Related Experiment Video
Updated: May 27, 2025
Using Micro-computed Tomography for the Assessment of Tumor Development and Follow-up of Response to Treatment in a Mouse Model of Lung Cancer
Published on: May 20, 2016
Computed Tomography Radiomics-based Combined Model for Predicting Thymoma Risk Subgroups: A Multicenter Retrospective
Yifei Liu1, Chao Luo1, Yongshun Wu2
1Department of Radiology, Sun Yat-sen University Cancer Center, State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, 651 Dongfeng Road East, Guangzhou 510060, Guangdong, China (Y.L., C.L., S.Z., G.R., H.L., L.L., T.Q.).
A new combined radiomics model accurately differentiates thymoma risk categories using computed tomography (CT) scans. This noninvasive tool aids in distinguishing low-risk from high-risk thymomas, improving patient risk stratification.
Area of Science:
- Oncology
- Radiology
- Medical Imaging
Background:
- Thymoma histological subtypes and risk categorization are challenging to distinguish.
- Accurate risk stratification is crucial for effective thymoma management.
Purpose of the Study:
- To develop a combined radiomics model for differentiating thymoma risk categories.
- To integrate computed tomography (CT) radiomics, clinical, and semantic features for improved accuracy.
Main Methods:
- Retrospective analysis of 360 thymoma patients from three centers.
- Development of a clinical-semantic model and a radiomics model (Rad_score) using AutoML.
- Construction of a combined model integrating radiomics, clinical, and semantic features.
Main Results:
- The combined model demonstrated superior performance over the clinical-semantic model in both training and validation sets.
- The combined model achieved higher accuracy (0.79) than the radiomics model (0.78) in the entire cohort.
- Specific radiomics features, like original_firstorder_median of venous phase, showed high importance.
Conclusions:
- Combined radiomics models show promise as noninvasive tools for differentiating thymoma risk classifications.
- This approach can aid in distinguishing low-risk from high-risk thymomas noninvasively.
