AI-Enabled Analysis of Images Meant to Catch One Disease Can Reveal Others

04.12.24 22:00 Uhr

NEW YORK, Dec. 4, 2024 /PRNewswire/ -- With the help of an AI tool, computed tomography (CT) scans taken originally to look for tumors or bleeding or infections, also revealed calcium buildup in arteries, a sign of worsening cardiovascular disease.

NYU Langone Health (PRNewsfoto/NYU Langone Health)

This is the result of a new study led by researchers at NYU Langone Health and an example of a new trend in "opportunistic screening," wherein radiologists repurpose existing medical images to diagnose illnesses beyond what the scan was originally designed to find.

Presented at the annual Radiological Society of North America (RSNA) meeting in Chicago on Dec. 4, the study featured a reanalysis of a large batch of scans of the abdomen, done for many reasons, to analyze a portion of the aorta, the major artery that runs from the heart through part of the abdomen. Using that data from commonplace scans, the study authors then used AI to measure the amount of aortic calcium, attach a standard score to the calcification level, and use it to predict a person's risk of having a major cardiovascular event, including a vessel blockage (heart attack).  

"Instead of relying on dedicated CT scans of coronary arteries, which are rare and not always covered by insurance, to find potentially fatal heart disease, we seek to use AI to help screen abdominal CT scans that are done for many reasons to opportunistically catch heart disease more often and earlier," said study senior investigator Miriam Bredella, MD, MBA. Bredella is the Bernard and Irene Schwartz Professor of Radiology and director of the Clinical and Translational Science Institute at NYU Grossman School of Medicine.

Specifically, researchers looked back at 3,662 CT scans between 2013 and 2023 of mostly older men and women in the New York area, where the same patients had both an abdominal scan (which captured part of the aorta) and a dedicated CT scan of their coronary arteries.

The researchers found that the AI-enabled measurement of calcification amounts in the aorta in abdominal scans done for other reasons enabled the team to accurately predict both calcification in the coronary arteries of the same person, and that person's risk of major cardiovascular events. This result, say the authors, suggested that the abdominal scan alone could be used to predict heart attacks or other cardiovascular events.

Those with aortic artery calcification were 2.2 times more likely after three years of monitoring to suffer a major heart attack or brain vessel blockage or need to undergo procedures to restore blood flow to the heart, which did indeed happen to 324 study participants, say the study authors. The study also showed early indications of arterial calcium buildup in 29% of study participants previously thought to have had none.

The new finding supports a previous study result published in September in the journal Bone about the use of opportunistic screening in diagnosing bone loss, also called osteoporosis.

In the earlier study, also with the aid of a fully automated AI algorithm, Bredella and her team from Massachusetts General Hospital and Harvard Medical School, performed a secondary analysis of CT scans done as part of screening for lung cancer in 3,708 patients, mostly older current and former smokers. By analyzing scans done to examine the lungs, but that also capture images of nearby bones, the researchers found severe signs of bone loss across all races and incomes in both men and women.

Osteoporosis, a disease underdiagnosed in both the general population but especially in racial minority groups, the research team reported, was present in 38 percent of Blacks, 55 percent of Asians, 56 percent of Hispanics, and 72 of Whites screened. Also detected by the opportunistic screening tool were high ratios of body fat, arterial hardening, and fatty liver, all of which are tied to bone loss.

"Our research demonstrates that opportunistic screening could help with diagnosing and treating osteoporosis in vulnerable groups who are at greater risk of the disease, in particular, the elderly and those who smoke," said Bredella. "This work establishes the foundation for using opportunistic screening to address the lack of access to osteoporosis and heart disease prevention, as well as to screening for cancer and diabetes."

More research, though, she notes, is needed to determine if the imaging data and analysis provide sufficient early identification of those at greater risk of major coronary disease or osteoporosis for treatment to prove effective at reducing illness and death.

Funding support for the study on aortic calcification was provided by National Institutes of Health grants UL1TR001445, R35HL144993, R01AG065330, and R01LM013344. Funding support for the study on osteoporosis was provided by National Institutes of Health grant K24DK109940.

Besides Bredella, other NYU Langone researchers involved in the study presented at RSNA are co-investigator Jeffrey Berger, MD; Soterios Gyftopoulos, MD, MBA, MSc; Bari Dane, MD; Eduardo Iturrate, MD; Michael Recht, MD; and Judy Zhong, PhD. Another study co-investigator is Malte Westerhoff, at Visage Imaging GmbH in Berlin, Germany.

Other co-investigators involved in the osteoporosis study are Florian Huber, MD; Katherine Bunnell; and Efren Flores, MD, at Massachusetts General Hospital and Harvard Medical School in Boston; Perry Pickhardt, MD, at the University of Wisconsin in Madison; and Ronald Summers, MD, PhD, at the National Institutes of Health Clinical Center in Bethesda, Md.

Note: This presentation (abstract #W7-SSCA08-5) at the annual meeting of the Radiological Society of North America (RSNA), Dec. 4, 2024, at 3 p.m. CST in Chicago is titled "Opportunistic Assessment of Aortic Artery Calcification Using Artificial Intelligence (AI) and Its Association with Coronary Artery Calcification and Cardiovascular Events."

Video animation about the work of study researcher Miriam Bredella, MD, MBA, is also available at: https://www.youtube.com/watch?v=MAQvZvHtwuo

Media Inquiries
David March
212-404-3528
David.March@nyulangone.org 

Cision View original content to download multimedia:https://www.prnewswire.com/news-releases/ai-enabled-analysis-of-images-meant-to-catch-one-disease-can-reveal-others-302317706.html

SOURCE NYU Langone Health System