Filter by input type
Select all
News
Pages
Events
Filter by category
Select all
AI ANALYTICS
Mobile Apps and Internet of Things
Advancement of science
big data
Connected communities
Coronavirus
Courses and training
DIAGNOSIS
Initial Editorial
Editorials
A world in the cloud
Events
Infographics
Artificial Intelligence and Science
IoT Apps
News
Digital platforms
Social networks
Review of scientific publications
Course Summary
Synopsis of essay
Overview of reference frames
Synopsis of recent publications
Use of Digital Platforms
Filter by input type
Select all
News
Pages
Events
Filter by category
Select all
AI ANALYTICS
Mobile Apps and Internet of Things
Advancement of science
big data
Connected communities
Coronavirus
Courses and training
DIAGNOSIS
Initial Editorial
Editorials
A world in the cloud
Events
Infographics
Artificial Intelligence and Science
IoT Apps
News
Digital platforms
Social networks
Review of scientific publications
Course Summary
Synopsis of essay
Overview of reference frames
Synopsis of recent publications
Use of Digital Platforms
Deep learning model diagnoses COVID-19 through CT scans

CovidCTNet, is an open-source deep learning-based platform that is capable of diagnosing COVID-19 using small CT scan images.

Timely and accurate diagnosis of COVID-19 is crucial to reduce not only the spread of the disease but also mortality. The most widely used conventional tests in the world for the detection of COVID-19 are polymerase chain reaction or PCR, one of the most reliable tests. However, its detection accuracy reaches 70 to 75%. Another way to know or make a diagnosis of a person for COVID-19, is by computed tomography (CT) images, these images have a higher sensitivity of almost 98%, however, the accuracy is 70%.

Medical researchers from universities in Iran, the United States, and Vietnam developed “an open-source framework, CovidCTNet, composed of a set of deep learning algorithms that accurately differentiates Covid-19 from community-acquired pneumonia (CAP) and other lung diseases.” This framework increases accuracy markedly in CT image detection, reaching 95%, far surpassing traditional radiology work that achieves 70% accuracy.

“CovidCTNet is designed to work with heterogeneous and small sample sizes independent of the CT imaging hardware.” This innovation aims to facilitate the detection of COVID-19 worldwide and also facilitate the work of radiologists in this process, which is why the developed algorithms as well as the model parameters are available as an open-source model.

“Open-source sharing of CovidCTNet enables developers to rapidly improve and optimize services while preserving user privacy and data ownership,” the authors explain in the research published in Nature.

In its results CovidCTNet, evaluated a data set of 20 mixed control cases, first using the developed algorithm and simultaneously by four independent radiologists. " The average reader performance of four radiologists showed a sensitivity of 79% for Covid-19 and specificity of 82.14%. The CNN classification of CovidCTNet, however outperformed the radiologists and achieved Covid-19 detection with sensitivity and specificity of 93 and 100%, respectively details the comparison of radiologist performance versus CovidCTNet,” the study explains.

Nevertheless, this technology is not intended to replace the work of radiologists, but rather to streamline their work, and carry out better diagnoses. 

 

The complete research is freely available in the journal Nature https://www.nature.com/articles/s41746-021-00399-3

Outstanding news

News by country

Share

Digital Health in the world

  • — Science Brief: Omicron (B.1.1.529) Variant/CDC updates
    See more
  • —Coronavirus resource center/Johns Hopkins
    See more
  • — Epidemiological tracing of COVID-19 contacts / Johns Hopkins Course
    See more
  • — SARS-CoV-2 infection behavior / FCS calculator
    See more
  • — Omicron SARS-CoV-2 variant: a new chapter in the COVID-19 pandemic/ Article The Lancet
    See more
  • —Genomic Epidemiology Tracker/GISAID
    See more
  • — Mexican Genomic Surveillance Consortium
    See more
Secured By miniOrange