Several recent findings show that the accuracy of coronary heart disease diagnosis and patient risk prediction is improved using artificial intelligence (AI) models developed by scientists from the Division of Artificial Intelligence in Medicine of Cedars-Sinai.
These advances, led by Piotr Slomka, PhD, director of imaging innovation at Cedars-Sinai and a researcher in the Division of Artificial Intelligence in Medicine and the Smidt Heart Institute, helps detect and diagnose one of the most common and deadly heart diseases.
Coronary heart disease affects the arteries that supply blood to the heart muscle. If left untreated, it can lead to heart attack or other complications such as arrhythmia or heart failure.
The disease, which affects approximately 16.3 million Americans aged 20 and older, is commonly diagnosed using photon emission computed tomography (SPECT) and computed tomography (CT). However, the images generated during the scan are not always easy to read.
“We continue to show that AI can improve image quality and reveal more information, leading to more accurate disease diagnoses,” said Slomka, who is also a professor of medicine and cardiology and lead author of three studies that were recently published involving the enhancement of cardiac imaging by AI.
Using AI to Improve Cardiac Imaging
The first study, published in The Journal of Nuclear Medicineuses AI technology for cardiac imaging that helps improve the diagnostic accuracy of SPECT imaging for coronary heart disease through advanced image corrections.
In SPECT imaging, it is important to have attenuation correction, which helps reduce artifacts in cardiac images, making them easier to read and more accurate. However, this requires an additional CT scanner and expensive hybrid SPECT/CT equipment, which is essentially two scanners in one.
Although CT attenuation correction has been shown to improve the diagnosis of coronary artery disease, it is currently only performed in a minority of exams due to the extra exam time, radiation, and limited availability of this expensive technology.
To help overcome these obstacles, Slomka and his team developed a deep learning model called DeepAC to generate corrected SPECT images without the need for expensive hybrid scanners. These images are generated by AI techniques similar to those used to generate “deep-fake” videos and are able to simulate high-quality images obtained by hybrid SPECT/CT scanners.
The team compared the diagnostic accuracy of coronary heart disease using uncorrected SPECT images; which are used in most places today; advanced hybrid SPECT/CT images and new AI-corrected images in never-before-seen data from centers never used in DeepAC training.
They found that the AI created images that were close to the same quality and enabled diagnostic accuracy similar to those achieved with more expensive scanners.
This AI model was able to generate DeepAC images in a fraction of a second on standard computer software and could easily be implemented in clinical workflows as an automatic pre-processing step. »
Piotr Slomka, PhD, Director of Imaging Innovation, Cedars-Sinai
Predict major adverse cardiac events
In the second study, published in the Journal of American College of Cardiology: Cardiovascular Imagingthe team demonstrated that deep learning AI can predict major adverse cardiac events, such as death and heart attacks, directly from SPECT images.
Investigators trained the AI model using a large multinational database that included five different sites with over 20,000 patient scans. It included images illustrating each patient’s cardiac perfusion and movement.
The AI model incorporates visual explanations for physicians, highlighting images with regions that contribute to the high risk of adverse events.
The team then tested the AI model at two separate sites with over 9,000 scans. They found that the deep learning model predicted patient risk more accurately than software currently used in the clinic.
“In the first study, we were able to demonstrate that AI can be used to make large image corrections without the need for expensive scanners,” Slomka said. “In the second, we show that existing images can be better used: predicting the patient’s risk of heart attack or death from the images, and highlighting the cardiac features that indicate this risk, in order to better inform clinicians. on coronary heart disease. “
“These findings represent proof of principle for how AI can augment clinical diagnostics,” said Sumeet Chugh, MD, director of the Division of Artificial Intelligence in Medicine. “Improvements in AI-powered SPECT imaging have the potential to improve the accuracy of diagnosing coronary heart disease, while doing so much faster and more cost-effectively than current standards.”
Reduce bias in AI models
The third study, published in the European Journal of Nuclear Medicine and Molecular Imagingdescribes how to train an AI system to perform well in all applicable populations; not just the population on which the system was trained.
Some AI systems are trained using high-risk patient populations, which can cause the systems to overestimate the likelihood of illness. To ensure the AI model works accurately for all patients and reduce any bias, Slomka and his team trained the AI system using simulated patient variations. This process, called data augmentation, helps to better reflect the mix of patients scheduled for imaging tests.
They found that models trained with a balanced mix of patients more accurately predicted the likelihood of coronary heart disease in women and low-risk patients, which can potentially lead to less invasive testing and more accurate diagnosis in women. .
The models also led to a drop in false positives, suggesting the system can potentially reduce the number of tests the patient undergoes to rule out the disease.
“The results suggest that improving training data is key to ensuring that AI predictions more closely reflect the population to which they will be applied in the future,” Slomka said.
Researchers are currently evaluating these new AI approaches at Cedars-Sinai and exploring how they can be integrated into clinical software and how they might be used in standard patient care.
The research was funded in part by the National Heart, Lung, and Blood Institute.
Shanbhag, AD, et al. (2022) Deep learning-based attenuation correction improves diagnostic accuracy of cardiac SPECT. Journal of Nuclear Medicine. doi.org/10.2967/jnumed.122.264429.
#models #improve #diagnostic #accuracy #coronary #heart #disease