Andrew Davies
Digital Health Lead, ABHI
Artificial intelligence (AI) excels at automatically recognising complex patterns and providing quantitative assessments of radiographic characteristics, which is improving the interpretation of medical imaging.
Medical imaging (X-rays, CT scans, MRIs and ultrasounds) plays a crucial role in diagnosing and monitoring a wide range of medical conditions. AI can provide quantitative, rather than qualitative, assessments and has shown tremendous potential in improving the accuracy and efficiency of interpreting these images.
Regulatory approvals for AI in radiology
Regulatory authorities have been actively working on frameworks to evaluate and approve AI-based medical devices. For example, in 2018, the Food and Drug Administration (FDA) cleared the first AI system — a machine learning algorithm — to be used for the detection of diabetic retinopathy in retinal images. Up to 2022, over 520 AI tools have been approved by the FDA, of which nearly 400 are in radiology.
Improving diagnosis efficiency and workflow
One of the most significant benefits of AI in radiology is its ability to enhance diagnostic accuracy by identifying subtle patterns and anomalies in medical images. AI doesn’t just improve accuracy; it also enhances radiology workflow. It can quickly triage and prioritise cases, allowing radiologists to focus on more complex diagnoses.
It can quickly triage and prioritise cases; allowing radiologists to focus on more complex diagnoses.
Challenges and ethical concerns
Despite the promise of AI in radiology, there are several challenges to overcome. Ensuring the privacy and security of patient data is a top priority. Additionally, the need for proper validation and regulatory oversight is essential to prevent errors and biases in AI algorithms. Addressing these concerns is vital for the responsible implementation of AI in radiology.
AI to become ‘business as usual’
Radiology is leading the field in the use of AI in clinical applications. In the UK, we are building the structures and processes that will enable us to move it from pilot to business as usual. The NHS AI lab, working alongside the Royal College of Radiologists, has established a Diagnostic Fund and is piloting an AI Deployment Platform that will make implementation more efficient and scalable.
Success for AI in imaging will be measured by increased diagnostic certainty, system efficiency, impact on workforce and patient care. Radiology can be the lead use case for deploying AI into routine clinical practice; and as research continues and technology evolves, AI is expected to play an increasingly vital role in helping radiologists detect and diagnose diseases.
