Saturday, February 24, 2024

Stockholm hospital to make use of AI to estimate breast most cancers danger, in world first – Euractiv

A brand new AI-based mannequin that may predict ladies’s breast most cancers danger will now be examined in scientific follow at a hospital in Sweden.

The brand new examine, referred to as SMART, will begin in April subsequent yr and contain 70,000 ladies between the ages of 40 and 74 present process breast most cancers screenings on the Södersjukhuses, one of many largest emergency care hospitals in Stockholm.

“It is going to be fully distinctive, as AI-based danger fashions in breast most cancers screenings haven’t but been examined in scientific settings globally,” Per Corridor, a professor of radiation epidemiology at Sweden’s Karolinska Institutet, advised Euractiv.

Of the contributors, 35,000 will likely be screened each two years, in keeping with the same old interval. The opposite group of 35,000 ladies will even be screened the identical approach however will even be supplied danger assessments utilizing the AI mannequin.

“The ladies who we discover to have at the very least 3 times the danger of breast most cancers in comparison with these with a basic danger will likely be invited to endure screenings with contrast-enhanced mammography,” explains Per Corridor.

Tailoring screening and remedy

At the moment, many established breast most cancers danger fashions are primarily based on way of life and genetic elements akin to household historical past, variety of kids, period of breastfeeding, smoking and alcohol habits, and BMI.

However for the previous eight years or so, he and his analysis group have been engaged on AI-based fashions to foretell particular person breast most cancers danger, with the purpose of tailoring screening and remedy for girls primarily based on their particular person breast most cancers danger.

“It’s a bit unusual that danger isn’t actually taken into consideration in at present’s basic mammography programmes as a result of some ladies have a excessive danger of breast most cancers and others have a low danger,” he mentioned.

The forthcoming examine will construct on new outcomes of a validation examine utilizing an image-based AI-risk mannequin developed by the Swedish postdoctoral researcher Mikael Eriksson, a member of Per Corridor’s group on the Swedish college.

In Eriksson’s examine, the AI element was tasked with calculating breast most cancers danger from mammography photographs of greater than 8,500 ladies in Spain, Italy and Germany, aged 45-69 years, who underwent breast most cancers screening between 2009 and 2020.

It’s skilled to search out and analyse tiny adjustments or new structural patterns in ladies’s mammography photographs.

“Hundreds of things are weighed collectively. The AI is ready to discover patterns which are individually weak as indicators however that the AI -tool can mix. Then the AI may also make an evaluation of how the breast is prone to develop sooner or later”, says Eriksson.

Initially, all ladies had adverse outcomes. Their two-year absolute danger of breast most cancers was then predicted and assessed at their follow-up examinations two years later.

For instance, the outcomes just lately revealed within the Lancet Regional Well being predicted that 4.7 per cent of the ladies had a most cancers danger seven occasions increased than the final danger group.

The AI mannequin was additionally capable of detect almost one in three clinically related stage II and better breast cancers in high-risk ladies.

What the long run could convey

The subsequent step is clinically evaluating the AI-based danger mannequin on the Södersjukhuset. This time the danger mannequin will likely be complemented by including particular person way of life danger elements to the breast most cancers danger calculation, in line with Per Corridor.

He’s optimistic in regards to the future approach of doing screenings primarily based on danger.

“What we actually hope is that we are able to discover cancers at an early stage, when the most cancers is tiny or small and subsequently simpler to remedy. We don’t not look forward to finding fewer tumours, however fewer tumours which are aggressive,” he mentioned.

[By Monica Kleja, edited by Vasiliki Angouridi]

Learn extra with Euractiv

Related Articles


Please enter your comment!
Please enter your name here

Latest Articles