Identification of clinical diagnostic and immune cell infiltration characteristics of acute myocardial infarction with machine learning approach.

Acute myocardial infarction (AMI) is a serious heart disease with high fatality rates. The progress of AMI involves immune cell infiltration. However, suitable clinical diagnostic biomarkers and the roles of immune cells in AMI remain unknown. Three datasets (GSE61145, GSE34198, and GSE66360) were used from Gene Expression Omnibus. Dysregulated expression of genes was screened and functionally analyzed. Weighted Gene Co-expression Network Analysis (WGCNA) was used to identify significant module genes associated with AMI. Machine learning algorithms (Support Vector Machine (SVM), Random Forest (RF) and Least Absolute Shrinkage and Selection Operator (LASSO)) were applied to identify hub genes. Subsequently, receiver operating characteristic curves (ROC) were generated to evaluate the risk of AMI patients. Finally, immune cell infiltration were assessed by CIBERSORT, correlation analysis and immunohistochemistry. A total of 134 upregulated and 25 downregulated genes were identified. Functional analysis showed that the dysregulated genes were involved in cytokine- and immune-related signaling. Ten hub genes were used to establish a diagnostic model. Immune cell infiltration analysis showed that ten genes were correlated with activation of various immune cells; specifically, naive B cells, activated CD4 memory T cells, and resting mast cells were significantly associated with AMI. Immunohistochemical staining indicated that FOS and IL18RAP were significantly upregulated in AMI, CD4 naive T and neutrophils were significantly infiltrated in the microenvironment of AMI. The hub genes involved in activating immune cell infiltration and developing AMI could act as promising diagnostic biomarkers and targets for clinical treatment of AMI.