Assessment of survival prediction after surgery in spinal metastases patients using the Global Spine Study Tumor Group (GSTSG) risk calculator; an external validation from a tertiary cancer hospital.

ABSTRACT

PURPOSE: We aim to validate the Global Spine Tumor Study Group (GSTSG) score compared to previous prognostic scoring systems in spinal metastasis. METHODS: We conducted a retrospective study from January 2013 to December 2022. The survival prediction was compared between the GSTSG, Tomita Score, Revised Tokuhashi Score, and Skeletal Oncology Research Group (SORG) Nomogram. Single-variable factors associated with survival rate were analyzed using univariate Cox regression and multivariable Cox proportional hazard model. Receiver operating characteristic was used for external validity analysis at 3, 6, 12, and 24 months. The overall survival rate was reported using the Kaplan-Meier survival curve. RESULTS: 248 spinal metastasis patients were included. The mean age was 59.23 ± 12.55 years. The mean duration of follow-up time was 470.29 ± 441.98 days. The external validity of GSTSG was the highest at all follow-up times (sufficiently accurate AUC > 0.7), which was about the same as SORG at 3 months (both AUC of GSTSG and SORG = 0.76) and higher than modified Tokuhashi and Tomita score at 12 months (AUC of GSTSG = 0.78, SORG = 0.71, Tomita = 0.64, and modified Tokuhashi = 0.61, respectively). CONCLUSION: From our study, the Multivariate Cox regression analysis indicates that the significant factors related to survival rate are regular analgesic use of weak opioids, lung metastasis, and previous chemotherapy. Compared to other traditional spinal metastases prognostic scoring systems, GSTSG shows the highest AUC for external validity in all follow-up times up to 24 months.