a burgundy zine

AI Joins the Fight Against COVID-19

Print Friendly, PDF & Email

By: burgundy bug

This transmission electron microscope image shows SARS-CoV-2, the virus that causes COVID-19, isolated from a patient in the U.S. Virus particles are shown emerging from the surface of cells cultured in the lab. The spikes on the outer edge of the virus particles give coronaviruses their name, crown-like.

Source: Novel SARS-CoV-2 | NIH Flickr

Last week, the National Institutes of Health announced that AI technology is underway to aid physicians across the United States in the global war against COVID-19.

These efforts are taking place with the Medical Imaging and Data Resource Center, which is comprised of research from physicists, physicians, radiologists, and scientists across the country.

Together, they’re developing AI algorithms to help physicians diagnose COVID-19 and develop “personalized therapies” for their patients. This technology will use medical imaging to detect the virus and utilize clinical data to support its findings, as well.

Recently, we spoke to NIBIB/NIH director of research Sciences Krishna Kandarpa, M.D., P.h.D. via email to learn more about the development of this technology and how it could impact the future of radiology.

Tell us a little about your background in medical research and your role with the MIDRC

I have a doctorate in engineering science with a bachelor’s degree in aerospace engineering. My area of research interest was fluid dynamics – how liquids and gases flow.

I am also a medical doctor who is board-certified in radiology and minimally invasive image-guided surgery (interventional radiology). My medical research was in the area of minimally invasive cardiovascular procedures.

I am currently the Director of Research Sciences at the NIBIB/NIH.

Tell us a little about the MIDRC and the researchers working with the center

The researchers consist of the nation’s leading physicists, physicians, and scientists from the Association of Physicists in Medicine (AAPM), the Radiological Society of North America (RSNA is an international society, in spite of its name), and the American College of Radiology (ACR).

The various researchers are scattered among leading academic medical centers across the United States.

How will AI be used for diagnosing and treating COVID-19? What can machine-learning algorithms reveal about the effects of the virus that can help physicians treat their patients?

AI algorithms will be used to detect the appearance of the disease on images of the lungs and heart and look for specific patterns as the COVID-19 progresses in severity.


The medical image information will be assessed along with symptoms and other clinical data at baseline, during treatment, and recovery.

Such information can inform the clinician on how best to manage an individual patient towards optimizing outcomes and will also be available for epidemiological studies.

What type of technology and data does it take to develop such advanced imaging tools?

It requires centers that have state-of-the-art computational facilities and physicists, engineers, data scientists, and physicians — especially radiologists — who have access to digital medical images.

It also requires those who know how to create artificial intelligence algorithms (computer programs) that can detect disease patterns within these images and can correlate these findings with other clinical information known about the patient.

This means having access to a large repository of images, including open access to various sites that have such information. It requires means of harmonizing the various sources of digital information for analysis while preserving patient privacy, as well.

Is the AI technology similar to anything that already exists? For example, will it screen medical imaging and detect features similarly to how facial recognition scans images and detects facial features?

Yes, there are technical similarities to ‘facial recognition’ algorithms.

But working with medical images would be tantamount to looking at a ‘face’ whose features constantly change in appearance as the disease progresses or responds to treatment.


The final diagnosis isn’t made from images alone but in conjunction with other clinical data.

How is the AI similar to or different from the AI technology used for diagnosing COVID-19 that was announced in other countries a few months ago?

Some of the AI algorithms could be similar. However, prior algorithms have been tested only at their specific sites on a small non-diverse set of images, and are not necessarily correlated with adjunct clinical information. Thus, they may not necessarily be universally applicable.

The MIDRC proposes to host a large repository of diverse images and clinical data from many sources, to test the universal applicability of these AI algorithms and facilitate means to regulatory approval so they can be made available to any US healthcare facility.

How long could it take to get this AI technology in hospitals and care facilities throughout the country?

Robust AI applications limited to specific sites are estimated to be made available within three to six months. But in order to be freely available to all facilities nationally, they have to pass regulatory muster, which could add a year or so.

What do we know so far about COVID-19 and other infectious diseases that will help develop this new technology?

The computational technologies themselves are independent of the disease. However, our growing knowledge of COVID-19 can help refine the algorithms and help in the management of future such pandemics.

The technologies that are developed can also be useful to detect and manage other diseases and organ systems.

How could these advancements in AI/radiology impact the treatment of other diseases in the future? For example, could this type of AI be used to analyze brain MRIs to diagnoses and predict the progression of Alzheimer’s disease? 

Yes, with respect to AD and brain diseases and more.

There is essentially no limit — AI applications to medical images can be made on any disease or organ system in the body where images show specific changes — such as with cancers, or neurological and metabolic ailments.


With there so much left to research and learn in medicine, and the rapid rate of technological advancement, is it possible that technology could reach a point where we’re learning more about infectious diseases from AI than we’re teaching the AI about the diseases — would the “student” at that point become the “teacher?”

Yes, anything is possible. If the question is whether AI can accelerate our knowledge of medical ailments and their treatment — there is every reason to be optimistic.

In the meantime, how can our readers continue to stay safe, stay healthy, and stay hopeful during these unprecedented times?

Follow the guidance posted by the CDC for COVID-19 precautionary measures.

Remember if it’s not just about you — your non-compliance with safe civic hygiene practices could result in some else’s death — don’t be selfish! 

Do you have any additional comments or final thoughts to share?

I would like to reiterate my previous answer — an ounce of prevention is worth more than a pound of cure — and thankfully there are many scientists working diligently on both. Go with the SCIENCE.

You can learn more about the NIBIB’s current research on COVID-19 here.

Donate to The Burgundy Zine

Interested in having content featured in an upcoming blog post or issue of The Burgundy Zine? Head on over to the submissions page!

For all other inquiries, please fulfill a contact form.

burgundy bug


A cynical optimist and mad scientist undercover, burgundy bug is the editor, graphic designer, webmaster, social media manager, and primary photographer for The Burgundy Zine. Entangled in a web of curiosity, burgundy bug’s work embodies a wide variety of topics including: neuroscience, psychology, ecology, biology, cannabis, reviews, fashion, entertainment, and politics. You can learn more about working with burgundy bug by visiting her portfolio website: burgundybug.com

View more posts from this author