Heart roundness may be early marker of atrial fibrillation, cardiomyopathy
Click Here to Manage Email Alerts
Key takeaways:
- An AI algorithm revealed heart roundness seen on cardiac MRI may predict risk for atrial fibrillation and heart failure.
- Researchers identified four loci associated with sphericity at genome-wide significance.
Healthy adults with spherical hearts, identified through a deep learning analysis of cardiac MRI images, were 31% more likely to develop atrial fibrillation and 24% more likely to develop cardiomyopathy, researchers reported.
In an analysis of more than 30,000 cardiac MRIs using an artificial intelligence algorithm, researchers found that sphericity may predict risk for certain CV conditions, even when heart size and function are normal.
“What clinicians have known is the heart gets rounder to compensate for injury or stress, for example after a heart attack,” David Ouyang, MD, a cardiologist and researcher at the Smidt Heart Institute at Cedars-Sinai, told Healio. “We wanted to quantify this with AI and then use that information to associate it with both genetics and CV outcomes.”
Ouyang and colleagues measured the left ventricle sphericity index (short axis length/long axis length) using deep learning-enabled image segmentation of cardiac MRI data for 38,897 participants from the UK Biobank. Researchers excluded participants with abnormal LV size or systolic function. Researchers assessed the relationship between LV sphericity and cardiomyopathy using Cox analyses, genome-wide association studies and two-sample Mendelian randomization.
The findings were published in Med.
Researchers found that a 1 standard deviation increase in sphericity index is associated with a 47% increased incidence of cardiomyopathy (HR = 1.47; 95% CI, 1.1-1.98; P = .01) and a 20% increased incidence of AF (HR = 1.2; 95% CI, 1.11-1.28; P < .001), independent of clinical factors and traditional MRI measurements.
“We used AI to quantify sphericity, essentially how round the heart is,” Ouyang said in an interview. “We saw that it was related to a history of or potentially developing HF and a history of developing AF in the future.”
The researchers also identified four loci associated with sphericity at genome-wide significance: PLN, ANGPT1, PDZRN3 and HLA DR/DQ.
The researchers noted that sphericity index may not fully capture phenotypic variation within the LV, as additional features simultaneously influence cardiac function.
Ouyang said the study demonstrates the utility of using deep learning and advanced imaging analysis to define nontraditional cardiac imaging risk biomarkers using large-scale data.
“The phrase ‘a picture is worth a thousand words’ is true,” Ouyang said. “There is so much more information in a picture than in a paragraph or even a page of text. It is the role of the scientist to understand all of those relationships, using AI to find other ways where we can predict risk factors for disease.”
Ouyang said researchers are currently reviewing other measurements that could be automated through AI, particularly related to the right ventricle.
For more information:
David Ouyang, MD, can be reached at david.ouyang@cshs.org; Twitter: @david_ouyang.
Perspective
Back to Top
Sean Pinney, MD, FACC, FAST, FHFSA
There is a great need to be able to identify patients at risk for developing HF. The strength of combining noninvasive imaging techniques with AI is an enhanced ability to identify at-risk populations that currently escape our clinical detection. Of course, once we are able to identify these populations, appropriate interventional trials will need to be designed to help elucidate treatments to prevent or forestall the development of clinical HF.
Collapse
Read more about
Click Here to Manage Email Alerts