According to a study published in Radiology, artificial intelligence (AI) is an excellent tool for fracture identification that has the ability to help physicians in congested emergency rooms.
Hierarchical summary receiver operating characteristic (HSROC) curves for (A) fracture detection algorithms and (B) clinicians with internal validation test sets. The 95% prediction region is a visual representation of between-study heterogeneity. Image Credit: Kuo, et al, Radiology 2022; 000:1–9, ©RSNA 2022.
Fracture diagnosis that is overlooked or delayed on X-Rays is a typical mistake that can have catastrophic consequences for the patient. The situation is exacerbated by a lack of ready access to expert advice as imaging volumes tend to outpace radiologist recruitment.
AI may be able to help solve this problem by functioning as a tool for radiologists, allowing them to diagnose fractures faster and more accurately.
A group of researchers in England analyzed 42 current studies comparing the diagnostic accuracy of AI and doctors in fracture diagnosis to understand more about the technology’s capabilities in the fracture situation. In 37 of the 42 investigations, X-Rays were used to diagnose fractures, whereas CT was utilized in five.
There were no statistically significant differences between clinician and AI performance, according to the researchers. The sensitivity of AI in diagnosing fractures was between 91 and 92%.
Technology May Reduce the Rate of Early Misdiagnosis
We found that AI performed with a high degree of accuracy, comparable to clinician performance. Importantly, we found this to be the case when AI was validated using independent external datasets, suggesting that the results may be generalizable to the wider population.
Dr. Rachel Kuo, MBBCh, Study Lead Author, Botnar Research Centre, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, Oxford University
According to Dr. Kuo, the findings of the study lead to various possible clinical and academic uses for AI in fracture identification. It has the potential to minimize the risk of early misdiagnosis in difficult emergency situations, such as when patients have numerous fractures. It also has the potential to be used as a training tool for junior clinicians.
It could also be helpful as a ‘second reader,’ providing clinicians with either reassurance that they have made the correct diagnosis or prompting them to take another look at the imaging before treating patients.
Dr. Rachel Kuo, MBBCh, Study Lead Author, Botnar Research Centre, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, Oxford University
Dr. Kuo noted that AI-based fracture diagnosis research is still in the early stages of development. Only a small percentage of the research that she and her colleagues investigated, assessed physician performance with AI aid, and there was only one example of an AI being tested in a systematic clinical study.
“It remains important for clinicians to continue to exercise their own judgment,” Dr. Kuo concluded. “AI is not infallible and is subject to bias and error.”
Journal Reference:
Kuo, R. Y. L., et al. (2022) Artificial Intelligence in Fracture Detection: A Systematic Review and Meta-Analysis. Radiology. doi.org/10.1148/radiol.211785.
Source: https://www.rsna.org/