LONDON – The Royal College of Radiologists is right to call for the medical imaging community to become actively involved in assessing the accuracy and credibility of AI algorithms, according to the head of one of the UK’s leading AI radiology companies.
Simon Rasalingham, CEO and Chairman of behold.ai, was responding to a warning by outgoing RCR President Dr. Nicola Strickland about ensuring that AI services coming onto the market were safe and fit for purpose. He welcomed the fact that the core curriculum of the Fellowship of the RCR has been recently rewritten to include the basic concepts of AI.
“I haven’t met a clinical or commercial person involved in the development and implementation of AI who doesn’t understand that the consultant radiologist is central to this process. The more the radiologist is involved in testing the efficacy and appropriate deployment of this technology the better it will be for the patient and for healthcare services generally”, he said.
It was pleasing that Dr. Strickland had pointed to the triage of worklists on PACS to prioritise automatically detected “urgent” cases as having promising potential.
“We are working on this very application and have a number of pilots either running or about to start in NHS hospitals – all of these integrations are being introduced under the close scrutiny of senior radiology personnel in each hospital”, he pointed out.
But Mr Rasalingham said it was important that, while taking a sceptical and quality-led stance, the RCR and other professional bodies should look to positively embrace AI as a game-changer in global radiology.
“Dr. Strickland said she thought that some countries have a tendency to try to get AI-related products into clinical practice as soon as possible, due to workload-related issues, particularly radiologist numbers versus escalating patient exams.
“Well, that is a perfect description of the UK! We have declining numbers of radiologists in this country, ever-increasing requests for scans and a new power in the land, the NHSX, determined to lever the benefits of digital technology for the benefit of patients.
“The pace of AI development in global radiology is astonishing and is bringing unique challenges. We have just been awarded ISO 13485 QMS certification – the gold standard of quality certification – but because of the advanced nature of our algorithm a lot of creative abstraction of concepts and ideas was required in order to build a QMS that met the necessary requirements.
“The audit itself was 7 man-days long and no area of the business was left unexamined. Developing AI applications for safe, verified and efficient use is not to be undertaken lightly whether it’s a company or a College”.
He added that the RCR could be proud if its members who were piloting the introduction of behold’s ‘red dot’ algorithm for instant triage radiology in NHS hospitals.
“They are highly-professional, incredibly thorough and completely committed to obtaining the most accurate and timely diagnoses for their patients. They are a credit to the College and a credit to radiology”.
Further information from John V. Wright on +44 7710 454214 Email: firstname.lastname@example.org
NOTES FOR EDITORS
Behold.ai is headquartered in London and has corporate and research facilities in New York and India.
We use cutting-edge science to pioneer new forms of diagnostic medicine; this innovative approach to healthcare is a step change in the treatment and prevention of diseases, delivering benefits today to patients across the NHS. The company is identified as a leader in AI image-based diagnosis in the influential Artificial Intelligence in Global Health: Defining a Collective Path Forward report produced by the Rockefeller Foundation and United States Agency for International Development’s (USAID) Center for Innovation and Impact (CII), in coordination with the Bill and Melinda Gates Foundation.
Behold.ai provides artificial intelligence, through its red dot® cognitive computing platform, to radiology departments. Behold.ai’s technology augments the expertise of radiologists to enable them to report with greater clinical accuracy, faster and more safely than they could before. This revolutionary combination helps to deliver greater performance in radiology reporting at a fraction of the price of outsourced reporting.