Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma

Abstract Background To develop a multiclass radiomics model for differentiating chondroid bone tumors using preoperative MRI. Methods This retrospective study included 120 patients (92 enchondromas, 16 low-grade chondrosarcomas, and 12 intermediate-to-high-grade chondrosarcomas) who underwent contra...

Full description

Saved in:
Bibliographic Details
Main Authors: Hyerim Park, Jooyeon Lee, Seungeun Lee, Joon-Yong Jung
Format: Article
Language:English
Published: BMC 2025-05-01
Series:BMC Cancer
Subjects:
Online Access:https://doi.org/10.1186/s12885-025-14330-6
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850127926075326464
author Hyerim Park
Jooyeon Lee
Seungeun Lee
Joon-Yong Jung
author_facet Hyerim Park
Jooyeon Lee
Seungeun Lee
Joon-Yong Jung
author_sort Hyerim Park
collection DOAJ
description Abstract Background To develop a multiclass radiomics model for differentiating chondroid bone tumors using preoperative MRI. Methods This retrospective study included 120 patients (92 enchondromas, 16 low-grade chondrosarcomas, and 12 intermediate-to-high-grade chondrosarcomas) who underwent contrast-enhanced MRI between 2009 and 2019. Tumor segmentation was manually performed by a musculoskeletal radiologist and validated by a senior radiologist. We used least absolute shrinkage and selection operator (LASSO) and random forest (RF) for feature selection and classification, with and without synthetic minority oversampling technique (SMOTE). Model performance was evaluated using five-fold cross-validation with average precision, accuracy, area under the curve (AUC), and weighted kappa statistics. Results The LASSO + RF model based on all sequences achieved the highest accuracy (0.826 ± 0.065) and AUC (0.967 ± 0.027). The highest mAP (0.750 ± 0.095) was observed in the SMOTE-enhanced T2WI-based model, highlighting the potential impact of class imbalance. Quadratic weighted kappa values ranged from 0.648 to 0.731 across models, indicating substantial agreement with pathological results. Conclusions Preoperative MRI-based radiomics provides a robust method for the classification of chondroid bone tumors, potentially enhancing clinical decision-making.
format Article
id doaj-art-e3689fc32d6c4f92b5303eb097fff6a7
institution OA Journals
issn 1471-2407
language English
publishDate 2025-05-01
publisher BMC
record_format Article
series BMC Cancer
spelling doaj-art-e3689fc32d6c4f92b5303eb097fff6a72025-08-20T02:33:31ZengBMCBMC Cancer1471-24072025-05-012511810.1186/s12885-025-14330-6Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcomaHyerim Park0Jooyeon Lee1Seungeun Lee2Joon-Yong Jung3Department of Radiology, College of Medicine, Soonchunhyang University Cheonan Hospital, Soonchunhyang University of KoreaDepartment of Biostatistics and Data Science, UTHealth Houston School of Public HealthDepartment of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of KoreaDepartment of Radiology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of KoreaAbstract Background To develop a multiclass radiomics model for differentiating chondroid bone tumors using preoperative MRI. Methods This retrospective study included 120 patients (92 enchondromas, 16 low-grade chondrosarcomas, and 12 intermediate-to-high-grade chondrosarcomas) who underwent contrast-enhanced MRI between 2009 and 2019. Tumor segmentation was manually performed by a musculoskeletal radiologist and validated by a senior radiologist. We used least absolute shrinkage and selection operator (LASSO) and random forest (RF) for feature selection and classification, with and without synthetic minority oversampling technique (SMOTE). Model performance was evaluated using five-fold cross-validation with average precision, accuracy, area under the curve (AUC), and weighted kappa statistics. Results The LASSO + RF model based on all sequences achieved the highest accuracy (0.826 ± 0.065) and AUC (0.967 ± 0.027). The highest mAP (0.750 ± 0.095) was observed in the SMOTE-enhanced T2WI-based model, highlighting the potential impact of class imbalance. Quadratic weighted kappa values ranged from 0.648 to 0.731 across models, indicating substantial agreement with pathological results. Conclusions Preoperative MRI-based radiomics provides a robust method for the classification of chondroid bone tumors, potentially enhancing clinical decision-making.https://doi.org/10.1186/s12885-025-14330-6RadiomicsChondrosarcomaEnchondromaMagnetic resonance imagingLeast absolute shrinkage and selection operator.
spellingShingle Hyerim Park
Jooyeon Lee
Seungeun Lee
Joon-Yong Jung
Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma
BMC Cancer
Radiomics
Chondrosarcoma
Enchondroma
Magnetic resonance imaging
Least absolute shrinkage and selection operator.
title Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma
title_full Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma
title_fullStr Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma
title_full_unstemmed Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma
title_short Grading chondroid tumors through MRI radiomics: enchondroma, low-grade chondrosarcoma and higher-grade chondrosarcoma
title_sort grading chondroid tumors through mri radiomics enchondroma low grade chondrosarcoma and higher grade chondrosarcoma
topic Radiomics
Chondrosarcoma
Enchondroma
Magnetic resonance imaging
Least absolute shrinkage and selection operator.
url https://doi.org/10.1186/s12885-025-14330-6
work_keys_str_mv AT hyerimpark gradingchondroidtumorsthroughmriradiomicsenchondromalowgradechondrosarcomaandhighergradechondrosarcoma
AT jooyeonlee gradingchondroidtumorsthroughmriradiomicsenchondromalowgradechondrosarcomaandhighergradechondrosarcoma
AT seungeunlee gradingchondroidtumorsthroughmriradiomicsenchondromalowgradechondrosarcomaandhighergradechondrosarcoma
AT joonyongjung gradingchondroidtumorsthroughmriradiomicsenchondromalowgradechondrosarcomaandhighergradechondrosarcoma