COVID-19 is taking its toll on the world, affecting more than 200 countries and regions. In the face of a surge in cases, rapid diagnosis and identification of high-risk patients for early intervention is crucial for controlling the pandemic.
The Hong Kong Polytechnic University (PolyU) is committed to the fight against COVID-19 and has been collaborating with various partners in the prevention and control of the disease.
Thus, PolyU announced the launch of a new AI system jointly developed with the Macau University of Science and Technology (M.U.S.T.) in conjunction with other Chinese institutions, to assist in the rapid diagnosis of COVID-19 pneumonia.
This research was published in the prestigious international journal Cell, in a paper entitled “Clinically Applicable AI System for Accurate Diagnosis, Quantitative Measurements, and Prognosis of COVID-19 Pneumonia Using Computed Tomography”.
The novel AI-assisted diagnostic system, based on computed tomography (CT) chest imaging, can efficiently distinguish COVID-19 pneumonia from other viral pneumonia, and predict the severity level in 20 seconds, with an accuracy of up to 90% or above. It assists radiologists and physicians in performing a quick diagnosis, especially when the health system is overloaded.
According to a Professor from the Faculty of Medicine of M.U.S.T. and the first author of the research paper, the research team used a large CT database of 3,777 patients to train and test the AI system to differentiate novel coronavirus pneumonia (NCP) from other common pneumonia and normal controls.
This system can provide accurate clinical prognosis that can aid clinicians when considering appropriate early clinical management and appropriate resource allocation. It has been developed to assist global clinicians in combating COVID-19.
The AI team uses deep learning, semantic segmentation and multiple neural network architectures to create an accurate AI diagnostic system comprising texture analysis, quantitative measurements and diagnosis classification.
In addition, based on the international standard for related research on lung lesions, lung, liver damage and relevant clinical indicators, they have also developed a comprehensive rating system to project the emergence of pneumonia severe lesions accurately for the follow-up treatment.
The scientists from PolyU and M.U.S.T. have also collaborated on other projects, including the seroconversion study in SARS-CoV-2 (COVID-19) patients in China, which helped pave the way for further diagnostic and therapeutic development to treat SARS-CoV-2.
The study “Seroprevalence of immunoglobulin M and G antibodies against SARS-CoV-2 in China” was also recently published in another prestigious scientific journal Nature Medicine.
The Deputy President and Provost of PolyU stated that by leveraging PolyU’s strengths and its collaborations with different research partners, the university is able to contribute to the well-being of humankind.
During the outbreak, university scholars developed a rapid automated diagnostic system to detect COVID-19 and have conducted various scientific experiments related to the pandemic.
Since the establishment of the biotechnology and translational medicine international collaborative platform in 2018, PolyU has been working closely with M.U.S.T. and other partners from the United States and the Greater Bay Area for high-impact research.
The Dean of the Faculty of Medicine of M.U.S.T. remarked that this AI system had been implemented in many hospitals in different countries.
M.U.S.T. is collaborating with several prominent universities including Baylor University, University of Buffalo and Stanford University to serve the US populations. He believed this application could soon be adopted in Hong Kong, more regions of Mainland China and worldwide.
COVID-19 is a new and highly contagious virus with over 7 million people infected (as of June 10). Researchers from PolyU and M.U.S.T. with their supportive partners will continue their efforts in COVID-19 research to use scientific technologies to combat this pandemic through prevention, diagnosis, treatment and control.
