Japanese AI Tool Predicts Health Age

River D'Almeida, Ph.D
3 min readAug 22, 2023

It also detects chronic disease risks from chest scans

Photo by Alexander Sinn on Unsplash

In a groundbreaking development, researchers at Osaka Metropolitan University in Japan have harnessed the power of artificial intelligence to predict a person’s chronological age and shed light on potential risks for chronic diseases through the examination of chest radiographs. This innovative approach not only holds the potential to redefine health assessment but also presents a significant stride towards personalized medicine. The team’s findings have been documented in a recent article published in The Lancet Healthy Longevity, marking a pivotal moment in the field of healthcare.

Accurate Chronological Age Prediction

The concept of chronological age, denoting a person’s actual years on this planet, is a fundamental parameter in health evaluation. However, the innovative tool devised by the Japanese researchers takes this assessment to new heights. The technology utilizes a sophisticated deep learning network, meticulously trained on a diverse dataset comprising chest radiographs collected from numerous institutions. With data from approximately 67,000 individuals, this network demonstrates remarkable accuracy in deciphering an individual’s chronological age from the intricate details present in chest scans.

The potency of this AI tool becomes evident through its staggering correlation coefficient of 0.95 when cross-referencing its predictions with actual chronological age. This high degree of correlation underscores the tool’s proficiency in precisely gauging a person’s age solely based on chest radiographs. This level of accuracy hints at a wealth of potential applications, from age verification in various settings to streamlining age-related medical assessments.

Potential Biomarker for Chronic Diseases

However, the innovation extends beyond the realm of age prediction. The researchers delved deeper into their dataset, comprising around 37,000 scans of individuals with documented chronic diseases. This exploration uncovered a remarkable revelation — a positive relationship between the disparities in an individual’s predicted age by the AI and their actual chronological age. Intriguingly, these disparities served as a…



River D'Almeida, Ph.D

Follow me for bite-sized stories on the latest discoveries and innovations in biomedical research.