AUTHOR=Li Wenle , Wang Gui , Wu Rilige , Dong Shengtao , Wang Haosheng , Xu Chan , Wang Bing , Li Wanying , Hu Zhaohui , Chen Qi , Yin Chengliang TITLE=Dynamic Predictive Models With Visualized Machine Learning for Assessing Chondrosarcoma Overall Survival JOURNAL=Frontiers in Oncology VOLUME=Volume 12 - 2022 YEAR=2022 URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.880305 DOI=10.3389/fonc.2022.880305 ISSN=2234-943X ABSTRACT=Chondrosarcoma is a malignant bone tumor with low incidence rate. Accurate risk evaluation is crucial for chondrosarcoma treatment. Due to the limited reliability of the existing predictive models, we intended to develop a credible predictor for clinical chondrosarcoma based on SEER data and four Chinese medical institutes. Three algorithms (BSR, UCOX and LASSO) were used for training. A normogram predictor including eight variables, Age, Sex, Grade, T, N, M, surgery and Chemotherapy, is constructed. The predictor provides good performance in discrimination and calibration, with AUC ≥ 0.8 in ROC curves of both internal and external validation. Specially, the predictor had very good clinical utility in terms of net benefit to patients at the 3 and 5 year points in both North America and China. A convenient web calculator based on the prediction model is available at https://drwenle029.shinyapps.io/CHSSapp, which is free and open to all clinicians.