Mobile Menu

AI predicts Covid-19 “wet lung” trajectories

Researchers in China and the US have developed an AI tool capable, they say, of accurately predicting which patients in the early stages of Covid-19 will go on to develop severe lung disease and which won’t. The tool could help doctors fairly prioritise patients for treatment as many countries are facing the realities of over-stretched healthcare systems.

The study applied machine learning algorithms to data of 53 patients in Chinese hospitals to assess what factors are critical in the development of acute respiratory disease syndrome (ARDS). In this complication the lungs fill with fluid and it proves fatal to 50% of coronavirus patients.

The three markers that proved to be the best indicators of future severity were: levels of the liver enzyme alanine aminotransferase (ALT), reported body aches, and haemoglobin levels. The reported accuracy of the model in predicting progression to ARDS was 80 percent.

The study authors were surprised that a previously considered hallmarks of Covid-19, including particular patterns in lung scans dubbed “ground glass opacity” weren’t valuable predictors of ARDS. Nor were other factors, including immune response, fever, sex, or patient age, reliable indicators.

Megan Coffee, co-author of the paper at NYU Grossman School of Medicine explained: “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.”

More on these topics

AI / Coronavirus / Disease Trajectories

Share this article