An Interpretable Integer Risk Score for Osteosarcoma Survival: Development and External Validation in a Multicenter Cohort
Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Background Osteosarcoma prognosis is heterogeneous, and existing risk stratification tools are limited. We aimed to develop and externally validate an interpretable integer‑based scoring system for predicting overall survival in patients with osteosarcoma. Methods This retrospective prognostic study used data from the Surveillance, Epidemiology, and End Results (SEER) database for model development (2010 to 2022) and an independent multicenter Chinese cohort for external validation (2010 to 2024). The SEER cohort was randomly split into training (70%) and internal validation (30%) sets. Candidate predictors included demographic, clinical, and treatment‑related variables available at diagnosis. The AutoScore-Survival framework was applied to derive an integer score. Model performance was assessed using C‑index, time‑dependent AUC, calibration metrics, and decision curve analysis. Results The development cohort included 2,755 patients (906 deaths, 32.9%) and the external validation cohort 816 patients (314 deaths, 38.5%). The final score (0 to 19 points) included seven variables: lung metastasis, tumor grade, age, tumor size, primary site, lymph node metastasis, and chemotherapy. In external validation, the score achieved a C‑index of 0.705 (95% CI, 0.675–0.735) and time‑dependent AUCs of 0.729, and 0.742 at 3, and 5 years. Calibration was good at 3 years (O/E ratio = 1.156), with some underestimation at 5 years (O/E = 1.368). Decision curve analysis showed positive net benefit at 3 and 5 years. Conclusions This integer risk score provides a practical and interpretable tool for osteosarcoma risk stratification with acceptable discrimination and calibration in external validation. Future prospective multi-center studies are needed to confirm its clinical utility across broader populations and to explore integration of molecular markers for improved model performance.
Source: Research Square
Language: English
Subfield: Oncology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7223788913