Study design: We conducted a prospective cohort study of 448 patients with a variety of spinal metastases. Objective. To compare the predictive value of the Tokuhashi scoring system (T12) and its revised edition (T15) for life expectancy both in the entire study group as well as in the various primary tumor subgroups. Summary of background data. In 1990 Tokuhashi and coworkers formulated a one point-addition-type prognostic scoring system with a total sum of 12 points for preoperative prediction of life expectancy as an adjunct in selecting appropriate treatment. Because the site of the primary tumor influences ultimate survival, the scoring system was revised in 2005 to a total sum of 15 points based on the origin of the primary tumor. Methods. This study included 448 patients with vertebral metastases, all of whom underwent surgical treatment during November 1992 to November 2009 at the Aarhus University Hospital. Data was retrieved from the Aarhus Spinal Metastases Algorithm. Scores based on the T12- and T15 scoring systems were calculated prospectively for each patient. All the patients were divided into three survival groups with different life expectancies according to their score points in both scoring systems. Furthermore, we divided all the patients into different groups dictated by the site of their primary tumor. The predictive value of each scoring system, both in the entire group and in each subgroup determined by the original tumour site, was evaluated by the Log-rank-test. The McNemar's test was used to compare the differences in accuracy rates between these two scoring systems. Survival curves were estimated using the Kaplan-Meier methods. Probability value less than 0.05 was considered statistically significant. Result. For the 448 patients with vertebral metastases, both the T12 and T15 scoring systems showed statistically significant predictive value (T12 group p<0.0001, T15 group p<0.0001, Log-rank test). The correlation between predicted survival period and real survival was significantly higher in T15 (p<0.0001, McNemar's test) than in T12. The further analyses by type of metastases showed that the predictive value of T12 and T15 was found in the prostate (p=0.0003), breast (p=0.0385), other tumor group (p<0.001) and lymphoma (p<0.05). Only T12 displayed predictive value in the colon group (p=0.0011). The accuracy rate of prognosis in T15 was significantly higher in those groups with spinal metastases originating from prostate (p=0.0032), breast (p<0.0001) and lung (p=0.0076). Conclusion. Both the T12 and T15 scoring systems have significant preoperative predictive value in terms of predicting the survival period for the entire vertebral metastases patient group as well as in the prostate, breast, lymphoma, and colon tumor groups. The accuracy rate was significantly improved in T15 for the total study group and in the subgroups of prostate, breast, and lung tumor