Discover award-winning AI at ECR 2024
Discover award-winning AI at ECR 2024

5 Key Takeaways on How to Leverage AI for Patient-Centric Public Healthcare

In a webinar run by the Royal Australian and New Zealand College of Radiologists, three leading radiologists shared their insights and experience with AI to support resilient and patient-focused public healthcare.

Here are five key takeaways from the presentations by:

  • Dr Justin Hegarty – Radiologist at Pacific Radiology and Te Whatu Ora Canterbury, New Zealand.
  • Dr Geraldine Dean – Consultant radiologist and AI lead for NHS Southwest London.
  • Professor Catherine Jones – Professor of Clinical Imaging Science at the University of Sydney
#1 To stay relevant, you must be ready

There is a pressing need to adopt clinical AI, even with the current limitations and obstacles. To preserve their professional autonomy, radiologists must position themselves as AI stewards. Rather than allowing other specialties or administrators to take the lead, radiologists should aim to build relationships with AI tools and play a crucial role in shaping their integration.

#2 Embrace AI as a collaborative tool, not a competitor

AI is already incorporating non-image inputs such as clinical history and investigation results. Radiologists need to adapt to this evolution, envisioning a future where AI tools become collaborators, not replacements, necessitating a shift in practice.

Peer-reviewed clinical research using chest X-rays has demonstrated that, in many cases, the radiologist’s performance improved with AI support. Radiology departments should prioritise AI integration as a matter of urgency to ensure their continued autonomy to the wider healthcare team.

#3 Start with a problem you want to solve

There are many different applications of AI across the care pathway that can add clinical value; start with a targeted clinical problem you need to solve. . Dr Dean’s AI adoption project, for example, explored whether an AI tool could help them meet the recommended timeframes of the United Kingdom’s National Optimal Lung Cancer Pathway.

Once you’ve decided on the primary problem, you can consider secondary objectives.

Dr Dean’s project team also:

  • assessed the impact of AI on reporting times
  • evaluated staff attitudes towards AI use in radiology and if this changed after real-world implementation
  • gathered a list of pain points and possible solutions to assist with future AI projects.
#4 Triage can be a great starting point

With the challenges involved with traditional worklist prioritisation, AI has significant potential to triage urgent cases more efficiently. Traditional methods – which rely on written referrals, source of patient (e.g. the emergency department), or urgency tags – often result in inefficiencies and delayed awareness of critical pathology.

AI, as a first look at imaging, can rapidly flag scans with suspected urgent findings, directing radiologists’ attention to complete time-critical reports first. This approach helps to prevent critical pathology from being overlooked or given routine priority, significantly reducing turnaround times and potentially saving lives.

#5 The advantage of comprehensive AI

A comprehensive tool – one with potential to impact patients across various stages of their medical journey – unlocks the full potential of AI in radiology. A tool’s comprehensiveness incorporates its:

  • Breadth of coverage – across modalities and findings
  • Depth of coverage – providing detailed insights into specific abnormalities
  • Reach – across diverse patient demographics and geographies.

Any solution that looks for a comprehensive range of findings can more effectively triage and prioritise worklists. For example, a chest X-ray tool that only looks for pneumothorax “is wonderful for the small percentage of patients who have a pneumothorax,” Professor Jones said. “But what about everybody else? It doesn’t help them at all.

Keen to find out more? Book a meeting with our team to see how AI can help your practice.

Seah JCY, et al. Effect of a comprehensive deep-learning model on the accuracy of chest x-ray interpretation by radiologists: a retrospective, multireader multicase study. Lancet Digit Health. 2021 Aug;3(8):e496-e506. doi: 10.1016/S2589-7500(21)00106-0. Epub 2021 Jul 1. PMID: 34219054.

*FDA website as of March 2023

A comprehensive decision-support tool. Powered by AI. Visit Web demo


Warning: Undefined array key "country_code" in /www/annalise_914/public/wp-content/themes/annalise/footer.php on line 154