Radiomics analysis in differentiating osteosarcoma and chondrosarcoma based on T2-weighted imaging and...

Area of Science:

• Radiology
• Oncology
• Medical Imaging Analysis

Background:

• Distinguishing osteosarcoma (OS) from chondrosarcoma (CS) is crucial for appropriate treatment.
• Magnetic resonance imaging (MRI) is a key modality, but differentiation can be challenging.
• Radiomics offers a quantitative approach to extract imaging features for improved diagnostic accuracy.

Purpose of the Study:

• To evaluate the diagnostic value of radiomics models for differentiating OS and CS.
• To compare the performance of models based on fat-suppressed T2-weighted imaging (T2WI-FS) and contrast-enhanced T1-weighted imaging (CET1).
• To develop a robust radiomics model for improved diagnostic accuracy in bone tumors.

Main Methods:

• Retrospective cohort study including pathologically confirmed OS or CS patients.
• Extraction of 530 radiomics features from T2WI-FS and CET1 MRI sequences.
• Application of Least Absolute Shrinkage and Selection Operator (LASSO) for feature selection and multivariate logistic regression for model construction.
• Receiver Operating Characteristic (ROC) curve analysis to assess diagnostic performance.

Main Results:

• The radiomics model incorporating T2WI-FS and CET1 features demonstrated high diagnostic accuracy.
• In the training cohort, Area Under the Curve (AUC) values were 0.894 for CET1 and 0.970 for T2WI-FS.
• In the validation cohort, AUC values were 0.821 for CET1 and 0.899 for T2WI-FS, indicating robust performance.

Conclusions:

• A radiomics model based on T2WI-FS and CET1 MRI sequences effectively differentiates between OS and CS.
• The developed model shows promising diagnostic performance, supporting its clinical utility.
• This approach can assist clinicians in diagnosis and potentially optimize healthcare resource allocation.