Retrospective evaluation of prognostic factors in metastatic spine disease: serum albumin and primary tumour type are key.

Abstract

Treatment decisions for metastatic spine disease are complex and depend on prognosis. Four prognostic systems in use are the Oswestry Risk Index, modified Bauer score (MBS), van der Linden score and New England Spinal Metastasis Score (NESMS). We aimed to determine the performance of these scoring systems in a New Zealand cohort of patients and develop a prognostic score specific to this demographic. A retrospective review of a patient cohort from 2009 to 2016 was undertaken. Scores and individual scoring items were evaluated with univariate and multivariate analyses. Significant items were used to design a simple, population-specific objective scoring system, which was then tested. A total of 106 patients receiving either surgery and radiotherapy (65) or radiotherapy alone (41) were included. Mean post-treatment survival time was 13.7 months. All scoring systems were significantly correlated with survival and had similar concordances. The MBS had the largest coefficient of determination (Cox and Snell's R2 = 0.18), followed by the NESMS (R2 = 0.14). On multivariate analysis, the lung cancer (MBS) and serum albumin (NESMS) items were significant. A modified Oswestry Risk Index primary tumour item and NESMS serum albumin outperformed the MBS (R2 = 0.20), providing the basis for a prognostic scoring tool specific to our demographic. Based on serum albumin and primary tumour type, we propose the 'Metastatic Spine Risk Index' as a simple and objective tool, specific to our population for predicting survival, which can be used in conjunction with other clinical information when considering treatment options.