Artificial intelligence (AI) may be able to help radiologists review tens of thousands of mammograms each year, according to a trial being conducted at Aberdeen Royal Infirmary.
June, a healthcare assistant and trial participant, is scheduled to have surgery as a result of the pilot's assistance in detecting early-stage breast cancer.
Mammograms are low-level X-rays that are used in breast cancer screenings to track and find changes that are too small to feel or see.
In the UK, they assist in saving about 1,300 lives a year, according to the NHS.
In Scotland, the number of women who attended a routine breast exam following an invitation rose over the three years leading up to 2022, but fewer radiologists are available to review the results.
The use of artificial intelligence (AI), a technology that allows computers to carry out specific tasks that would typically require human intelligence, is already pervasive in a number of different industries.
While prominent experts' concerns that AI could wipe out humanity have recently made headlines, the more practical applications of the technology are already being demonstrated in the field of healthcare.
Many scientists and physicians view AI as a powerful tool to work alongside practitioners rather than replace them because of its potential to accelerate the process of discovering new drugs and treatments for diseases.
Data from Public Health Scotland (PHS) show that between 2021 and 2022, there were 1,830 new cases of breast cancer among women of all ages in Scotland.
Of the 5,000 mammograms that radiologists examine on average each year, between 250 and 300 patients will be called back, with 30 to 40 of them needing more in-depth care.
Dr. Gerald Lip, clinical director of the North East Scotland Breast Screening Programme, told BBC Click that there was a possibility that with that number, cancers might go undetected.
Gemini, a project by NHS Grampian in partnership with the University of Aberdeen and private industry, was mentioned in Scotland's AI Strategy when it was first published in 2021.
The trial's AI model Mia was created by Kheiron Medical Technologies, and Microsoft provided the cloud computing services that made it possible.
Dr. Lip and other radiologists are experimenting with AI as an additional check at the end of mammogram scan reviews because the National Screening Counsel's rules currently prohibit its automatic deployment in screenings.
After Dr. Lip described how the AI tool helped identify a problem area, June, a trial participant who has undergone similar surgery before, had a biopsy performed to remove a small portion of her breast tissue for testing.
Using anonymized mammogram results, he demonstrated to BBC Click how the software functions.
We're currently observing a woman who had mammograms on both her left and right sides. "You're looking for differences," he said.
Radiologists can view and examine deviations between the two scans detected by the AI by pressing a button.
Dr. Lip indicated the primary area of concern by pointing to one area that the AI program had circled.
"In screening, you want to catch things early on before they grow large," he continued.
June revealed to Click a few weeks after her biopsy that the use of AI, as opposed to another human pair of eyes, made the procedure feel less intrusive.
She said, "You know your pictures are on the screen and people are looking at them. "In contrast, the slight sense that someone is looking has vanished when it's an artificial intelligence. ".
June will require surgery once more as a result of the biopsy's findings.
June stated, "They've definitely caught it at an earlier stage this time. The biopsy showed that I do have an early-stage cancer.". But I'm going to have a mastectomy because I've had experience with it before.
"I don't want this treatment. It's encouraging that it's being caught, though, at the same time. ".
Numerous radiology and advance practitioner staff members may be approaching or have reached retirement age, according to a significant review of Scotland's breast screening program that was published last year.
It warned that the loss of "super reader" radiologists, who are needed by the service to review a large number of results, could make it "vulnerable.".
Patients are having to wait too long for necessary tests and cancer treatments, according to a warning from the Royal College of Radiologists that the UK is experiencing "chronic staff shortages.".
The Scottish government's review found that using AI to replace one human reader could "cover half of the screening image reading burden," which amounts to about 1 point 72 million images read annually. Two radiologists are required to read the images and report the results.
Can the technology being tested in Aberdeen eventually take the place of employees?
"I think the goal of this evaluation is to see what's the best way we can work with AI," said Dr. Lip. "Whether it's replacing one of the radiologists, whether it's part reading some of the normal mammograms, or whether it's improving our cancer detection as a safety net.".
According to Peter Kecskemethy, co-founder of Kheiron, more than 30 NHS trusts in the UK will use the technology, making it available to millions of people.
Furthermore, with trusts in England already looking into how AI can support better, quicker outcomes for breast cancer patients, the technology seems to be continuing to play a significant role in aiding physicians in saving lives.
See the full report and more examples of how AI is transforming various industries. on this week's Click episode.
