Artificial Intelligence: AI Identifies Covid’s Tell-Tale Traces From Lung Ultrasounds
The system was 100% accurate in distinguishing COVID-19 lungs from non-coronavirus pneumonia.
Researchers at Lawson Health Research Institute, London, have developed an AI-based computer system to pinpoint distinctive COVID-19 markers in the ultrasound scans of lungs of virus-afflicted patients. These hidden traces are indistinguishable even to highly trained and experienced medical personnel. (The London Free Press)
The scientists trained an AI model on a dataset of lung ultrasounds from 243 patients, including 612 video loops and nearly 121,400 frames.
The scientists tested their new tool on a mix of lung ultrasounds of both COVID and non-COVID patients to verify whether it could tell them apart.
According to Robert Arntfield, a Lawson researcher and medical director of London Health Sciences Centre’s critical care trauma center, the system was able to identify the COVID scans with perfect accuracy.
This is a significant achievement given that the ultrasounds of both categories look highly similar, even to specialists.
In fact, when 61 doctors were shown 25 ultrasound clips of patients with COVID-19 and non-coronavirus lung conditions, the results were mixed (“no better than a coin toss”), but again, the computer could accurately pick one from the other.
COVID-19’s “QR Code”
Dr. Robert Arntfield, Lawson Researcher and Medical Director of the Critical Care Trauma Centre at LHSC said: “The neural network was able to identify the unique characteristics among different scans, and exceed human-level diagnostic specificity. Our study of over 100,000 ultrasound images showed that while trained physicians could, as expected, not distinguish between different causes of lung disease, the AI had nearly perfect accuracy in making the diagnosis. It’s almost like the AI sees a QR code that we cannot see, unique to the disease.”
Project Deep Breathe
The results from the AI experiment have encouraged the Lawson Health Research Institute to green light Dr. Arntfield’s “Project Deep Breathe” project.
The project will venture beyond COVID-19 and get the AI tool to analyze other, multiple conditions using lung ultrasounds.
Related Story: AI Detects COVID From Cellphone Recorded Coughs
Image Credit: Lawson Health Research Institute
Latest Alternative Investment News
In a unique plug for bitcoin’s adoption, MicroStrategy (NASDAQ: MSTR), the business information firm, has decided to pay its non-employee directors in bitcoin. The company has been a trendsetter, setting…
Alternative Investments/Digital: Yet Another Bitcoin ETF Application; Galaxy Digital Is The Latest Hopeful
Galaxy Digital, the cryptocurrency investment firm led by Mike Novogratz, has filed an application with the SEC to launch a bitcoin ETF that will trade on the NYSE Arca. The…
Microsoft (NASDAQ: MSFT) may buy Nuance Communication (NASDAQ: NUAN) at a valuation of $56 per share, according to Bloomberg, which cited unnamed people who were familiar with the matter. The…
Venture Capital: Supervest, A Platform That Matches Loan-Seeking Merchants With Investors, Is Seeking Funds
Supervest calls it a new asset class. Merchant cash advances (MCA) are a form of short-term financing for merchants in the form of a sales agreement in which capital is…