Analysis of seven prognostic scores in patients with surgically treated epidural metastatic spine disease.

Background: Prognostic scores have been proposed to guide the treatment of patients with metastatic spine disease (MSD), but their accuracy and usefulness are controversial. The aim of this study was to evaluate seven such prognostic scoring systems. The following prognostic scores were compared: Tomita, Van der Linden (VDL), Bauer modified (BM), Oswestry Spinal Risk Index (OSRI), Tokuhashi original (T90), Tokuhashi revised (TR05), and modified Tokuhashi revised (TR17).

Methods: We retrospectively reviewed all our patients who underwent surgery for spinal metastases, February 2008-January 2015. We classified all 223 patients into the predicted survival-time categories of each of the 7 scoring systems and then tallied how often this was correct vis-à-vis the actual survival time. Accuracy was also assessed using receiver operating characteristic (ROC) analysis at 1, 3, and 12 months.

Results: The median (95% CI) survival of the 223 patients was 13.6 (7.9-19.3) months. A groupwise ROC analysis showed sufficient accuracy for 3-month survival only for TR17 (area under the curve [AUC] 0.71) and for 1-year survival for T90 (AUC 0.73), TR05 (AUC 0.76), TR17 (AUC 0.76), Tomita (AUC 0.77), and OSRI (AUC 0.71). A pointwise ROC score analysis showed poor prognostic ability for short-term survival (1 and 3 months) with sufficient accuracy for T90 (AUC 0.71), TR05 (AUC 0.71), TR17 (AUC 0.71), and the Tomita score (AUC 0.77) for 1-year survival.

Conclusion: The TR17 was the only prognostic system with acceptable performance here. More sophisticated assessment tools are required to keep up with present and future changes in tumor diagnostics and treatment.