Facebook says that it has created artificial intelligence (AI) software that is able to predict if a Covid-19 patient’s health will continue to deteriorate, as well as whether they will need oxygen, simply by scanning the individual’s chest X-rays.
In collaboration with New York University (NYU), the tech giant said that its system can make such a calculation and offer a four-day advance warning of deterioration. The research duo, writing in a blog post, explain that they constructed machine-learning models that can help doctors in better preparing to care for patients, as global coronavirus cases continue to spike.
The two institutions claim that one model is able to predict the worsening of a patient’s condition by analysing a single chest X-ray, while another provides a similar function but instead makes its judgement by looking at a series of X-rays. A final model, the blog authors claim, is able to forecast how much oxygen assistance a patient could need in days ahead, based on their current condition.
Both Facebook and NYU say that they have five researchers working on digital Covid-19 prediction modeling software.
“Our model using sequential chest X-rays can predict up to four days (96 hours) in advance if a patient may need more intensive care solutions, generally outperforming predictions by human experts,” the authors of the software write.
The software was fine-tuned by analysing 26,383 NYU chest X-ray images taken from 4,914 patients over time. This allowed the models to document the telltale signs of the worsening of a patient’s condition over a period of 24, 48, 72, and 96 hours.
“Based on reader studies that we conducted with radiologists at NYU Langone, our models that used sequences of X-ray images outperformed human experts at predicting ICU needs and mortality predictions, and overall adverse event predictions in the longer term (up to 96 hours). Being able to predict whether a patient will need oxygen resources would also be a first, and could help hospitals as they decide how to allocate resources in the weeks and months to come,” the team write.
Currently, there are over 93 million documented cases of human Covid-19, with at least two million deaths caused by the virus.