Metastatic disease of the spine is an increasingly common public health problem. Surgery should be an integral component of the overall cancer treatment plan and, importantly, must neither delay not jeopardize any of the other components. The prognosis governs the choice of the surgical strategy. Tokuhashi et al. developed a prognostic score in 1990, then revised it in 2000 and 2005. Here, our objective was to evaluate the performance of the Tokuhashi score in a cohort of 260 patients and to look for other variables that might improve preoperative outcome prediction.

We retrospectively established a single-centre cohort of 260 patients who underwent spinal metastasis surgery between 1998 and 2008. For each patient, the following data were collected prospectively: socio-demographic features, history of the malignancy, variables needed to determine the Tokuhashi score, and treatments used. SAS 9.0 software was chosen for the statistical analysis. Variables were described as mean±SD, overall survival was estimated using the Kaplan–Meier method, and survivals in subgroups were compared by the log-rank test. To assess agreement between survival predicted by the Tokuhashi score and observed survival, we computed Cohen's kappa and interpreted the results according to Landis and Koch.

There were 143 females and 117 males with a mean age of 59years and overall median survival of 10months. Median observed survivals in the three Tokuhashi score categories (<6, 6–12, and >12months predicted survival) were 5, 10, and 36months, respectively. These survival times differed significantly (P<0.0001). Cohen's kappa indicated moderate agreement between predicted and observed survivals. Other factors associated with significant survival differences were time from cancer diagnosis to metastasis diagnosis (synchronous, <2years, 2–5years, or >5years; P<0.0001) and age (<70years or ≥70years, P=0.0053).

Our cohort study supports the validity and reproducibility of the Tokuhashi score. Our finding that shorter time to metastasis diagnosis and age≥70years were also significantly associated with survival in our population invites further efforts to improve and update the Tokuhashi score.

IV, retrospective study.