Cardiac imaging and AI: 5 things to know

Seven radiological organizations have issued a joint scientific statement regarding the use of AI in cardiac imaging, published Jan. 28 in Radiology. 

The statement was drafted by the European Society of Cardiovascular Radiology, the European Society of Medical Imaging Informatics, the North American Society for Cardiovascular Imaging, the Society of Cardiovascular Computed Tomography, the Society for Cardiovascular Magnetic Resonance, the Society for Imaging Informatics in Medicine and the Radiological Society of North America.

Here are five things to know from the statement:

1. The organizations provided an assessment of various AI applications for cardiac imaging and assigned each application a technology readiness level score. The scores reflect the application's maturity within cardiac CT and MRI.
   
   
2. The technology readiness levels were:
   
   
   • TRL 1: Basic principles reported
   • TRL 2: Technology concept and/or application formulated
   • TRL 3: Analytical and experimental function and/or proof of concept
   • TRL 4: Component validation in experimental setting
   • TRL 5: Component validation in relevant setting
   • TRL 6: System/subsystem or prototype demonstration in relevant setting
   • TRL 7: System prototype demonstrated in clinical setting
   • TRL 8: Actual system completed, tested and demonstrated in clinical setting
   • TRL 9: Actual system proven in clinical setting
     
     
3. The applications and technology readiness scores shared in the statement were:
   
   
   • Patient and imaging test selection: TRL 1-2
   • Cardiac imaging workflow: TRL 1-2
   • Scanning protocol selection: TRL 1-2
   • Prognostication for major adverse cardiac events: TRL 1-5
   • Large language models for communication between cardiac imagers, referring clinicians and patients: TRL 3-8
   • Cardiac image acquisition and reconstruction: TRL 5-9
   • Coronary CT angiography analysis: TRL 5-9
   • Cardiac MRI quantitative analysis: TRL 5-9
   • Non-cardiac findings detection on cardiac imaging: TRL 6-9
   • Large language models for cardiac imaging reports: TRL 6-9
   • Cardiac findings detection on non-cardiac imaging: TRL 7-9
     
     
4. AI applications that have already been successfully deployed within cardiac CT and MRI are: coronary CT angiography deep learning reconstruction, coronary CT angiography plaque analysis, CT-derived fractional flow reserve, cardiac MRI biventricular function assessment and incidental CAC detection.
   
   
5. In addition to technology readiness, environmental sustainability, ethics and legal implications should be considered when incorporating AI into cardiac imaging, the statement authors said. 

Read the full scientific statement, including more in-depth analysis of each application, here.