An artificial intelligence (AI) system has been developed which can interpret and prioritise abnormal chest X-rays with critical findings, potentially reducing backlogs, according to a study in the journal Radiology.
In the United Kingdom, there are an estimated 330,000 X-rays at any given time that have been waiting more than 30 days for a report. Deep learning technology, a type of AI, has been proposed as an automated means to reduce this backlog and identify exams that merit immediate attention, particularly in publicly funded healthcare systems.
Researchers at King's College London, Guy's & St Thomas' Foundation and the University of Warwick developed the system using 470,388 chest radiographs X-rays undertaken from 2007 to 2017. Radiology reports were pre-processed using an in-house natural language processing system, which classified X-rays as critical, urgent, non-urgent or normal. The system was tested in a simulation using an independent set of 15,887 radiographs.
The AI system detected normal radiographs with a sensitivity of 71%, specificity of 95%, positive predictive value of 73% and negative predictive value of 94%.
The average reporting delay was reduced from 11.2 to 2.7 days for critical imaging findings (P<.001) and from 7.6 to 4.1 days for urgent imaging findings (P<.001) in the simulation compared with historical data.
The researchers now plan to study the technology in a much larger sample size and deploy more complex algorithms for better performance. Future research goals include a multi-centre study to prospectively assess the performance of the triaging software.