Abstract
BACKGROUND: We aimed to develop and validate a comprehensive prognostic model for melanoma bone metastasis. METHODS: Data on melanoma bone metastasis patients were obtained from the SEER database and Fujian Cancer Hospital. Cox regression analysis was conducted to identify independent prognostic factors and to establish a Nomogram to predict the overall survival rate of patients. RESULTS: We generated a Nomogram chart incorporating factors such as Age, Site, AJCC T stage, AJCC N stage, Surg Prim, Systemic and Sur Seq, Surg or Rad Seq, DX Brain, DX Liver, DX Distant LN, Tumor number, First malignant primary, Marital status, and Urban. The 1-year, 3-year, and 5-year OS rate AUCs for the training cohort were 0.715, 0.711, and 0.714, respectively, with a c-index of 0.656; the 1-year, 3-year, and 5-year OS rate AUCs for the internal validation cohort were 0.695, 0.725, and 0.719, respectively, with a c-index of 0.650; the 1-year, 3-year, and 5-year OS rate AUCs for the external validation cohort were 0.714, 0.791, and 0.842, respectively, with a c-index of 0.710. Calibration curves showed the consistency between the Nomogram's observed and predicted prognostic outcomes. CONCLUSION: Our model can be used to predict the prognosis of melanoma bone metastasis.