To facilitate the identification of fractures on X-rays, researchers from Boston University School of Medicine (BUSM) have developed an algorithm that uses artificial intelligence (AI).
According to Gadgetendency, the researchers explain that their system is capable of detecting broken bones automatically and quickly, and then reporting them to radiologists. The algorithm is also capable of indicating areas where fractures are likely to occur.
With its capabilities, this algorithm can help radiologists prioritize reading X-ray images with proven fractures, said Ali Guermazi, head of radiology at BUSM. . Using algorithms can also save patients time when receiving fracture diagnoses and help doctors provide care sooner.
About 24% of human diagnostic errors that can harm patients are due to misinterpretation of broken bones, the researchers explain. With an AI-based algorithm, they hope to reduce this error rate thanks to an expanded radiographic data set.
To test the effectiveness of the algorithm, several experts have established the gold standard for comparing the performance of many different human reading sources (radiologists, orthopedic surgeons, emergency physicians, etc.) , family doctor …) with and without artificial intelligence support . After testing, the researchers note that their algorithm has allowed readers to reduce the number of missed fracture diagnoses by 29% as well as increase detection sensitivity by 16%. For tests involving some fractures, it can be seen that with the help of AI, their accuracy has increased by 30%.
Also according to Ali Guermazi, AI can be a powerful tool for medicine. It can help not only radiologists but also other clinicians to be more efficient and faster in their work.