Patients with metastatic major salivary gland carcinoma (SGCs) always end with a poor prognosis, and survival time is a major concern for clinicians and patients, but effective predictive tools are lacking in clinical practice.

Clinical information on patients diagnosed with metastatic major SGCs was extracted from the SEER database. Cox analysis was applied to identify clinicopathological characteristics associated with patient overall survival (OS). A random survival forest (RSF) algorithm was used to establish an accurate prognostic prediction model for these patients.

Cox analysis revealed that age, T stage, N stage, pathology type, bone and liver metastasis, primary tumor surgery, chemotherapy, and radiotherapy were independent factors for OS among patients with metastatic major SGCs. Our RSF model has a C-index of 0.657 in the test set and 0.701 in the external validation set, and the area under the curve (AUC) values at 1, 3, and 5 years range from 0.715–0.802 in the test set and 0.655–0.918 in the external validation set. Patients were divided into high-risk and low-risk groups based on the risk score of the RSF model, and patients in the low-risk group had significantly better OS than those in the high-risk group, and chemotherapy did not benefit patients in the low-risk group.

In this study, a prognostic prediction model was constructed for patients with metastatic major SGCs using RSF algorithm, and the validation results indicate that the model has the potential to be a useful tool for clinicians in predicting survival and designing individualized treatment.