Machine learning combined with PET/CT can predict heart attack risk
Merging machine learning with information obtained from PET/CT angiography scans can predict a patient's likelihood of experiencing a future heart attack, according to new research published in the Journal of Nuclear Medicine.
Predicting a patient’s risk of having a heart attack can be difficult, even when they have a condition like coronary artery disease, which is known to increase the probability of major cardiac events. Though cardiovascular risk scores are valuable assessments of a person’s cardiac health, they simply are not enough to sufficiently assess risk, experts noted.
Advanced imaging modalities and developments in artificial intelligence have shown promise for identifying patients who are most at risk of having a heart attack, and a new study evaluates how combining the two technologies could fine tune cardiac risk stratification.
“Our goal in the study was to investigate whether the information provided by 18F-NaF PET and CT angiography is complementary and could improve prediction of heart attacks with the use of artificial intelligence techniques,” Jacek Kwiecinski, with the Division of Artificial Intelligence in Medicine at Cedars-Sinai Medical Center, and co-authors explained.
The researchers examined 293 participants with known coronary artery disease. Each underwent a coronary 18F-NaF PET and CT angiography exam on a hybrid PET/CT scanner. The information obtained from those scans were then combined with the patients’ clinical data to train the machine learning model to identify abnormalities and risk factors.
“Using the information from these approaches and by leveraging machine learning, we were able to build an integrated model for prediction of events in patients with established coronary artery disease,” the experts clarified.
During the 53 month follow-up period, 22 out of 293 patients experienced a myocardial infarction. When independently evaluated, only 18F-NaF coronary uptake was a predictive indicator of heart attack, while clinical data in conjunction with machine learning had a meager performance. But when all the available data was integrated together the machine learning model’s performance exceeded all other approaches.
“We showed that risk prediction does not depend on cardiovascular risk scores, stenosis severity or CT calcium scoring,” the experts said. “Rather, the risk of myocardial infarction is primarily governed by the analysis of plaque type and plaque burden provided by coronary CT angiography and assessments of disease activity by 18F-NaF PET.”
You can view the detailed research in the Journal of Nuclear Medicine.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She joined Innovate Healthcare in 2021 and has since put her unique expertise to use in her editorial role with Health Imaging.