Researchers at Edith Cowan University have developed software to quickly analyze bone density scans to detect abdominal aortic calcification (AAC), a predictor of cardiovascular events and other health risks. The software processes images with up to 80% agreement with experts, revolutionizing early disease detection in routine clinical practice.
Bone density scans can now quickly identify cardiovascular health risk indicators. With artificial intelligence, we will soon be able to predict our future risk of serious health illnesses at the push of a button. Abdominal aortic calcification (AAC) is the deposition of calcium in the wall of the abdominal aorta. It may signal an increased risk of cardiovascular events, including heart attack and stroke.
It can also predict your risk of falls, fractures and dementia later in life. Conveniently, regular bone density machine scans used to detect osteoporosis can also detect AAC. However, highly trained experts are required to analyze the images, and the analysis process can take 5-15 minutes per image.
But researchers from Edith Cowan University's (ECU) School of Science and the School of Medicine and Health Sciences have collaborated to develop software that can analyze scanned images much faster: around 60,000 images a day.
Associate Professor Joshua Lewis, researcher and Heart Foundation Future Leaders Fellow, said this huge increase in efficiency is crucial for widespread use of AAC in research and for helping people avoid health problems later in life.
"Because these images and automated scoring can be quickly and easily obtained at the time of bone density testing, this may in the future lead to new methods for early cardiovascular disease detection and disease monitoring in routine clinical practice," he said.
The results come from an international collaboration between Edith Cowan University, the University of Western Australia, the University of Minnesota, the University of Southampton, the University of Manitoba, the Marcus Institute on Aging and Hebrew Senior Living Harvard Medical School. This is a truly multidisciplinary global collaboration. While this is not the first algorithm to assess AAC from these images, this study is the largest of its kind, is based on the most commonly used bone density machine model, and is the first to be tested in the real world using images as part of routine bone density testing.
More than 5,000 images were analyzed by the expert and research team's software. After comparing the results, the experts and the software came to the same conclusion about the degree of AAC (low, medium or high) 80% of the time - an impressive number considering this is the first version of the software. Importantly, only 3% of people thought to have high AAC levels were incorrectly diagnosed by the software as having low AAC levels.
Professor Lewis said: "This is noteworthy because these are the people with the most severe disease and the highest risk of fatal and non-fatal cardiovascular events and all-cause mortality. While the accuracy of the software still needs to be improved compared to manual readings, these results were generated by version 1.0 of our algorithm and our latest version has greatly improved the results."
"Automated assessment of the presence and extent of AAC with an accuracy similar to that of imaging experts opens up the possibility of large-scale screening for cardiovascular disease and other diseases - even before someone develops any symptoms. This will enable people at risk to make necessary lifestyle changes earlier and enable them to live healthier later in life."