Machine Learning for Clinical Decision-Making: Challenges and Opportunities in Cardiovascular Imaging

Area of Science:

• Cardiology
• Medical Informatics
• Artificial Intelligence

Background:

• Machine learning (ML) holds significant potential to revolutionize clinical decision-making in cardiology.
• Understanding clinicians' conventional decision-making pathways is crucial for successful ML tool implementation.

Purpose of the Study:

• To discuss the state-of-the-art and clinical status of ML in cardiovascular imaging.
• To identify challenges in ML interpretation and decision support within cardiology.

Main Methods:

• Review of current ML applications in cardiovascular imaging.
• Analysis of ML's impact across data acquisition, feature extraction, interpretation, and decision support.
• Focus on challenges in interpretation, decision support, learning processes, auditability, infrastructure, and clinical integration.

Main Results:

• ML can enhance data acquisition and feature extraction, reducing clinician workload.
• ML offers potential for complex data interpretation and augmented patient status understanding.
• ML can support clinical decisions by predicting outcomes and recommending interventions.

Conclusions:

• ML presents a paradigm shift in cardiovascular imaging, particularly in interpretation and decision support.
• Addressing challenges in learning, auditability, infrastructure, and integration is key for widespread clinical adoption.
• Further research and development are needed to fully realize ML's potential in clinical cardiology.