Development and validation of a machine learning model for predicting vulnerable carotid plaques using routine blood biomarkers and derived indicators: insights into sex-related risk patterns
Authors' conclusion
A Random Forest model integrating routine blood markers and derived indices accurately predicted vulnerable carotid plaques. The results underscore the importance of sex-specific risk assessment, highlighting differential effects of key biomarkers across genders and value intervals.
Study context
- Population
- Hospitalized patients at Suzhou Municipal Hospital (2019–2020) who underwent carotid US, selected from an initial cohort of 10,028from initial cohort of 10,028
- Participants
- 1,701 patients
- Exposure
- Five machine learning models compared for predicting vulnerable carotid plaques
- Outcome
- Prediction of vulnerable carotid plaques (predefined US imaging criteria); performance by AUC, sensitivity and specificity
Methodological detailssingle-centre, machine learning prediction model development and validation, 7:3 train-test split, cross-validation, SHAP interpretation, restricted cubic spline modeling
- Single-centre
- Machine learning prediction model development and validation
- 7:3 train-test split
- Cross-validation
- SHAP interpretation
- Restricted cubic spline modeling
Main finding
As reported in the abstract
Predictive performance for vulnerable carotid plaques
| Group | AUC | Specificity | Sensitivity |
|---|---|---|---|
| Random Forest | 0.847 (95% CI 0.791-0.895) | 89.4% | 64.2% |
Random Forest
AUC0.847 (95% CI 0.791-0.895)
Specificity89.4%
Sensitivity64.2%
Observational data: associations, not effects.
Further findings
Top predictors (SHAP analysis)
- Top predictors: gender, age, fibrinogen, neutrophil-to-lymphocyte ratio (NLR), creatinine, fasting blood glucose, UHR, TyG, systemic inflammation response index (SIRI) and lymphocyte count
Sex-specific differences
- Significant sex-specific differences in SHAP values for key biomarkers, including age, UHR, TyG, SIRI and others
- Restricted cubic spline (RCS) modeling revealed distinct sex-related patterns in plaque vulnerability across biomarker value ranges
Original publication
Development and validation of a machine learning model for predicting vulnerable carotid plaques using routine blood biomarkers and derived indicators: insights into sex-related risk patterns. Cardiovasc Diabetol. Published 10 August 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →