Email
Password
Remember meForgot password?
    Log in with Twitter

Essential Science: AI can reduce X-ray screening times

By Tim Sandle     Jan 28, 2019 in Science
A new study shows how machine learning can be applied to reduce the amount of time needed to process abnormal chest X-rays. The system was developed through reviewing thousands of medical images.
A study from the University of Warwick (U.K.) has shown how a novel artificial intelligence platform can significantly lower the time required to ensure that abnormal chest X-rays, with serious findings, receive an expert radiologist opinion sooner.
This method can reduce the average delay from eleven days to less than three days. The types of X-rays in scope are those of the chest. Such X-rays are performed to allow radiographers to diagnose and monitor conditions impacting upon the lungs, heart, bones and soft tissues.
AI in healthcare
Creative Commons image by SurfaceWarriors
Creative Commons image by SurfaceWarriors
Creative Commons image by SurfaceWarriors
Artificial intelligence is increasingly used to assist medics with healthcare data processing and with making decisions. One application is drug discovery, such as the use of small systems-of-interest specific datasets to improve drug development and personalized medicine.
In a different area, a new computer program has been developed to analyze images of patients' lung tumors, specify cancer types and even identify altered genes driving abnormal cell growth.
A third example is by applying artificial intelligence to help lung doctors interpret respiratory symptoms accurately and make a correct diagnosis. This is using machine learning to assess pulmonary function tests like spirometry.
Speeding up X-ray reviews
Atomic-resolution structure of an in vivo-grown cockroach protein - III.
Atomic-resolution structure of an in vivo-grown cockroach protein - III.
International Union of Crystallography
With the new application of artificial intelligence, the British researchers teamed up with medical staff from London hospitals — Guy’s and St Thomas' NHS Hospitals. This gave the researchers access to a dataset composed of half a million anonymised adult chest radiographs (as X-rays). Using these images, the researchers were able to develop an artificial intelligence platform for computer vision.
The new platform can recognise radiological abnormalities in the X-rays in real-time. Furthermore, the system is capable of suggesting how quickly these exams should be notified by a radiologist.
To aid the development of the artificial intelligence, the science team constructed and tested a Natural Language Processing algorithm that can read a radiological report. The language also allows the artificial intelligence to understand the findings described by the radiologist. From this the platform can determine the priority level of the exam.
The system was ‘taught’ by applying the algorithm to the historical exams. This involved processing a large volume of data for the artificial intelligence system to understand those visual patterns in X-rays that were predictive of medical importance. With the test runs, the artificial intelligence was shown to have a 73 percent reliability.
Commenting on the research and development process, lead researcher Professor Giovanni Montana, who has expertise in data science, said: “Artificial intelligence led reporting of imaging could be a valuable tool to improve department workflow and workforce efficiency.”
He adds: “The increasing clinical demands on radiology departments worldwide has challenged current service delivery models, particularly in publicly-funded healthcare systems. It is no longer feasible for many Radiology departments with their current staffing level to report all acquired plain radiographs in a timely manner, leading to large backlogs of unreported studies."
The research paves the way for other alternative models of care, such as computer vision algorithms, to be used to identify abnormal images, such as with computed tomography, magnetic resonance imaging and positron emission tomography.
Research paper
The research has been published in the journal Radiology, with the research paper titled “Automated Triaging of Adult Chest Radiographs with Deep Artificial Neural Networks.”
Essential Science
Lab-grown beef.
Lab-grown beef.
Mosa Meat / media photo
This article is part of Digital Journal's regular Essential Science columns. Each week Tim Sandle explores a topical and important scientific issue. Last week we looked at the potential for lab-grown meat, that is meat grown in cell culture instead of inside animals – what is referred to as 'cellular agriculture'.
The week before, we outlined how a new compound, identified in coffee, has been shown to be a potential treatment in the fight against neurodegenerative conditions like Parkinson’s disease and Lewy body dementia, according to new research.
More about Xrays, Artificial intelligence, Medical
 
Latest News
Top News