Assessing a lung scan for a COVID-19 patient can take anywhere from five to 15 minutes. Determining whether a patient is at risk of becoming critically ill from the coronavirus could take even longer. That’s why some scientists are now trying to speed the process up with the help of artificial intelligence.
Scientists in China and tech giant Tencent worked together to develop a new AI model that calculates the risk of a patient’s health severely deteriorating from COVID-19. Development of the tool also got help from renowned infectious disease expert Zhong Nanshan, known for helping manage the early-2000s SARS epidemic, who set up an AI lab together with Tencent in February.
The findings of the project were published in scientific magazine Nature last week. Tencent also made the tool available on its website, where it’s officially called the “Calculation Tool for Early Triage of Critically-ill COVID-19 Patients using Deep Learning.” The hope is that it could help the most at-risk patients get the care they need faster and advise hospitals on how to best allocate their resources.
The tool is based on a deep learning model using data from nearly 1,600 patients from different medical institutions. To predict a patient’s risk level, the scientists narrowed down a list of features to what they found to be the 10 most relevant ones. These include X-ray abnormalities, age, cancer history, and more. These were then used to establish the AI model for risk prediction.
The tool is another example of tech giants using AI for medical applications. In the US, IBM has Watson for Health and Google has DeepMind Health. In addition to Tencent, other Chinese tech giants, including Alibaba, have been using AI for everything from better diagnostics to drug research. Deep learning models have already been used to diagnose skin cancer and other types of carcinoma.
Both Alibaba and Tencent have offered their AI gene sequencing tools to help fight the spread of the coronavirus.
Tencent has also developed a COVID-19 self-screening tool and a platform for pandemic updates. Another solution coming from its lab is an AI pattern recognition tool that analyses CAT scans for quicker diagnosis.
This article was first published by Abacus.
