is simultaneously simple in clinical use.