New study finds deep learning model predicts heart, lung, and frailty changes more accurately than DNA-based aging clocks

Artificial intelligence may be able to reveal how fast your body is aging by analyzing a chest X-ray, according to a new study published in The Journals of Gerontology. Researchers found that a deep learning model was able to detect subtle, age-related changes in the heart, lungs, and overall health more effectively than leading DNA-based “epigenetic clocks.”

The study, entitled “Deep learning chest X-ray age, epigenetic aging clocks, and associations with age-related subclinical disease in the Project Baseline Health Study,” compared the AI model — known as CXR-Age — to two widely used biological age measures derived from DNA methylation: Horvath Age and DNAm PhenoAge. Researchers analyzed data from 2,097 adults participating in the Project Baseline Health Study, a multi-site U.S. research initiative designed to better understand health and disease over time.

CXR-Age showed strong associations with early signs of heart and lung aging, including coronary calcium, worsening lung function, greater frailty, and elevated levels of proteins linked to neuroinflammation and aging. By contrast, the DNA-based clocks showed weaker or no associations — especially among middle-aged adults.

“These findings suggest that deep learning applied to common medical images can reveal how our organs are aging — information that might one day help clinicians identify people at risk of age-related disease before symptoms develop,” said Douglas P. Kiel, MD, MPH, director of the Musculoskeletal Research Center at the Hinda and Arthur Marcus Institute for Aging Research, a co-author of the study. “AI tools like this could become an important complement to traditional risk assessments.”

The researchers concluded that AI-derived CXR-Age may serve as a better indicator of cardiopulmonary aging than existing epigenetic aging clocks, highlighting the potential of medical imaging and machine learning to advance personalized, preventive medicine.