Artificial intelligence tool detects COVID-19 in lung ultrasound images: Study
Representative image. Photograph:( Others )
Story highlights
The university said, 'The findings culminate an effort that started early in the pandemic when clinicians needed tools to rapidly assess legions of patients in overwhelmed emergency rooms'
The university said, 'The findings culminate an effort that started early in the pandemic when clinicians needed tools to rapidly assess legions of patients in overwhelmed emergency rooms'
The coronavirus (COVID-19) pandemic has subsided, but its impact can still be seen as long Covid. Hence, regular monitoring of cases and symptoms, in the form of testing, is necessary.
As we all know samples for Covid diagnostic tests are typically collected using an anterior nares (nasal) swab sample. Some diagnostic tests involve other samples such as mid-turbinate, nasopharyngeal, oropharyngeal, or saliva.
Meanwhile, in the latest development, Johns Hopkins University has highlighted research that shows artificial intelligence (AI) can spot Covid in lung ultrasound images.
Also read: Don't let menopause pause you, embrace it with confidence and dismantle stigmas
Researchers published a study on March 11 in Communications Medicine that aims to explain the creation of an AI system that analyses lung ultrasound pictures.
The scans detect patterns known as B-lines, which show as bright, vertical anomalies and suggest inflammation in patients with pulmonary problems.
In a news release, the university said, "The findings culminate an effort that started early in the pandemic when clinicians needed tools to rapidly assess legions of patients in overwhelmed emergency rooms."
Also read: Rising cases of silent heart attacks linked to long Covid? WION decodes
The study further noted that the AI was trained by mixing computer-generated images with actual ultrasounds of patients, including some of those who sought care at Johns Hopkins.
As quoted, first author Lingyi Zhao, PhD, who developed the software during postdoctoral fellowship work, said, "Early in the pandemic, we didn't have enough ultrasound images of COVID-19 patients to develop and test our algorithms, and as a result, our deep neural networks never reached peak performance."
Further, the researchers taught the AI to learn from a combination of real and simulated data in order to identify abnormalities in ultrasound scans that suggest a person has COVID-19.
Zhao explained that the tool is a deep neural network, a sort of artificial intelligence designed to act similarly to the interconnected neurons that allow the brain to recognise patterns, process speech, and perform other complicated tasks.
(With inputs from agencies)
