Having older AI-estimated heart age than actual age is linked with cardiovascular/dementia-related death

Key takeaways:

• An artificial intelligence (AI) model was developed that could estimate heart age from ultrasound images.
• Researchers used the AI model to estimate the heart age of 568 people aged 50 years or older from a general population cohort.
• Having an AI-estimated heart age older than the person's actual age was associated with higher cardiovascular- and dementia-related mortality over 20 years.
• Prospective studies are warranted to assess whether AI-based models of heart age may be able to identify individuals at high risk or monitor the effects of treatment.

Basel, Switzerland - 13 November 2026: Having an older heart age estimated using artificial intelligence (AI) ultrasound analysis compared with actual age was associated with worse outcomes, according to results presented today at the ESC Digital & AI Summit 2026 [1].

Deep learning AI models can be used to estimate heart age from ultrasound (echocardiography) images. "Previously, it has been shown that a discrepancy between AI-estimated heart age and actual age is linked with worse outcomes in patients with cardiovascular disease [2], but there is a gap in the evidence regarding whether this relationship exists for the general population," said study presenter, Doctor Jeffrey G. Malins from Mayo Clinic, Rochester, USA. "We conducted an analysis where we calculated the cardiac age gap between the AI estimation and the individual's actual age in a community-based cohort and evaluated this against long-term data on death from cardiovascular or dementia-related causes," he explained.

This retrospective analysis used data from a community-based cohort of 568 adults aged 50 years or older who underwent echocardiography between 2002 and 2004. A previously developed AI-age model [3] was applied to echocardiography clips of the heart and an average AI-predicted age was obtained for each participant. The echocardiographic cardiac age gap was calculated as AI-predicted (biological) age minus chronological age. Data on cardiovascular- and dementia-related mortality were collected from the cohort over a 20-year period.

Each five-year increase in echocardiographic cardiac age gap was associated with a 61% increase in 20-year cardiovascular/dementia-related mortality (hazard ratio [HR] 1.61; 95% confidence interval [CI] 1.24-2.07; p<0.001), when adjusted for the participants' age and sex.

When adjusted for age, sex, body mass index and left ventricle ejection fraction, the association was mildly attenuated but was still significant: each five-year increase in echocardiographic cardiac age gap was associated with a 46% increase in 20-year cardiovascular/dementia-related mortality (HR 1.46; 95% CI 1.13-1.90; p=0.004).

"Results from a community-based cohort indicate that AI-derived cardiac aging signatures may serve as a marker to identify people with long-term vulnerability," said Doctor Malins. "These types of analyses may also be useful to study if lifestyle or medical interventions can reduce the cardiac age gap," he noted.

Cardiovascular disease is increasingly recognised as a contributor to dementia-related mortality and aging-related decline. "We evaluated cardiovascular- and dementia-related mortality together for practical reasons in our study but it would be useful to perform separate AI age analyses to provide insights into the relationship between cardiovascular and neurodegenerative diseases," explained Doctor Malins. "Wider prospective evaluations of AI-derived cardiac aging signatures in larger, independent cohorts are now warranted," he concluded.

ENDS