Evaluation of scoring systems and prognostic factors in patients with spinal metastases from nasopharyngeal carcinoma.

BACKGROUND CONTEXT The decision for operative treatment of patients with spinal metastases is dependent on the patient's predicted survival. Tokuhashi, Tomita, Bauer, and Oswestry scores have been devised for survival prediction; however, none of these systems have been evaluated in nasopharyngeal carcinoma (NPC). PURPOSE To investigate the accuracy of these scoring systems in predicting survival and to identify prognostic factors for survival of the patients with spinal metastases from NPC. STUDY DESIGN Retrospective analysis of the patients with spinal metastases from NPC who were treated in our institution. PATIENT SAMPLE The study included 87 patients with spinal metastases from NPC. OUTCOME MEASURES The primary outcome measure was the survival time of these patients. The potential prognostic factors that are known to influence survival such as general condition, extraspinal bone metastases, vertebral bone metastases, visceral metastases, and neurologic assessment based on Frankel score were also studied. METHODS The predicted survival according to the four scoring systems were calculated and labeled as "A" scores. These patients were then rescored by assigning NPC as a good prognostic tumor and labeled as "B" scores. The predicted survival of scores A and B were compared with actual survival. Potential prognostic factors of survival were investigated using univariate and multivariate Cox regression analyses. For all scoring systems, Kaplan-Meier survival estimates and log-rank tests were done; the predictive values were calculated using postestimation after Cox regression analyses. RESULTS The median overall survival for the whole cohort was 13 (range 1-120) months. In multivariate analysis, general condition (p<.01), visceral metastases (p<.01), and vertebral metastases (p<.01) showed significant association with survival. The absolute score of all scoring systems was significantly associated with actual survival, which extended to the different prognostic subgroups of each scoring systems. Log-rank test revealed significant differences in survival between the different prognostic subgroups of all scoring systems (p<.01). Predictive value of survival by modified Tokuhashi score was the highest among all four scoring systems. CONCLUSIONS Patients with spinal metastases from NPC have relatively good survival prognosis. All four scoring systems could be used to prognosticate these patients. The modified Tokuhashi score is the best in doing so.