A group of Chinese and US researchers have jointly developed a new artificial intelligence tool to predict which newly infected coronavirus patients would have a higher risk for severe complications from Covid-19.
Researchers at the NYU Grossman School of Medicine and the Courant Institute of Mathematical Sciences at New York University, along with their counterparts at the Wenzhou Central Hospital and Cangnan People’s Hospital, worked on a study that designed computer models, which made decisions based on data from the two Chinese hospitals in the eastern coastal city of Wenzhou.
Their study, which was published online on March 30 in the journal Computers, Materials & Continua, focused on an AI tool that could help accurately predict which coronavirus patients would go on to develop acute respiratory distress syndrome (ARDS) - the fluid build-up in the lungs that can be fatal in the elderly.
“While work remains to further validate our model, it holds promise as another tool to predict the patients most vulnerable to the virus, but only in support of physicians’ hard-won clinical experience in treating viral infections,” said study author Megan Coffee, a clinical assistant professor in the Division of Infectious Disease & Immunology within the Department of Medicine at NYU Grossman School of Medicine, in a statement released on Monday in the US.
Their findings come at a time when the worldwide coronavirus death toll soared towards 37,000 amid new waves of outbreaks in the US, where there is increasing pressure on overwhelmed hospitals like those in New York that are bracing for the peak of the pandemic.
It also reflects how the Covid-19 outbreak has led to a surge in the use of AI applications to help contain the spread of the disease. In mainland China, some of the initial applications range from robotic cleaners spraying disinfectant at segregated wards and apps that can track people’s travel history to AI voice assistants calling people to give advice on home quarantine.
Combined with other factors, the new AI tool jointly developed by researchers from New York City and Wenzhou was able to predict risk of ARDS with up to 80% accuracy, according to their study.
It used demographic, laboratory and radiological findings collected from 53 patients in Wenzhou, as each tested positive for Covid-19 at the two Chinese hospitals in January. Symptoms typically included cough, fever and upset stomach. In a minority of patients, however, severe symptoms like pneumonia developed within a week.
The new AI tool found that mild increases in deep muscle aches, enzyme alanine aminotransferase (ALT) and haemoglobin levels were useful in predictive analysis.
Liver-damaging diseases like hepatitis often show a dramatic increase in ALT levels, while high level of haemoglobin - the iron-containing protein that enables blood cells to carry oxygen to bodily tissues - were linked to later respiratory distress.
“We hope that the tool, when fully developed, will be useful to physicians as they assess which moderately ill patients really need beds, and who can safely go home, with hospital resources stretched thin,” said study co-author Anasse Bari, a clinical assistant professor in computer science at the Courant Institute, in a statement.
Limitations of their research, according to the authors, included the relatively small data set and the slight clinical severity of disease in the number of patients studied.
Still, there are other ongoing AI initiatives that also aim for the accurate diagnosis of coronavirus patients. Researchers from the Huazhong University of Science and Technology and Tongji Hospital in Wuhan, said they have developed an AI diagnostic tool that can quickly analyse blood samples to predict survival rates.
The developers claim the AI tool achieved 90% accuracy on the fatality and survival rates of more than 400 patients, based on blood samples collected from Tongji Hospital. Their findings were released on the preprint server Medrxiv.org, a platform scientists worldwide use to release non-peer reviewed research on Covid-19.
- SCMP