- According to recent studies, having a myocardial infarction, often known as a heart attack, increases the likelihood of having complicated blood vessels in the retina of the eye.
- The finding was achieved by analyzing data that represented a sizable group of people using artificial intelligence and "deep learning."
- When the algorithm takes into account demographic information, the researchers can properly forecast heart attacks when combined with new genetic findings.
Retinal scans might soon be used to foretell heart attacks. According to recent studies, the human retina's blood vessels' reduced complexity serves as an early indicator of myocardial infarction.
In an interview with Medical News Today, ophthalmologist Dr. Howard R. Krauss remarked, "For decades, I've always preached that the eye is not just the window to the soul, but the window to the brain and the window to the body as well."
"Deep learning" combined with AI [artificial intelligence] is demonstrating that to be the case, he added.
Dr. Rigved Tadwalkar, a cardiologist who was not involved in the study, told MNT that the results were intriguing.
The features of the ocular vasculature can be exploited for individual risk prediction for myocardial infarction, the author noted, "although we have known that evaluation of the retinal vasculature might yield insights on cardiovascular health."
''This [research] adds another tool to the arsenal for figuring out who could benefit from earlier preventative action (in the case of heart attacks).''
— Rigved Tadwalkar, M.D.
The photographic station may be remote from the clinician and maybe even available via a smartphone in the future, according to Dr. Krauss, who was not involved in the study.
On June 12, the study was presented to the European Society of Human Genetics.
Blood vessels and retinal scanning
Following a news release, the experiment used information from the UK Biobank, which houses retinal scans and more than 500,000 people's demographic, epidemiological, clinical, and genetic data. Age, se_, smoking status, systolic blood pressure, and body mass index were all listed under demographic information (BMI).
About 38,000 white British volunteers with retinal scans who later suffered heart attacks were identified by the researchers. For these people, the biobank supplied genetic data and retinal fundus photos.
There are two sizable networks of blood veins, or vasculature, at the rear of the retina, on either side of where it links to the optic nerve. Each has a similarly intricate fractal design and resembles a tree branch in a healthy person.
But for other people, branching is remarkably simplified and this intricacy is essentially gone.
In this study, a relationship between poor retinal vascular complexity and coronary artery disease was found using artificial intelligence (AI) and a deep learning model.
The strength of AI
''The benefit of using AI and deep learning, according to Dr. Krauss, is that as the database grows, one may discover relationships and the predictive value of retinal assessments in ways that we would not even be aware of at this time.''
Dr. Krauss said that using AI and deep learning for this type of study has benefits.
"AI is capable of analyzing a photo and identifying the gender of the subject with a 97 percent accuracy. Beyond guessing, he continued, "no ophthalmologist can look in the eye, or look at a picture, and tell you if it's male or female.
When only taking vascular density into account, the AI model performed rather well. Although its accuracy was "substantially decreased when compared with a model that contained demographic data, and with existing risk models," Villaplana-Velasco claimed otherwise.
"We discovered a considerable improvement," she added, "even when we merely introduced age and se_ to retinal vascular complexity."
Certain genetic areas
The researchers' model's prediction ability was further enhanced by a third element.
Lead author and Ph.D. candidate Ana Villaplana-Velasco stated, "our genetic study demonstrated that four genetic areas associated with retinal vascular complexity have a function in biological processes connected to MI."
She stated that her team was interested in working with other in-vitro experiment-focused research organizations to learn more about this connection.
Dr. Tadwalkar added, "The findings make sense because a real link between fractal dimension [complexity] and incident cardiovascular disease was discovered.
The polygenic risk score, which may considerably increase accuracy on its own, is also integrated into the model, he said.
Recognizing symptoms of other illnesses
According to Dr. Krauss, the retina may include information about several systemic disorders.
Since the development of the ophthalmoscope more than a century ago, modern medicine has utilized the retina to identify several systemic disorders. We can now detect retinal alterations before an ophthalmologist would normally, thanks to the application of AI and deep learning networks, he added.
Villaplana-Velasco told MNT, "These alterations are not exclusive for increased MI risk." She went on to say that these findings may also apply to other conditions.
"We think that each disorder could have a different profile of retinal vascular changes."
— Lead author Ana Villaplana-Velasco
The results, according to Dr. Tadwalkar, "are a step in the right direction and [give] at least some explanation." However, "we would need to see further studies that replicate the findings."
Drs. Tadwalkar and Krauss both voiced worry about the narrow variety of people covered in the data. According to Dr. Krauss, the majority were "white U.K. inhabitants who were not gendered stratified."
''The population under research has intrinsic limitations on the data employed in the study,'' according to Dr. Tadwalkar.
Future research should concentrate on replicating results in additional cohorts and/or with more patients, he continued, since this would not only verify the findings but also enhance risk prediction.
You must be logged in to post a comment.