Related Experiment Video
Updated: Jun 27, 2026
06:31
Erosion Identification in Metacarpophalangeal Joints in Rheumatoid Arthritis using High-Resolution Peripheral Quantitative Computed Tomography
Published on: October 6, 2023
High-Frequency Ultrasound Radiomics Combined with Clinical Features for Detecting OMERACT-Defined Metacarpophalangeal
Minghui Yao1, Wenxue Li1, Yuwei Xin1
1Department of Ultrasound, Peking University People's Hospital, Beijing 100044, China.
Diagnostics (Basel, Switzerland)
|June 26, 2026
Summary
High-frequency ultrasound radiomics combined with clinical data accurately assesses metacarpophalangeal joint cartilage damage in early rheumatoid arthritis (RA). This quantitative model offers a valuable extension to current diagnostic methods for RA patients.
Area of Science:
- Medical Imaging
- Rheumatology
- Artificial Intelligence
Background:
- Early rheumatoid arthritis (RA) diagnosis and monitoring are crucial for effective treatment.
- Quantitative assessment of metacarpophalangeal (MCP) joint cartilage damage is challenging with conventional methods.
Purpose of the Study:
- To develop and validate a high-frequency ultrasound radiomics model for quantitative assessment of MCP joint cartilage damage in early RA.
- To compare the performance of radiomics, clinical, and combined models in detecting cartilage damage.
Main Methods:
- Radiomics feature extraction from high-frequency ultrasound images of 656 MCP joints (99 early RA patients, 65 controls).
- Evaluation of five machine learning classifiers, with Support Vector Machine selected for the final radiomics model.
- Construction and assessment of radiomics, clinical, and combined diagnostic models using SHapley Additive exPlanations (SHAP) for interpretation.
Main Results:
- A radiomics model using eight stable features achieved an AUC of 0.804.
- The combined model (radiomics + age, disease duration, DAS28) demonstrated superior diagnostic performance (AUC = 0.855).
- Elevated radiomics scores were observed in high-risk, early RA (OMERACT grade 0) joints, with PerimeterSurfaceRatio being a key feature.
Conclusions:
- High-frequency ultrasound radiomics, when integrated with clinical factors, shows strong potential for detecting MCP joint cartilage damage in early RA.
- This quantitative approach can serve as a valuable adjunct to traditional semiquantitative assessments in RA management.
