How Artificial Intelligence is Being Used in Lung Imaging, with Rachel Eddy, PhD

The medical world has been steadily transformed by the use of artificial intelligence (AI) and machine learning, with some experts in fields such as pulmonology having uncovered benefits from AI’s application and exploring new options for utilization.

During the September 2024 episode of Lungcast, a conversation about AI in lung imaging featured Rachel Eddy, PhD, assistant professor in the departments of radiology and pediatrics at the University of British Columbia, and an investigator at the BC Children’s Hospital. Eddy joined to guide listeners through the current and future uses of AI in lung imaging.

In this episode, Eddy and Rizzo highlight a variety of functions and ongoing research related to the implementation of AI in imaging. During the above segment of Lungcast spoke about recent use of generative AI as well as terms commonly used.

"'Deep learning' combines math and stats as well, but with these complex neural networks to map the start to the finish compared to machine learning, deep learning especially requires really large data sets to learn accurate mappings of that start to finish," Eddy explained. "However, we don't always know what that mapping is and how the network generated that mapping. This is why deep learning is often called a 'black box,' where we can't actually see inside that box and how the machine learned the mapping."

Additionally, Eddy covers her own work and its implementation of AI technology. She also notes the impact on patients with COPD or asthma.

"...(My) work so far has used AI algorithms to derive measurements from CTS of the lungs, especially of the of the airways," she said. "And we've used those to tell us something about asthma and patients with asthma."

Lungcast is a monthly respiratory health podcast series which is hosted by Albert Rizzo, MD, the chief medical officer of the American Lung Association (ALA), and produced by the ALA and HCPLive.

Subscribe to Lungcast on Spotify here, or listen to the episode below.