AXREM is the UK trade association representing the interests of suppliers of diagnostic medical imaging, radiotherapy, healthcare IT and care equipment in the UK. The group has formed the AXREM AI SFG (special focus group) to promote the adoption of beneficial AI technology and to identify and seek to overcome hurdles to adoption encountered by suppliers.
Most recent AI solutions are based on a process of deep learning in which an artificial neural network is presented with annotated examples of a particular kind of pathology and gradually learns to recognise what that pathology looks like. This process is data dependent, and so the work of data scientists and curators is as important as the AI scientists and software engineers who develop the actual algorithms.
A key part of developing safe and reliable AI solutions is the use of a representative dataset while performing the AI training. “There are well publicised examples where bias has been introduced during the training process leading to highly inaccurate results,” AXREM states. “The answer to this is to use data drawn from the target population, which means a generic AI solution needs to be trained on a representative subset of the entire population. This objective seems daunting, but it can be achieved using a network of ‘trusted research environments’, coupled to a technology called federated learning that can facilitate building a generic model safely and effectively.
“The potential gains that clinicians and managers could expect from applying AI to the process of triaging a backlog of unread x-ray images to prioritise the work of the radiologist are real, and the technology has recently matured to a level at which that is a realistic task for AI. However, it is not realistic to think that AI will ‘replace radiologists’ any time soon as some disruptors in the field have claimed,” AXREM continues.
There are many examples emerging where AI could speed up radiologists in their day-to-day work, allowing them to spend more time where their expertise is needed most. In a research environment, it is now possible to use AI in clinical trials to accelerate the process of measuring tumour volumes during a drug trial for a potential cancer cure to track changes in pathology.
The AXREM AI SFG is committed to engaging with relevant stakeholders in the sector to ensure that when AI is adopted, it is used appropriately and effectively.
Picture: Representative data must be used when performing AI training.
Published on page 3 of the June 2022 issue of RAD Magazine.