Abstract: Objective · To explore the prognostic factors for the breast cancer patients with spinal metastasis, and establish a prognostic scoring model. Methods · A total of 160 breast cancer patients with spinal metastasis in Ruijin Hospital, Shanghai Jiao Tong University School of Medicine from January 2008 to January 2016 were retrospectively identified. The clinical characteristics and prognosis were analyzed by univariate and multivariate survival analysis to explore the prognostic factors. And then a prognostic scoring model was developed according to the regression coefficient for each independent prognostic factor. Results · The 160 breast cancer patients with spinal metastasis whose average age was 56.8 years (range 22-82 years) were identified, and the median follow-up was 40 (24, 55) months. The multivariate Cox analysis showed that the patients' general condition, hormone receptor expression, visceral metastasis, and serum carbohydrate antigen 125 (CA125) level significantly influenced survival (P<0.05). According to the regression coefficients, a survival prediction scoring model comprising these factors was established, which ranged from 0 to 6 points. Three risk groups with different prognoses were identified : low risk group (0-1 point), intermediate risk group (2-4 points), and high risk group (5-6 points). Conclusion · The general condition, hormone receptor expression, visceral metastasis, and serum CA125 level were independent prognostic factors for the breast cancer patients with spinal metastasis. And the prognostic scoring model comprising these four clinical factors can effectively predict the patients' prognoses.

Key words: breast cancer, spinal metastasis, prognostic scoring model, estrogen receptor, progesterone receptor, visceral metastasis, carbohydrate antigen 125 (CA125)