Abstract

This work aims to provide a thorough examination of the existing research on the use of pre-trained artificial intelligence models for predicting and categorizing myocardial infarction, emphasizing the various models, methodologies, and findings documented in the literature. This research is a literature review examining the use of pre-trained artificial intelligence models for myocardial infarction prediction and classification. We utilized research publications from academic sources dealing with myocardial infarction and artificial intelligence applications. The technique covered artificial intelligence models (particularly BERT, ResNet, and XGBoost), data sources (EHR, ECG, and CMR), methods (transfer learning, deep learning, and machine learning), and the quality of the literature review was evaluated using the Scale for the Assessment of Narrative Review Articles (SANRA). Pre-trained artificial intelligence models, particularly CNN and transformer architectures, have significant promise in the prediction and categorization of myocardial infarction. These models provide elevated accuracy, prompt detection, and tailored methodologies, while optimizing data and computing resource use. This research thoroughly studied the use of pre-trained artificial intelligence models to predict and classify myocardial infarction. Our findings indicate that models based on CNN and transformer topologies, in particular, provide considerable benefits in the diagnosis and treatment of myocardial infarction, with the possibility for early detection and a tailored strategy. However, concerns such as data quality, model interpretability, and the requirement for thorough validation must be addressed in clinical practice. Future research addressing these constraints and concentrating on the practical and ethical implementation of AI-based solutions in cardiology has the potential to enhance patient outcomes and usher in a new age of precision medicine. This review was created using the Scale for the Assessment of Narrative Review Articles (SANRA) criteria, which included the topic's relevance, defined goals, a literature search, and acceptable source citing.

Key words: Myocardial infarction, artificial intelligence, transfer learning, scale for the assessment of narrative review articles