Could a SELFIE predict whether you’ll survive cancer? Scientists develop AI tool that can spot patients at greatest risk of dying simply by looking at their faces

A simple selfie could soon help doctors predict whether a cancer patient will survive.

Scientists have developed an artificial intelligence (AI) tool that can spot patients at greatest risk of dying – simply by analysing a photograph of their face.

In tests, more than half of those identified as being at highest risk died within 90 days.

By contrast, just two per cent of those in the lowest-risk group died over the same period.

Experts believe the technology could help doctors decide which patients are well enough for aggressive cancer treatment, and which might need a gentler approach.

The tool, called FaceSurvival, was developed by US researchers at Mass General Brigham.

It builds on earlier research showing that cancer patients who look older than their actual age are more likely to die.

But rather than estimating a patient’s biological age, the new AI looks for subtle facial clues that could reveal their chances of survival.

The findings, published in The Lancet Digital Health, could transform how doctors assess patients before treatment.

Doctors already make quick judgements about a patient’s health simply by looking at them.

Known as the ‘eyeball test’, this helps them assess how frail someone appears.

‘Clinicians already intuitively assess patient health based on appearance,’ said study co-author Dr Grace Lee, an expert in colon and rectal surgery.

‘Our tool provides a way to standardise and quantify this intuitive practice, potentially enhancing prognostic accuracy and clinical decision making.’

To develop the technology, researchers trained an AI system using more than 40 million publicly available facial photographs.

They then taught FaceSurvival to predict mortality using pictures of more than 34,000 cancer patients.

When tested on a separate group of 3,820 patients, the tool gave each a risk score between zero and one.

Of the 122 patients given the highest scores, 64 died within three months.

That compared with just 20 of the 975 patients in the lowest-risk group.

Overall, those in the highest-risk quarter were more than three times as likely to die as those in the lowest-risk quarter, even after accounting for age, sex and cancer type.

The findings were also tested in nearly 4,900 cancer patients in the Netherlands.

The scientists also developed a second tool, called FAHR-FaceAge, which estimates a person’s biological age from their face.

Unlike chronological age, biological age reflects how much the body has aged.

In earlier research, the team demonstrated the idea using photographs of two Hollywood actors, both taken when they were 50.

Paul Rudd, now 57, star of Ant-Man and The 40-Year-Old Virgin, was estimated to have a biological age of just 42.6.

By contrast, Wilford Brimley, who starred in the 1985 film Cocoon and died aged 85 in 2020, was estimated to have a biological age of 69.

In the latest study, the researchers tested their new biological-age tool on more than 38,000 cancer patients.

Those whose faces appeared at least five years older than their actual age had a 21 per cent higher risk of death.

Those who looked at least five years younger had a 17 per cent lower risk.

Smoking, heavy drinking and drug use were also linked to an older-looking face.

Smoking showed the strongest association, with current smokers tending to look biologically older than former smokers and those who had never smoked.

The researchers believe the two tools could eventually be used together to help doctors assess patients more quickly and cheaply.

However, the technology remains experimental and more studies are needed before it can be used in hospitals.

Scientists are also investigating whether cosmetic surgery affects its accuracy.

And while the AI can identify patients at greater risk of dying, it is not yet known whether using it to guide treatment would improve survival.