A nomogram to predict metastasis of soft tissue sarcoma of the extremities

Abstract

Soft tissue sarcoma (STS) of the extremities are a rare tumor. Metastases develop in about 40%-50% of patients, most of whom die from their disease. We sought to identify potential risk factors associated with metastatic diseases upon presentation for patients with STS and established a reliable nomogram model to predict distant metastasis of STS at presentation. The current study retrospectively analyzed 3884 STS of the extremities or trunk patients from the Surveillance, Epidemiology, and End Results (SEER) database between 2010 and 2015. Based on patient registration, all patients were randomly allocated to training sets and validation sets (2:1). Then, univariate and binary logistic regression analysis was used to determine the significantly correlated predictors of metastasis. Finally, the nomogram model was established, using these predictors and validated it. 311 (8.21%) of the cases experienced distant metastatic disease was present at the time of presentation. The nomogram was developed from age, histology subtype, primary site, tumor size, grade and depth. Encouragingly, the nomogram showed favorable calibration with C-index 0.790 in the training set and 0.801 in validation set. The DCA showed that the novel model was clinically useful. This nomogram model had a high precision to predict the metastasis of soft tissue sarcoma of the extremities. We expect this model could be used in different clinical consultation and established risk assessment.