New insights from the AI-ENCODE examine confirmed synthetic intelligence (AI) efficiently allowed the automated extraction of key purposeful and physiological knowledge from routine angiograms. The late-breaking outcomes have been introduced in the present day on the Society for Cardiovascular Angiography & Interventions (SCAI) 2024 Scientific Periods.
Coronary angiography is a broadly 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 skills of standard coronary angiography by increasing its diagnostic scope which might enhance medical decision-making and positively affect affected person outcomes.
The AI-ENCODE examine leveraged superior machine-learning methods to develop the vary of knowledge 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 operate, intracardiac filling pressures, and cardiac index from one to 2 angiographic movies. Echocardiograms carried out near the angiogram served because the ‘gold customary’ 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 typically have required extra assessments similar to echocardiograms and/or proper coronary heart catheterization.
Conventional diagnostic instruments in cardiovascular drugs harbor huge info, however a lot stays underutilized. The AI-ENCODE mission proved that AI may be leveraged to unlock and ship a broader, extra significant spectrum of medical findings from current angiograms. This examine actually exhibits us AI’s prowess in revealing insights past what the human eye can see. It will be important that we leverage AI’s capabilities as a diagnostic instrument to offer the very 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 look ahead to 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, Might 2, 2024; 9:45-9:52 AM PT
- Lengthy Seashore Conference Heart, 104A, First Stage