E183 ‘Radiology’s AI Revolution’ with The Mater Hospital’s Peter MacMahon

Today’s guest is Peter MacMahon, Consultant Radiologist at the Mater Hospital. Following the Mater Hospital’s nomination in the Best Application of AI in Healthcare category at the 2024 AI Awards, Peter joins the show to discuss the integration of AI into radiology and healthcare at large. This conversation provides a window into how AI is being used to enhance patient care.

Peter offers a first-hand account of how the transition from analog to digital images in radiology was a significant transformation, paving the way for the integration of AI. As radiologists transitioned from film to digital data, it opened new opportunities for analysis and image processing. The wealth of digital information accumulated over the past decades has laid the groundwork for AI tools that assist radiologists in interpreting complex images more efficiently and accurately.

AI’s transformative journey is depicted vividly through the experience of adopting the software from California-based company Rapid, which began assisting radiologists with complex data analysis as early as 2015. With AI handling the intricate task of analysing perfusion data, physicians were able to make quicker, more accurate decisions in time-critical cases such as strokes.

Since then, the Mater Hospital has experimented with various AI tools, including those developed by companies like ADOC and Gleamer. These tools flag possible emergency conditions like brain haemorrhages and fractures, serving as reliable ‘wingmen’ for human radiologists by providing additional data points and ensuring that nothing is missed, especially during high-pressure scenarios.

The integration of AI into the Mater’s clinical workflows has shown promising results. It provides a safety net for human error, and while AI is not infallible, its consistent accuracy helps reduce diagnostic errors. For instance, AI routinely analyses CT scans and flags cases needing urgent attention, allowing radiologists to act swiftly and confidently even in stressful environments. Peter emphasises that AI currently serves as a co-pilot, augmenting human decisions rather than replacing them.

While the current use of AI in healthcare helps improve diagnostic precision, long-term studies are underway to evaluate its impact on actual patient outcomes. Peter and his team are keen to explore whether AI tools could streamline patient care processes – such as reduced waiting times or more rapid initiation of treatment – thereby justifying their broader implementation across healthcare systems.

However, the journey to integrating AI in healthcare is fraught with challenges. Data security concerns, especially in the aftermath of cyberattacks, require painstaking efforts to maintain patient privacy while utilising cloud computing resources. Healthcare providers must navigate these technological and regulatory landscapes carefully to ensure secure and efficient adoption of AI.

Peter’s insights into the use of AI in healthcare illustrate how AI tools can serve as vital allies in diagnostics. With ongoing efforts to evaluate their impact on patient outcomes and address integration challenges, AI stands poised to revolutionise healthcare further. To stay ahead in this rapidly evolving field, healthcare professionals must embrace innovation while keeping ethical considerations at the forefront.