Development of the interpretable typing prediction model for osteosarcoma and chondrosarcoma based on machine learning and radiomics - a multicenter retrospective study

Aug 6, 2026

Accurate differentiation between osteosarcoma and chondrosarcoma is essential for treatment selection and prognosis, yet conventional radiological assessment is hindered by overlapping imaging patterns. A multicenter retrospective study trains and validates six machine learning classifiers using CT-derived radiomics features and patient pathology labels, and quantifies model interpretability with SHAP. The RF model achieves near-perfect discrimination (AUC ~1.00), supported by key wavelet-transformed GLCM and first-order predictors.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Introduction

Background and clinical challenge

Limitations of biopsy and role of CT imaging

Methods

Retrospective cohort and data sources

Radiomics feature extraction and models

Model validation and SHAP interpretability analysis

Results

Discrimination performance of machine learning models

Feature importance and interpretability findings

Conclusion

OPEN ACCESS EDITED BY Veronica Aran, Instituto Estadual do Cérebro Paulo Niemeyer (IECPN), Brazil REVIEWED BY Weihang Li, Fourth Military Medical University, China Run Meng, Nantong University, China *CORRESPONDENCE Xiao-Bin Tian [txb6@vip.163.com](txb6@vip.163.com) RECEIVED 16 September 2024 ACCEPTED 30 October 2024 PUBLISHED 20 November 2024 CITATION Long Q-Y, Wang F-Y, Hu Y, Gao B, Zhang C, Ban B-H and Tian X-B (2024) Development of the interpretable typing prediction model for osteosarcoma and chondrosarcoma based on machine learning and radiomics: amulticenter retrospective study. Front. Med. 11:1497309. doi: 10.3389/fmed.2024.1497309 COPYRIGHT © 2024 Long, Wang, Hu, Gao, Zhang, Ban and Tian. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. TYPE Original Research PUBLISHED 20 November 2024 DOI 10.3389/fmed.2024.1497309 Development of the interpretable typing prediction model for osteosarcoma and chondrosarcoma based on machine learning and radiomics: a multicenter retrospective study Qing-Yuan Long 1,2, Feng-Yan Wang 2, Yue Hu3, Bo Gao 2, Chuan Zhang 1, Bo-Heng Ban4 and Xiao-Bin Tian 2* 1The Second Affiliated Hospital of Guizhou Medical University, Kaili, China, 2School of Clinical Medicine, Guizhou Medical University, Guiyang, China, 3Guang'anmen Hospital, China Academy of Chinese Medical Sciences, Beijing, China, 4Qiannan State Hospital of Traditional Chinese Medicine, Duyun, China Background: Osteosarcoma and chondrosarcoma are common malignant bone tumors, and accurate differentiation between these two tumors is crucial for treatment strategies and prognosis assessment. However, traditional radiological methods face diagnostic challenges due to the similarity in imaging between the two. Methods: Clinical CT images and pathological data of 76 patients confirmed by pathology from January 2018 to January 2024 were retrospectively collected from Guizhou Medical University Affiliated Hospital and Guizhou Medical University Second Affiliated Hospital. A total of 788 radiomic features, including shape, texture, and first-order statistics, were extracted in this study. Six machine learning models, including Random Forest (RF), Extra Trees (ET), AdaBoost, Gradient Boosting Tree (GB), Linear Discriminant Analysis (LDA), and XGBoost (XGB), were trained and validated. Additionally, the importance of features and the interpretability of the models were evaluated through SHAP value analysis. Results: The RF model performed best in distinguishing between these two tumor types, with an AUC value close to perfect at 1.00. The ET and AdaBoost models also demonstrated high performance, with AUC values of 0.98 and 0.93, respectively. SHAP value analysis revealed significant influences of wavelettransformed GLCM and First Order features on model predictions, further enhancing diagnostic interpretability. Conclusion: This study confirms the effectiveness of combining machine learning with radiomic features in improving the accuracy and interpretability of osteosarcoma and chondrosarcoma diagnosis. The excellent performance of the RF model is particularly suitable for complex imaging data processing, providing valuable insights for the future. KEYWORDS osteosarcoma, chondrosarcoma, machine learning, typing prediction, interpretability Frontiers in Medicine 01 [frontiersin.org](frontiersin.org) 1 Introduction Osteosarcoma, a malignant tumor originating from mesenchymal stem cells, has an annual incidence rate ranging from 1 to 4 cases per million people and can occur at any age, wit

FAQs

Why is differentiating osteosarcoma and chondrosarcoma clinically important?

Accurate typing guides treatment strategy and prognosis assessment. Similar imaging patterns make differentiation difficult, and misdiagnosis may delay optimal care or lead to unnecessary treatments.

What data and features were used to train the prediction models?

The study retrospectively collected CT images and pathology-confirmed data from 76 patients, extracted 788 radiomic features covering shape, texture, and first-order statistics, and trained six machine learning models.

How was model interpretability evaluated?

SHAP value analysis was used to identify which features most influenced model predictions, highlighting contributions from wavelet-transformed GLCM and first-order features.

Which model performed best and what was its key result?

The Random Forest model performed best, reaching an AUC close to 1.00 for distinguishing between osteosarcoma and chondrosarcoma.

Aug 6, 2026

Upload by Ava Thompson