Machine learning models for positron emission tomography myocardial perfusion imaging.

Area of Science:

• Cardiology
• Medical Imaging
• Artificial Intelligence

Background:

• Machine learning (ML) models are being developed to analyze myocardial perfusion imaging (MPI) from positron emission tomography (PET).
• These models aim to automate the detection of ischaemia and scar tissue in cardiac PET scans.
• The goal is to assist in the interpretation of PET scans, identifying potential abnormalities in patients and specific coronary vessels.

Discussion:

• The integration of ML-driven PET imaging analysis into routine clinical workflows presents challenges.
• Standardizing reporting and ensuring seamless incorporation of these advanced tools are crucial for adoption.
• The impact of these automated assessments on clinical decision-making requires further investigation.

Key Insights:

• ML algorithms show promise for automating the assessment of ischaemia and scar in cardiac PET MPI.
• Automated identification of abnormalities can potentially streamline the reporting process for PET scans.
• These tools may help flag patients and specific vessels with a higher likelihood of cardiac abnormalities.

Outlook:

• Further research is needed to determine the optimal methods for integrating ML-based PET analysis into clinical practice.
• Clinical trials are necessary to evaluate the real-world impact of these technologies on patient management and clinical outcomes.
• The future may see widespread use of AI in cardiac imaging, enhancing diagnostic accuracy and efficiency.