New insights from the AI-ENCODE examine confirmed synthetic intelligence (AI) efficiently allowed the automated extraction of key practical and physiological knowledge from routine angiograms. The late-breaking outcomes have been offered as we speak on the Society for Cardiovascular Angiography & Interventions (SCAI) 2024 Scientific Classes.
Coronary angiography is a extensively used diagnostic process for figuring out coronary artery illness (CAD), carried out in over a million sufferers yearly in america. At present, knowledge extracted from coronary angiograms are confined to detecting blockages within the coronary arteries. Nonetheless, AI expertise has the potential to broaden the diagnostic talents of typical coronary angiography by increasing its diagnostic scope which may enhance medical decision-making and positively affect affected person outcomes.
The AI-ENCODE examine leveraged superior machine-learning strategies to develop the vary of information obtained from routine coronary angiograms. Utilizing a library of greater than 20,000 angiograms carried out on the Mayo Clinic from 2016 to 2021, the examine group developed and validated a number of AI algorithms to extract knowledge on left and proper ventricular perform, intracardiac filling pressures, and cardiac index from one to 2 angiographic movies. Echocardiograms carried out near the angiogram served because the ‘gold normal’ for the comparability.
The AI fashions predicted left ventricular ejection fraction (LVEF), LV filling pressures, proper ventricular dysfunction, and CI with an space underneath a receiver working curve (AUC) of 0.87, 0.87, 0.80, and 0.82, respectively. These outcomes present that novel AI fashions have been in a position to extract key diagnostic knowledge that would normally have required further exams corresponding to echocardiograms and/or proper coronary heart catheterization.
Conventional diagnostic instruments in cardiovascular drugs harbor huge data, however a lot stays underutilized. The AI-ENCODE undertaking proved that AI might be leveraged to unlock and ship a broader, extra significant spectrum of medical findings from current angiograms. This examine really exhibits us AI’s prowess in revealing insights past what the human eye can see. It is necessary that we leverage AI’s capabilities as a diagnostic device to supply the absolute best for our sufferers.”
Mohamad Alkhouli, MD, Division Chair of Analysis and Innovation at Mayo Clinic Faculty of Drugs, and lead writer of the examine
The examine authors stay up for additional analyses of AI’s capabilities within the cardiovascular area.
Session particulars:
- “Synthetic Intelligence for Extracting Non-Coronary Information from Angiography: The AI-ENCODE Research”
- Thursday, Could 2, 2024; 9:45-9:52 AM PT
- Lengthy Seashore Conference Heart, 104A, First Degree
