Construction of machine learning diagnostic models for cardiovascular pan-disease based on blood routine and biochemical detection data
Authors' conclusion
The present study constructed multiple models using 50 features from the blood routine and biochemical detection data for the diagnosis of various circulation system diseases. At the same time, the unique hematologic features of various circulation system diseases, including some metabolic-related indicators, were also explored. This cost-effective work will benefit more people and help diagnose and prevent circulation system diseases.
Study context
- Population
- Healthy people and circulation system disease patients with blood routine and biochemical detection data
- Participants
- 25,794 vs 32,822
- Comparison
- Machine learning models for diagnosis of circulation system disease
- Healthyn = 25,794reference
- Circulation system diseasen = 32,822
- Outcome
- Circulation system disease prediction, XGBoost model (best performance)
Methodological detailsmachine learning models, SHAP interpretation
- Machine learning models
- SHAP interpretation
Performance
As reported in the abstract
| Test or model | AUC | Accuracy | Sensitivity | Specificity | PPV | NPV | MCC | F1 score |
|---|---|---|---|---|---|---|---|---|
| XGBoost | 0.9921 (0.9911-0.9930) | 0.9618 (0.9588-0.9645) | 0.9690 (0.9655-0.9723) | 0.9526 (0.9477-0.9572) | 0.9631 (0.9592-0.9668) | 0.9600 (0.9556-0.9644) | 0.9224 (0.9165-0.9279) | 0.9661 (0.9634-0.9686) |
XGBoost
AUC0.9921 (0.9911-0.9930)
Accuracy0.9618 (0.9588-0.9645)
Sensitivity0.9690 (0.9655-0.9723)
Specificity0.9526 (0.9477-0.9572)
PPV0.9631 (0.9592-0.9668)
NPV0.9600 (0.9556-0.9644)
MCC0.9224 (0.9165-0.9279)
F1 score0.9661 (0.9634-0.9686)
Further findings
Distinguishing dilated cardiomyopathy from other circulation system diseases (best subtype model)
- AUC
- 0.9267 (0.8663-0.9752)
- Most models distinguishing various circulation system diseases also had good performance
SHAP feature importance
- Biochemical features (K, total protein (TP), albumin (ALB), NBIL) made major contributions to predicting circulation system disease
- Top 4 features subdividing circulation system diseases: RBC, K, direct bilirubin (DBIL), glucose (GLU)
Original publication
Construction of machine learning diagnostic models for cardiovascular pan-disease based on blood routine and biochemical detection data. Cardiovasc Diabetol. Published 28 September 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →