Distinct immune-metabolic phenotypes underlie poor coronary collateral circulation

Area of Science:

• Cardiology
• Molecular Biology
• Genomics

Background:

• Coronary collateral circulation (CCC) is crucial for myocardial perfusion in coronary artery disease (CAD).
• The molecular basis of poor CCC remains poorly understood.
• Understanding CCC heterogeneity is vital for improving patient outcomes.

Purpose of the Study:

• Identify distinct molecular phenotypes in patients with poor CCC.
• Validate these phenotypes using clinical data and assess their prognostic value.
• Evaluate the potential for personalized therapeutic strategies.

Main Methods:

• Proteomic profiling of 149 patients (69 with poor CCC).
• Unsupervised clustering to identify molecular subtypes within poor CCC.
• Machine learning (XGBoost) for clinical data modeling and SHAP value interpretation.
• External validation using the MIMIC database and survival analysis for MACE.

Main Results:

• Two phenotypes identified: Complement-Driven Vascular Remodeling (CDVR) and Immuno-Thrombotic Myocardial Dysfunction (ITMD).
• An XGBoost model using fasting glucose, eosinophil percentage, and HbA1c achieved high discrimination (AUC > 0.91).
• ITMD phenotype showed significantly higher MACE incidence and upregulated platelet activation, diabetic cardiomyopathy, and metabolic pathways.

Conclusions:

• Poor CCC is characterized by distinct immune-metabolic phenotypes.
• Integrated proteomic-clinical modeling accurately classifies these phenotypes.
• This classification improves risk stratification and may guide personalized therapies for CAD patients with inadequate collateralization.