AI age teller: Using chest X-ray to tell your age

The Japanese have created an uncanny AI model that can estimate your true age from the looks of your chest X-ray. It can help doctors in the early detection of chronic disorders.
Rupendra Brahambhatt
Composition on the subject of future technologies
Composition on the subject of future technologies


Have you ever wondered why some people look much older than their chronological age? A new study from Japan’s Osaka Metropolitan University (OMU) suggests this could be a sign of a disease they don’t know yet.  

The study authors have developed an AI program that can accurately calculate an individual's age by reading their chest X-ray. This model estimates age, unlike various previously reported AI programs that examine radiographs to detect lung anomalies. Then researchers use this information to predict body ailments further.

For instance, they claim that if the age of a person calculated by this AI is significantly higher than their actual age, the person is likely to have chronic conditions like COPD (chronic obstructive pulmonary disease), hypertension, or hyperuricemia (above normal uric acid levels in the blood). 

This new AI model is probably the first program that helps scientists establish a direct connection between age and chronic medical conditions. According to the researchers, it can therefore emerge as an effective tool for the early detection of chronic disorders, which is crucial for their effective treatment. 

“Chronological age is one of the most critical factors in medicine. Our results suggest that chest radiography-based apparent age may accurately reflect health conditions beyond chronological age,” said Yasuhito Mitsuyama, one of the study authors and a graduate student at OMU.

Testing the AI age teller  

The researchers fed large amounts of data to their program to develop an AI model that could accurately estimate age from chest radiographs. They collected 67,099 chest X-rays of more than 36,000 individuals from three health check-up centers. All these X-rays were taken between 2008 and 2021 and were of healthy people. 

When researchers calculated the correlation coefficient (a numerical parameter used to check the linear relation between two variables) between the estimated and actual age, it came out to be 0.95. This means that the AI-estimated ages and the precise ages of patients are strongly aligned. 

In the next phase, they tested their AI model using another 34,197 chest X-rays collected from two medical centers. However, this time all the radiographs belonged to individuals who suffered from known medical conditions. 

When the AI model calculated the age by looking at these X-rays, Mitsuyama and his team noticed a surprising positive correlation between the AI-estimated age and the possibility of chronic conditions. 

They observed that when AI calculated a person’s age higher than their chronological age, it mostly turned out that the same individual also has ongoing health conditions like high blood pressure or pulmonary disorders like chronic bronchitis. 

“The higher the AI-estimated age compared to the chronological age, the more likely individuals were to have these diseases,” the study authors note

The researchers hope that with further research and development, they can employ their AI model for more advanced tasks such as predicting a person’s lifespan, calculating their chances of survival from a particular disease, and suggesting the best treatment for their various medical conditions.   

The study is published in the journal The Lancet Healthy Longevity.

Add Interesting Engineering to your Google News feed.
Add Interesting Engineering to your Google News feed.
message circleSHOW COMMENT (1)chevron
Job Board