The AI system can also help in decreasing backlogs in hospital in the future. According to the researchers, there is a possibility of large backlogs related to chest X-rays, as it makes up about 40 % of all diagnostic imaging across the globe.
“Currently, there are no systematic and automated ways to triage chest X-rays and bring those with critical and urgent findings to the top of the reporting pile,” Giovanni Montana, study co-author, explained. He is currently working at the University of Warwick in Coventry, England and before that, he was associated with King’s College London.
Montana and his co-workers utilized over 470,300 adult chest X-rays to create an AI system that could recognize rare results.
A simulation was used to assess the performance of the system in prioritizing X-rays by checking on a distinct group of 15,887 chest X-rays. To protect patient privacy, every classifying data was eradicated from the X-rays.
Researchers claimed that the system produced highly accurate results in recognizing abnormal from normal chest X-rays. Simulations demonstrated that critical results got an expert radiologist opinion in an average of 2.7 days with the AI system, whereas in real practice an average of 11.2 days would produce the same results.
Radiology journal posted the study results on 22 January.
“The initial results reported here are exciting as they demonstrate that an AI system can be successfully trained using a very large database of routinely acquired radiologic data,” Montana claimed.
He further said: “With further clinical validation, this technology is expected to reduce a radiologist’s workload by a significant amount by detecting all the normal exams, so more time can be spent on those requiring more attention.”
The experts claimed that the coming plan is to test many numbers of X-rays and to perform a multi-center training to evaluate the performance of the AI system.