Artificial intelligence can objectively determine cardiac calcium scores faster than doctors
![A study published this week in the Journal of the American College of Cardiology (JACC): Cardiovascular Imaging shows artificial intelligence (AI) algorithms can more rapidly and objectively determine calcium scores in computed tomographic (CT) and positron emission tomographic (PET) images than physicians.[1] The AI also performed well when the images were obtained from very-low-radiation CT attenuation scans. https://doi.org/10.1016/j.jcmg.2022.06.006](/sites/default/files/styles/top_stories/public/2022-09/ai_cac_calcium_score_reading_study_for_pet_ct.jpg.webp?itok=GoiezUxV)
A study published this week in the Journal of the American College of Cardiology (JACC): Cardiovascular Imaging shows artificial intelligence (AI) algorithms can more rapidly and objectively determine calcium scores in computed tomography (CT) and positron emission tomography (PET) images than physicians.[1] The AI also performed well when the images were obtained from very-low-radiation CT attenuation scans.
The aim of this study was to develop a deep learning (DL) model capable of fully automated coronary artery calcium (CAC) definition from PET CT attenuation correction (CTAC) scans. CT attenuation correction scans are always acquired with cardiac PET imaging to calibrate the nucleating imaging, so many canters have started using the data from these scans for CAC scoring as an added value for patients.
The deep learning CTAC scores predict cardiovascular risk similarly to standard CAC scores quantified manually by experienced operators from dedicated ECG gated CAC scans and can be obtained almost instantly, with no changes to PET/CT scanning protocol.
"Using these artificial intelligence and deep learning techniques requires less imaging, less radiation and lower costs," explained senior study author Piotr Slomka, PhD, in a statement. He is a research scientist in the Smidt Heart Institute at Cedars-Sinai in Los Angles, director of Innovation in Imaging and professor of cardiology and medicine in the Division of Artificial Intelligence in Medicine.
Calcium scoring has become a standard-of-care for evaluating a patient's risk of future heart attacks. The calcium shows past areas that healed after a soft plaque became inflamed and ruptured. The higher the calcium score, the higher the risk of a future cardiac event. Patients with a score of zero have extremely low risk of developing coronary disease. Cardiologists often will use the CAC score to determine if a patient needs to go on statin therapy.
Calcium soaring is increasingly being used as as a risk assessment in patients and is being incorporated into PET/CT scans, and can be read as an incidental finding on many types of chest CT scans. Automating this review with AI was been of interest by vendors and hospitals as a way to make low-cost coronary CT screening faster and more economically viable. While a couple vendors offer this type of automation, validation in clinical trials is what is needed to show these algorithms can be relied on.
The novel deep learning model, originally developed for video applications, was adapted to rapidly quantify coronary artery calcium. The model was trained using 9,543 expert-annotated CT scans and was tested in 4,331 patients from an external cohort undergoing PET/CT imaging with major adverse cardiac events. Same-day paired electrocardiographically (ECG) gated CAC scans were available in 2,737 patients.
The CT attenuation maps were obtained with PET/CT scans and could be processed by artificial intelligence techniques for rapid and objective determination of coronary calcium score without additional scan and radiation.
This research was supported in part by grant R01HL089765 from the National Heart, Lung, and Blood Institute/National Institutes of Health (principal investigator, Piotr Slomka).
Dave Fornell has covered healthcare for more than 17 years, with a focus in cardiology and radiology. Fornell is a 5-time winner of a Jesse H. Neal Award, the most prestigious editorial honors in the field of specialized journalism. The wins included best technical content, best use of social media and best COVID-19 coverage. Fornell was also a three-time Neal finalist for best range of work by a single author. He produces more than 100 editorial videos each year, most of them interviews with key opinion leaders in medicine. He also writes technical articles, covers key trends, conducts video hospital site visits, and is very involved with social media. E-mail: dfornell@innovatehealthcare.com