Cluster analysis of clinical, angiographic, and laboratory parameters in patients with ST-segment elevation...

Area of Science:

• Cardiology
• Machine Learning in Healthcare
• Biostatistics

Background:

• ST-segment elevation myocardial infarction (STEMI) is a severe manifestation of coronary artery disease.
• Accurate risk stratification is crucial for guiding STEMI treatment and discharge planning.
• Hierarchical agglomerative clustering (HAC) offers a novel machine learning approach for patient categorization.

Purpose of the Study:

• To investigate the utility of HAC in stratifying STEMI patients into distinct phenotypic clusters.
• To compare the clinical outcomes, specifically mortality, between the identified patient clusters.

Main Methods:

• A cohort of 3205 STEMI patients diagnosed between 2015 and 2023 was analyzed.
• Hierarchical agglomerative clustering (HAC) was applied to categorize patients into two distinct clusters.
• Outcomes, including mortality, were compared between the clusters using statistical analyses (chi-square, log-rank, Cox regression).

Main Results:

• STEMI patients were divided into Cluster 1 (2731 patients) and Cluster 2 (474 patients).
• Cluster 2 exhibited significantly higher mortality rates (23%) compared to Cluster 1 (5.4%) (P < 0.01).
• Survival analysis and Cox regression confirmed a substantially increased risk of death in Cluster 2 (HR = 3.51, P < 0.001), even after adjusting for age and sex.

Conclusions:

• Hierarchical agglomerative clustering (HAC) effectively categorizes STEMI patients into groups with differing prognoses.
• The HAC method shows potential as a predictive tool for one-month mortality in STEMI patients.
• This machine learning approach may aid in refining risk assessment and treatment strategies for STEMI.