Ever wondered what the future holds for your health? Well, it might soon be just a button press away! How, you ask? Artificial Intelligence, or AI, our multi-talented digital ally, is at it again, this time promising to shed light on potential health issues before they take a toll on us.
The Magical Mix of AI and Medical Imaging
Let’s start by talking about Abdominal Aortic Calcification, or AAC. Sounds like a mouthful, doesn’t it? Simply put, AAC is a build-up of calcium deposits in the walls of your abdominal aorta, and it’s like the fortune teller of health risks – predicting potential heart attacks, strokes, late-life dementia, and even your risk of falls and fractures. Pretty important stuff to know about, right?
Here’s where it gets even more interesting. Your regular bone density scans, the same ones used to check for osteoporosis, can also spot AAC. But there’s a small catch. These images need to be meticulously analyzed by highly trained experts, a process that can take somewhere between 5 to 15 minutes per image. That’s a lot of time, especially when you consider the sheer volume of images generated every day.
AI to the Rescue
But what if this tedious task could be completed in a fraction of the time? That’s exactly what a team of brilliant researchers from Edith Cowan University’s (ECU) School of Science and School of Medical and Health Sciences have achieved. They’ve created a software, powered by AI, that can churn through a whopping 60,000 images in a single day! Yes, you read that right.
This significant upgrade in efficiency doesn’t just mean less waiting time for results. It paves the way for routine and widespread use of AAC in research and potentially early disease detection. Imagine getting a heads-up on your risk of serious health problems during a regular health check-up. Now that’s what I call ‘preventive healthcare’.
The Power of Collaboration
Achieving such a monumental task was not an individual feat. This was an international and multidisciplinary collaboration involving the ECU, the University of WA, University of Minnesota, Southampton, University of Manitoba, Marcus Institute for Aging Research, and Hebrew SeniorLife Harvard Medical School.
Now, before you start thinking, “Wait, isn’t this the first time something like this is being tried?”, let me assure you it’s not. But this study stands out for its scope, using the most commonly used bone density machine models, and testing the software in a real-world setting using images taken as part of routine bone density testing.
Proving the Concept
More than 5000 images were fed to the software for analysis. The results? Pretty impressive, to say the least! The software managed to reach the same conclusion as human experts about the extent of AAC (categorized as low, moderate, or high) 80 percent of the time. This high level of accuracy is especially noteworthy considering this was just the first version of the software.
Now, the software wasn’t perfect. It did misdiagnose about 3 percent of the cases, stating that people with high AAC levels had low levels. However, this is the first iteration of the algorithm, and the team is confident that they can make significant improvements in subsequent versions.
A Bright Future Ahead
The research offers a glimmer of hope for the future of healthcare. Just think about it. An AI tool that can automatically evaluate the extent of AAC with accuracy levels similar to imaging specialists opens up the possibility of large-scale screenings for cardiovascular disease and other conditions. And the best part? This can happen even before you experience any symptoms.
Such advancements would enable us to make necessary lifestyle changes far earlier, setting us up for healthier years ahead. Now isn’t that a future we all would want?
The project was funded by the Heart Foundation, and the findings were published in the eBioMedicine journal. As we continue to blend the power of AI with medical research, we’re unlocking the potential to revolutionize healthcare as we know it.
Naeha Sharif, Syed Zulqarnain Gilani, David Suter, Siobhan Reid, Pawel Szulc, Douglas Kimelman, Barret A. Monchka, Mohammad Jafari Jozani, Jonathan M. Hodgson, Marc Sim, Kun Zhu, Nicholas C. Harvey, Douglas P. Kiel, Richard L. Prince, John T. Schousboe, William D. Leslie, Joshua R. Lewis. Machine learning for abdominal aortic calcification assessment from bone density machine-derived lateral spine images. eBioMedicine, 2023; 104676 Link
