Table 1_Development and external validation of an interpretable machine learning model for

A machine learning model developed and validated using clinical data from 2,327 patients across two Chinese hospitals. The model uses nine routinely available clinical parameters to diagnose angiographically confirmed coronary heart disease. The work by Chunxia Deng was last updated in May 2026 and includes a web-based calculator for risk prediction.

Use Cases

• Diagnosing coronary heart disease based on nine clinical predictors including total cholesterol and chest distress.
• Validating machine learning models for clinical use based on external validation cohort performance.
• Assessing individualized CHD risk based on features like age, sex, and cardiac troponin levels.
• Enhancing clinical interpretability of ML models based on SHAP analysis for patient communication.

Strengths

• Model performance is quantified with an AUC of 0.896 in internal validation and 0.865 in external validation.
• The dataset underpinning the model includes 1,269 patients for development and 1,058 for external validation.
• Model interpretability is addressed using SHAP analysis on nine selected clinical features.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying patient data is unknown, which may limit suitability assessment.
• Data may reflect geographic bias inherent to the two Chinese hospital sources.

Provenance

Source

Chunxia Deng via figshare.

Collection Method

Clinical data collected from patients at two Chinese hospitals.

Freshness

Last updated 2026-05-15 04:26:31; freshness should be verified.

Geography

China