27 de fev de 2019

Adoption of AI in breast cancer screening is imminent

SAÚDE 3 minutos de leitura

Artificial intelligence is booming in healthcare, especially when it comes to cancer screening. One company, named Kheiron Medical, developed a groundbreaking breast cancer algorithm, which received a CE mark mid last year. Now that it can be tested in a real clinical setting, the ‘machine’ is getting real.

Kheiron Medical Technologies’ mission is to help clinicians detect breast cancer earlier and more accurately. They developed a unique software algorithm that automates the second reading of mammograms. We spoke to Dr. Hugh Harvey, clinical director, about how artificial intelligence is shaping radiology.

What are the benefits of AI in breast radiology?

"AI will have a significant impact on the radiologist’s workflow. In the UK, there’s currently a 13% shortage of radiologists. One in five senior radiologists will retire within the next 5 years. At the same time, the volume of imaging in the NHS (the UK’s National Health Service) is skyrocketing. In breast screening particularly, which is a double-reading process in Europe, automation can relieve some of the burden on radiologists."

- Dr. Hugh Harvey


Today, mammograms are read by two different radiologists, however, the AI software can function as an independent reader. This means that in the future you’ll need just half of the resources that are required today. AI will be essential in order to help radiologists cover their heavy workload.

How does AI fit into a radiologist’s clinical workflow?

"The results from both the primary and secondary mammogram readings are recorded. The findings are then paired and the radiologist will be notified for closer inspection only when there’s a discrepancy between both results. The results from the clinical trials have been irrefutable. The detection accuracy of the algorithm exceeds all of the recommended performance standards for human radiologists."

- Dr. Hugh Harvey


Next to handling second readings, the software can also function as a triage system. Thanks to the algorithm, critical cases that require urgent attention can be prioritized. Thirdly, the software runs as a quality-control mechanism, which allows departments to analyze their historic performance against a consistent standard. Needless to say, this will have a vital impact on health outcomes for patients.

What is Barco’s role in all this?

Barco provided the mammography displays (Coronis Uniti®) which we use to show how the machine-learning algorithm thinks. Especially the SpotView™ tool is particularly helpful because it can highlight specific areas on a mammogram. Because of the high quality of the displays and the spotlight, you can map out subtle image details on the exact pixel level. We’re always using Barco mammography displays at our conferences to educate radiologists about the capabilities of AI in breast screening.

Next steps: evidence and trust

We are now collecting data from clinical trials at test bed sites to prove that AI works in a clinical setting. We are working with government bodies in the UK to raise awareness and acceptance among radiologists and ensure engagement with the public.

I believe we’re just a just a year or so away from our AI technology being adopted across the board by the NHS. By then, we will have been able to collect all of the outcomes data from clinical trials and there will be evidence within the NHS that AI really works. It will be game-changing. This next-gen technology has the ability to improve healthcare on a global scale.


At Barco, we encourage medical research, especially in the field of cancer detection. For this AI test case, Kheiron Medical relies on Barco’s Coronis Uniti® displays and clinical tools to provide doctors with accurate insights related to artificial intelligence in breast screening. Because reliable and detailed visual information is vital to effective decision-making.