The efficacy of artificial intelligence in predicting mortality rate and cardiogenic shock in acute coronary syndrome patients.

OBJECTIVE

In 2023, the Ministry of Health of the Republic of Indonesia reported that deaths due to coronary heart disease reached 245,343 per year. Currently, many studies are focused on developing AI or to predict mortality rates and complications from Acute Coronary Syndrome (ACS). Therefore, the purpose of this systematic review and meta-analysis is to summarize recent research findings regarding the efficacy of AI in predicting mortality rates and the incidence of cardiogenic shock in ACS patients.

DATA SOURCES: A systematic literature search using across five databases-PubMed, Cochrane, Plos One, Scopus, and Springer-up to May 3, 2024, with additional studies included through hand searching and reference list checking.

STUDY SELECTION: Study selection was done independently by two reviewers using pre-specified inclusion and exclusion criteria.

DATA EXTRACTION: Critical appraisal was conducted using the Prediction model Risk Of Bias Assessment Tool (PROBAST) tool for AI diagnostic studies.

DATA SYNTHESIS: The reporting of this article follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement. We included 7 studies and analyzed the results using Revman v.5.4, MetaDTA, and STATA v.17 to construct forest plots for sensitivity and specificity.

CONCLUSION

The 91 % (95 % CI: 0.87-0.94) pooled specificity and 0.94 (95 % CI: 0.92-0.96) pooled AUC highlight the potential of ML models to provide more accurate risk stratification to predicting mortality rate and cardiogenic shock in ACS patients.