Share this:

Like this:

AI Machine Learning Tool predicts COVID-19 survival from blood samples from critically ill patients

While the Omicron variant is raging across the United States and the rest of the globe, we see hospitals everywhere maximizing their capacity to treat critically ill COVID-19 patients - not all of whom have an obvious risk factor. In the intensive care unit, it is often difficult to determine who can survive and beat the infection, and who can end up succumbing to the disease.

That may change very soon thanks to a new artificial intelligence tool that can predict the survival outcome of severe cases of COVID weeks in advance - all from a single blood test. The new model, described in a peer-reviewed paper published in PLOS Digital Health Tuesday, could help physicians make more informed treatment plans for COVID patients during initial hospitalization, to minimize the odds of mortality and improve patient care along the way.

"The clinical picture of COVID-19 is unusually diverse, ranging from asymptomatic infection to very serious illness and death," said Florian Kurth, a clinical researcher at Charity-University Medicine in Berlin and lead author of the new study, in a statement. "For physicians, it is difficult to assess the individual risk for a patient of deterioration and / or death due to COVID-19." The new AI model "may well predict the likelihood that an individual patient will die or survive COVID-19."

The new study can be divided into two parts. First, the researchers examined hundreds of blood samples from 50 critically ill COVID patients treated in Germany and Austria to learn how the levels of 321 different proteins changed during the course of the infection. All patients were on intensive care and required ventilation and additional organ replacement therapy. Fifteen patients died and 35 survived.

The researchers found that there were 14 proteins most strongly associated with either COVID survival or death, and that these protein levels were altered early by the disease. The most important were proteins involved in blood coagulation and antibody function.

In the second part, the new AI model was built and trained to make prognoses for COVID patients based on the levels of these 14 proteins in a single blood sample. Kurth and his colleagues tested this model on samples from a new cohort of 24 critically ill patients. Of this cohort, 19 patients survived and five died - and the model was able to accurately predict the results for all but one of these patients. These predictions were "far better" than those used in current clinical risk assessments, Kurth said.

The current study is based on an extremely small sample size, so the authors want to run the model through much broader tests to prove that it could be a valuable prediction for real-world COVID-19 outcomes in hospital settings.

It will take some time, but they also hope that the results of the paper can help other researchers better understand the mechanisms behind COVID and what kinds of treatments should be advanced to stop disease progression in general. Kurth said there is hope that the lessons here could open the doors to a new future for medicine, where we make early predictions about a disease based solely on a blood test.


Leave a Reply

Your email address will not be published. Required fields are marked *

Share this:

Like this:

%d bloggers like this: