Abstract
BACKGROUND: The use of nomograms for prognostication of individual cancer patients has been recommended in order to facilitate precision medicine. However, models for patients with soft tissue sarcomas (STSs) are limited because of the rarity and heterogeneity of such cancers. In addition, no model has been developed on the basis of an Asian cohort. Here, we attempted to develop and internally validate nomograms for patients with localized STSs of the trunk and extremity. METHODS: This study retrospectively extracted 2827 patients with primary trunk and extremity STSs after definitive surgery using the Bone and Soft Tissue Tumor Registry, which is a nationwide sarcoma database in Japan. We developed three nomograms predicting the probability of local recurrence-free survival (LRFS), distant metastasis-free survival (DMFS) and disease-specific survival (DSS) at 2 years after surgery, using the Cox multivariate model. The nomograms were internally validated for discrimination and calibration using bootstrap resampling and assessed for their clinical applicability by decision curve analysis (DCA). RESULTS: Local recurrence, distant metastasis and disease-specific death occurred in 241 patients (8.5%), 554 patients (19.6%) and 230 patients (8.1%), respectively. Histological diagnosis, grade and tumor size strongly influenced all three endpoints. The nomograms predicted accurately the probability of LRFS, DMFS and DSS (concordance index: 0.73, 0.70 and 0.75, respectively). DCA demonstrated that our nomograms had clinical applicability. CONCLUSION: We have developed the first nomograms for STSs based on an Asian cohort. These nomograms allowed accurate prediction of LRFS, DMFS and DSS at 2 years after definitive surgery, and can be used as a guide by clinicians for appropriate follow-up and counseling of patients.