Early atherosclerosis detection
Software can accurately identify plaque formation in arteries
The merging of artificial intelligence and decades-old cardiology technology has the potential to make early atherosclerosis detection more accurate and detailed. Atherosclerosis is an inflammation characterized by the formation of plaque—made from fat, calcium, and other substances—on coronary artery walls and elsewhere in the human body.
Researchers from the University of Campinas (UNICAMP), in partnership with the Federal University of Santa Catarina (UFSC), have developed an AI algorithm capable of automatically identifying changes in the thickness of the carotid artery’s tunica intima—the thinnest wall of the vessel that transports blood and oxygen to the brain—using ultrasound images.
According to one of the researchers responsible for the breakthrough, cardiologist Wilson Nadruz Junior, of UNICAMP, the image processing and analysis software uses techniques based on topological extinction values and mathematical morphology to restrict regions of interest and highlight desired aspects. “The algorithm allows ultrasonographers to use routine clinical equipment to automatically obtain precise tunica intima measurements, which are generally taken by trained professionals using research equipment.”
The technology is protected by the UNICAMP Innovation Agency, and has a patent filed with the Brazilian National Institute of Industrial Property (INPI). Currently, this method of targeting the carotid artery’s tunica intima is part of UNICAMP’s technology portfolio and is available for licensing.
Nadruz Junior explains that, at present, one of the most precise ways to detect atherosclerosis is through a very specific measurement taken of the artery wall, more specifically of a structure known as the intima-media complex. “Basically, it is a measurement of the thinnest layer—intimal tissue—and a thicker part of the carotid artery.” The problem is that atherosclerosis is detected using invasive imaging methods, such as catheterization and angiotomography, comments the researcher, emphasizing that these techniques are also more expensive.
A neck ultrasound, on the other hand, is a noninvasive method that has been used for more than 30 years to identify plaques likely to increase the risk of cardiovascular disease. Additionally, the intima-media complex can be measured using ultrasonography. However, this exam cannot differentiate between changes caused by increased tunica intima thickness (the most specific measure of atherosclerosis) and increased tunica media thickness, which is caused by muscle cell growth, and which has less prognostic value.
Studies conducted at UNICAMP and foreign institutions show that changes in the tunica intima, which has more contact with the blood, would be associated with more serious cardiovascular issues. “The greatest challenge is detecting, in a quick and routine manner, the exact dimensions of the tunica intima, because of how thin it is,” emphasizes Nadruz Junior.
“The breakthrough is that we were able to develop an algorithm that, when fed carotid images obtained using routine ultrasound machines, outputs the tunica intima measurement. We analyzed approximately 200 images and, recently, obtained a patent for the technology,” states Nadruz Junior.
The research project relied on the contributions of Professors José Roberto Matos Souza (of the UNICAMP School of Medical Sciences), Rangel Arthur (of the UNICAMP School of Technology), and Alexandre Silva (of the Federal University of Santa Catarina).
“It is important to clarify that this tool can help with risk stratification, which goes hand-in-hand with clinical assessment. In no way is it something that will replace the physician’s role,” warns Nadruz Junior. “Clinical assessment of the patient will never cease to be important.”
At the same time, this is not to say that the traditional carotid exam, which is always accompanied by blood tests and other imaging reports, will not soon be a thing of the past. “The idea is that, in the future, this algorithm will be embedded in ultrasound scans. Therefore, the traditional exam, despite its importance over the last three decades, may be replaced,” says the researcher.
The group that developed the technology believes that commercial application of the software will be attractive, especially because it can be used on any type of device and requires no special training. “The technology could have enormous potential in the prevention, early detection, and reduction of future cardiovascular diseases,” says the cardiologist.
Tunica intima measurements are currently taken by specially trained professionals and require research equipment only available in academic laboratories. “It has taken five long years of work and the artificial intelligence researchers have done very well. It’s not easy doing what they did. They managed to greatly improve the interpretation of ultrasound images,” says Nadruz Junior.