Validation of a model with which to predict the survival prognosis of patients with spinal cord compression resulted from metastatic cancers

Abstract

PurposeThis study was to create and validate a new therapeutic strategy which can guild clinicians to select the optimal therapy for patients with spinal cord compression resulted from metastatic cancers. MethodsThe entire cohort of 206 consecutive patients was randomly divided into two groups: a training group and a validation group. For the patients randomized to the training group, we retrospectively analyzed 12 preoperative factors. The receiver operating characteristic (ROC) curve c-statistics were calculated to measure the capability of the score and the revised Tokuhashi score in the validation group. ResultsFour prognostic factors, primary site (P < 0.01), preoperative ambulatory status (P = 0.02), visceral metastases (P < 0.01), and preoperative chemotherapy (P = 0.04), were included in the scoring model. The prognostic scores ranged between 0 and 9 points, and three prognostic groups were designed. The median survival were 3.4 months for patients with 0–2 points, 7.2 months for those with 3–5 points, and 18.3 months for those with 6–9 points, respectively (P < 0.01). In the validation group, the corresponding median survival was 3.8 months, 7.1 months, and 16.3 months, respectively (P < 0.01). The ROC curve c-statistics for the scores as a predictor of 3, 6, and 12 months survival rates were 0.74, 0.78, and 0.83 respectively, and the c-statistics for the revised Tokuhashi scores were 0.73, 0.75, and 0.75, respectively. ConclusionsWe created and validated a new scoring model for predicting survival and function outcome of patients with spinal cord compression resulted from metastatic cancers after surgical decompression. This scoring model can help select the optimal therapy for those patients.