AI helps radiologists distinguish coronavirus from community-acquired pneumonia in chest CT

The symptoms of coronavirus are very similar to other common lung diseases such as Asthma, Pneumonia or Sleep apnea. The major symptoms of Covid-19 include cough, fever, tiredness, difficulty breathing.

Researchers in China have developed an AI-powered model that has helped radiologists in China to distinguish COVID-19 from community-acquired pneumonia and other lung diseases in chest CT imaging.

The researchers have collected a dataset of 3506 patients with chest CT exams. After exclusion, 3,322 eligible patients were included for the model development and evaluation in this study.

The datasets were collected from six hospitals between August 2016 and February 2020, scientists refined the model using 4,356 exams from 3,322 patients.

The deep learning framework call COVNet developed by researchers were able to extract meaningful visual features from volumetric chest CT exams for the detection of COVID-19.

The COVNet framework consists of a RestNet50 as the backbone, which takes a series of CT slices as input and generates features for the corresponding slices.

The extracted features from all slices are then combined by a max-pooling operation. The final feature map is fed to a connected layer and the softmax activation function to generate a probability score for each type.

The COVID-19 Detection Neural Network—or COVNet for short—scored high marks, notching 90% sensitivity and 96% specificity for diagnosing coronavirus, the research published in Radiology.

“These results demonstrate that a machine learning approach using convolutional networks model has the ability to distinguish COVID-19 from community-acquired pneumonia,” concluded Lin Li, from the Department of Radiology at Wuhan Huangpi People’s Hospital in China, and colleagues.

Li and colleagues included hundreds of CT scans in the dataset, which displayed community-acquired pneumonia and other lung ailments. Their model also scored high marks in differentiating such diseases from novel coronavirus, with an 87% sensitivity rate and 92% specificity rate.

With radiologists cautioning about the overlap between coronavirus imaging findings and other lung issues, Li and colleagues believe AI can provide a useful assist to radiologists concerned about specificity.

“There is overlap in the chest CT imaging findings of all viral pneumonia with other chest diseases that encourages a multidisciplinary approach to the final diagnosis used for patient treatment,” they added.

More in AI

How China built facial recognition for people wearing masks

Researchers Use AI to Create Music through Proteins

Google developing a tool to help triage individuals for Covid-19 testing

Google Ph.D. fellowship programme: Indian students can apply by April 15

Leave a Reply

Your email address will not be published.