Related Experiment Video
Updated: Jun 11, 2025
08:33
Author Spotlight: Methodologies and Advancements of Chronic Pain Management Research
Published on: January 5, 2024
1.1K
Identification of texture MRI brain abnormalities on Fibromyalgia syndrome using interpretable machine learning
Hongyang Jiang1,2, Aihui Liu3, Zhenhua Ying4,5
1Medical College of Soochow University, Suzhou, China.
Scientific Reports
|October 9, 2024
Summary
Objective diagnostic markers for fibromyalgia symptoms (FMS) were developed using radiomics and machine learning. These models aid in diagnosing chronic pain (CP) and distinguishing FMS subgroups with improved accuracy.
Area of Science:
- Neuroimaging
- Radiomics
- Machine Learning
Background:
- Fibromyalgia symptoms (FMS) lack objective diagnostic markers, leading to diagnostic challenges.
- Chronic pain (CP) diagnosis and FMS subgroup differentiation require improved methods.
Purpose of the Study:
- To develop interpretable machine learning models for objective FMS diagnosis.
- To create radiomics-based nomogram models for differentiating FMS subgroups.
Main Methods:
- Extreme gradient boosting (XGBoost) models were trained on radiomics features from brain imaging.
- Feature selection utilized Mann-Whitney U, Spearman's rank correlation, and LASSO.
- Shapley Additive exPlanations (SHAP) provided model interpretability.
- A nomogram integrated radiomics scores and clinical predictors for subgroup diagnosis.
Main Results:
- The XGBoost model demonstrated stable performance, indicating low overfitting for CP diagnosis.
- The nomogram model, incorporating radiomics scores, effectively distinguished typical from sub-clinical FMS.
- The combined radiomics and clinical model outperformed clinical factors alone in subgroup differentiation.
Conclusions:
- Interpretable radiomics-based XGBoost models offer objective diagnostic potential for chronic pain.
- Machine learning-derived radiomics scores enhance nomogram models for precise fibromyalgia subgroup diagnosis.
