This AI Tool Predicts Which COVID-19 Patients Will Develop 'Wet Lung'
Researchers in China and the U.S. reported on Monday they've developed an artificial intelligence (AI) tool capable of accurately predicting which people recently infected with the coronavirus will also develop severe lung disease.
AI versus the COVID-19 coronavirus
Once activated, the algorithm might help doctors deciding which patients to prioritize amid exceedingly-stretched health care systems, said Megan Coffee, a physician and professor at New York University's Grossman School of Medicine — who also co-authored a paper on the new findings in the journal Computers, Materials & Continua — to phys.org.
The AI tool found numerous interesting indicators that strongly predict which people infected with the novel virus later developed so-called acute respiratory disease syndrome (ARDS), a severe complication that often follows the COVID-19 illness, which fills the lungs with fluid, killing roughly 50 percent of coronavirus patients who get it.
The team behind the research applied a machine-learning algorithm to data collected from 53 coronavirus patients populating two hospitals in Wenzhou, China, and found that changes in three critical features — liver enzyme alanine aminotransferase (ALT) levels, hemoglobin levels, and reported body aches — most accurately predicted subsequent, severe disease.
AI helps doctors cope with flood of coronavirus patients
Contrasting with this were characteristics thought to be related to COVID-19, including a specific pattern in lung images called "ground glass opacity," strong immune responses, and fevers, which weren't useful factors to predict which patients with initially mild symptoms would progress to full-blown ARDS.
Sex and age weren't useful predictors either, despite other studies showing that men more than 60 years old were at higher risk.
"It's been fascinating because a lot of the data points that the machine used to help influence its decisions were different than what a clinician would normally look at," said Coffee to AFP, according to TechXplore.
Of course, using AI in medical settings isn't a new concept — another AI-based tool exists that helps dermatologists predict which patients will later develop skin cancer.
This is significant because doctors are learning about COVID-19 on-the-fly, can helps doctors decide — diagnostically — which way to go, and gives them more criteria upon which to decide which patients should take priority when hospitals around the world are overhwhelmed from the COVID-19 pandemic, according to co-author Anasse Bari, a computer science professor at NYU, reports TechXplore.
With the coronavirus outbreak deepening its grip in global hubs of business, science, and culture like New York City, every advantage new technologies can provide in the fight against the pandemic is more than welcome. It's necessary.
Can AI rescue lab rats and guinea pigs? Science’s long-debated ethical question may finally have some answers.