Using AI and Ultrasound to Diagnose COVID-19 Faster

Posted on August 11, 2020 by aium1952

Coronavirus disease 2019 (COVID-19) is a newly identified virus that has caused a recent outbreak of respiratory illnesses starting from an isolated event to a global pandemic. As of July 2020, there are over 2.8 million confirmed COVID-19 cases in the U.S. and over 11.4 million worldwide. In the United States alone, over 130,000 Americans have died from COVID-19, with no end in sight. A major cause of this rapid and seemingly endless expansion can be traced back to the inefficiency and shortage of testing kits that offer accurate results in a timely manner. The lack of optimized tools necessary for rapid mass testing produces a ripple effect that includes the health of your loved ones, jobs, education, and on the national level, a country’s Gross Domestic Product (GDP), but artificial intelligence and ultrasound may help.

STATE OF ART IN DIAGNOSIS

Currently, there are two types of tests that are conducted by healthcare professionals–diagnostic tests and antibody tests. The diagnostic test, as the name implies, helps diagnose an active coronavirus infection in a patient. The ideal diagnostic test and the “gold standard” according to the United States Center for Disease Control (CDC) is the Reverse Transcription Polymerase Chain Reaction, or simply, RT-PCR. RT-PCR is a molecular test not only capable of diagnosing an active coronavirus infection, but it can also indicate whether the patient has ever had COVID-19 or were infected with the coronavirus in the past. However, the time required to conduct the test limits its effectiveness when mass deployed.

A much faster but less reliable diagnostic test alternative to RT-PCR is an antigen test. Much like the gold standard, the antigen test is capable of detecting an active coronavirus infection in a much shorter timeframe. Although antigen tests produce rapid results, usually in about an hour, the results are deemed highly unreliable, especially with patients who were tested negative according to the US FDA.

In contrast, the antibody test is designed to search for antibodies produced by the immune system of a patient in response to the virus and is limited by its ability to only detect past infections, which is less than ideal to prevent an ongoing pandemic.

THE PROBLEM

To combat the rapid expansion of an airborne virus such as COVID-19, or future variations of a similar virus, rapid and reliable solutions must be developed that aim at improving the limitations of current methods. Although highly accurate, methods such as RT-PCR do not meet the speed requirements needed for testing on a large scale. Depending on the location, diagnosis of an active coronavirus infection with RT-PCR may take anywhere between several hours and up to a week. When the number of daily human-to-human interactions are considered, the lack of speed in diagnosing an active coronavirus patient could be the difference between a pandemic or an isolated local event.

As an alternative to molecular tests, Computed Tomography (CT) scans of a patient’s chest have shown promising results in detecting an infection. However, in addition to not being recommended by the CDC to diagnose COVID-19 patients, there are many unwanted consequences with the use of CT scans. With CT scans used to diagnose multiple illnesses, some of which relate to serious emergencies such as brain hemorrhaging, they cannot be used as the primary tool for diagnosing COVID-19. This is especially true in rural areas where the healthcare infrastructure is underfunded. Mainly due to the required deep cleaning of the machine and room after each patient, which usually requires 60 to 120 minutes, many institutions are unable to provide CT scans as a viable primary diagnostic tool. Ultimately, given the need for CT scanners for several other health complications combined with limited patient capacity at each hospital, alternative methods must be developed to diagnose an active coronavirus patient.

THE SOLUTION

Recently Point-of-Care (POC) devices have started to be adopted by many healthcare professionals due to its reliability and portability. An emerging popular technique, which adopts improvements made in mobile ultrasound technology, allows for healthcare professionals to conduct rapid screenings on a large scale.

Working since mid-March, when early cases of physicians adopting mobile ultrasound technology emerged, the research team at The Ohio State University, Dr. Alper Yilmaz and PhD student Shehan Perera, started developing a solution that can automate an already well-established process. Dr. Yilmaz is the director of the Photogrammetric Computer Vision lab at Ohio State. Dr. Yilmaz’s expertise in machine learning, artificial intelligence, and computer vision combined with the research experience of Shehan Perera laid a strong foundation to tackle the problem at hand. As it stands, the screening of a new patient, with the use of a mobile ultrasound device takes about 13 minutes, with the caveat that it requires a highly trained professional to interpret the results generated by the device. With the combination of deep learning and computer vision, the research team was able to use data generated from the ultrasound device to accurately identify COVID-19 cases. The current network architecture, which is the product of many iterations, is capable of detecting the presence of the virus in a patient with a high level of accuracy.

Many fields have been revolutionized with modern deep learning and computer vision technologies. With the methods developed by the research team, this technology can now allow any untrained worker to use a handheld ultrasound device, and still be able to provide a service that rivals that of a highly trained doctor. In addition to being extremely accurate, the automated detection and diagnosis process takes less than 10 minutes, which includes scanning time, and sanitation is as simple as removing a plastic seal that covers the device. The benefits of this technology can not only be useful for countries such as the United States, with a well-established healthcare system, but, more importantly, can significantly help countries and areas where medical expertise is rare.

CONCLUSION

The United States healthcare system is among the best in the world, yet we are failing to provide the necessary treatment patients clearly need. The developments made in artificial intelligence, deep learning, and computer vision offer proven benefits, which can not only be leveraged to improve the current state of the global pandemic but can lay the foundation to prevent the next. Alternative testing methods such as mobile ultrasound devices combined with novel artificial intelligence algorithms that allow for mass production, distribution, and testing could be the innovation that could help decelerate the spread of the virus, reducing the strain on the global healthcare infrastructure.

Feel Free to Reach the Authors at:

Photogrammetric Computer Vision Lab – https://pcvlab.engineering.osu.edu/
Dr. Alper Yilmaz, PhD
Email: Yilmaz.15@osu.du
LinkedIn: https://www.linkedin.com/in/alper-yilmaz

Shehan Perera
Email: Perera.27@osu.edu
LinkedIn: https://www.linkedin.com/in/shehanp/

References

https://www.fda.gov/consumers/consumer-updates/coronavirus-testing-basics

https://www.whitehouse.gov/articles/depth-look-COVID-19s-early-effects-consumer-spending-gdp/#:~:text=BEA%20estimates%20that%20real%20GDP,first%20decline%20in%20six%20years.&text=This%20drop%20in%20GDP%20serves,in%20response%20to%20COVID%2D19.

Interested in learning more about COVID-19 or AI? Check out the following posts from the Scan:

The Place of POCUS in Prevention of Physician Burnout

Posted on September 10, 2019 by aium1952

Doctors’ jobs, in the hospital or clinic, have been getting more demanding and less rewarding in the last several years. Well-meaning changes including the rise of electronic medical records and attempts to improve how we do our jobs through quality measures have made us sad and tired and supply none of the joy that we can get from a satisfied patient or a diagnostic puzzle cleverly solved. We may find ourselves aging, with multiplying frown lines and receding hairlines, sitting at our computers finishing our documentation, while our families have vacations and parties without us. Although we make enough money, strangely, it doesn’t buy happiness.

When we are tired and sad; we lack the creativity to make job changes. Fear eclipses courage.

Sometimes we do stupid things involving alcohol or indiscretions, or buying something expensive on credit… family members give us “that look.”

We feel inadequate.

We get grumpy and stop doing that extra little bit to connect with the patient or unravel the mysterious illness. The precious little job satisfactions of working well with our team or taking our patients’ point of view become rarer.

We are burning out. There’s that telltale smell of smoke as our soul shrivels and our dreams fry.

What do we need? Probably a vacation, maybe even a stint working in global medicine, to change our perspective. Counseling and confiding in friends can help. If we keep doing the same job, perhaps we need a scribe to take care of the paperwork. Also learning a new skill could make us wake up and love medicine again. Enter point-of-care ultrasound.

I don’t want to trivialize the pain of burnout. It can be devastating, making us depressed, ending marriages, wrecking careers and friendship, collapsing us inward, and sometimes leading to suicide. Somehow we need to jump off of that horrific course and better sooner than later. I got close to burning out early in my career and ever since that time I’ve done everything I can to stay in love with my job. For me, learning to do point-of-care ultrasound enriched my practice and, along with a major career adjustment, kept me from getting all charred and crispy.

Doing point-of-care ultrasound, for a physician who is already skilled in practice but has no ultrasound experience, can be life-altering. As I matured in my practice, some of my physical exam skills improved but others atrophied for lack of use and because I knew that I couldn’t trust them. A fluid wave doesn’t predict ascites. Dullness in the base of the lung doesn’t lead me to suspect a pleural effusion. Splenomegaly, if not massive, is so hard to detect in my super-adequately nourished patients. Learning basic point-of-care ultrasound brought me back to paying good attention to my patients’ bodies. And they were fascinated and appreciated the extra care. I also was able to more quickly solve their medical mysteries and shorten previously prolonged evaluations. Seeing patients got more fun.

Burnout is an awful feeling and is preventable. It happens when we get ourselves into situations that are not sustainable and don’t feed our souls. We physicians have vast options and we need to recognize when we are trying to do a job that is wrong for us. And before we quit the profession entirely, we need to try learning something that makes it fun again. Point-of-care ultrasound, for instance.

How do you avoid burnout? Do you have your own experience to share? Comment below, or, AIUM members, continue the conversation on Connect, the AIUM’s online community.

Janice Boughton MD, FACP, RDMS, is an internist Moscow, Idaho. She practices hospital medicine and rural primary care as well as teaching point of care ultrasound techniques in the US and Africa. She also writes about healthcare economics in her blog (www.whyisamericanhealthcaresoexpensive.blogspot.com.)

Dr. Boughton graduated from the Johns Hopkins School of Medicine in 1986 and completed residency training at the Johns Hopkins Hospital and the University of Washington. She started doing bedside ultrasound in 2011.

Artificial Intelligence and Point-of-Care Ultrasound

Posted on May 21, 2019 by aium1952

One of the greatest ongoing challenges of POCUS (point-of-care ultrasound) is educating existing physicians, residents, students and others. There are not even enough teachers to teach everyone who wants to learn. Clinicians would like to get the results from POCUS performed on their patients but have difficulty investing the effort required to learn, practice, and then become credentialed. Further complicating things for some is the dreaded self-doubting period, which could last months or years, where providers worry they may make a mistake and be ridiculed for it, or worse. Blaivas

One potential answer is thought to be artificial intelligence (AI); kind of like it seems to be for everything in medicine today. What good is AI in POCUS anyway? What if the education required was simply to find the correct spot on the body to apply the probe? Then the algorithm would do the rest and it would be more accurate than the best POCUS masters. Not only would training be truly minimized, maybe to minutes, but the examination would be shortened as well. A few sweeps through organs, whether it is the liver and gallbladder or the heart, may be enough for the AI algorithm to do its thing. This would mean all those busy clinicians really would get a great return on their time investment. If the algorithm is that accurate and expert, providers will not be questioned easily when they document an AI US finding.

AI is an inescapable topic of sensational news stories and movies alike. AI is simply a machine approximation of human-like intelligence in task performance. The type most associated with image interpretation is deep learning. How does it work? Programmers develop software architectures roughly resembling levels of neurons in the cerebral cortex, with multiple connections. The levels of neurons have specific functions and transmit messages to neurons in the next row via mathematical functions. They are also capable of sending messages in reverse as feedback. Such a deep network is often termed a convolutional neural network (CNN; or some variant on the name). It can learn to interpret images, whether CXR, head CT, or ultrasound, by scanning each image one tiny part at a time, then pooling all of the neuronal-like reactions to those tiny parts and coming up with an answer. Give it enough training data and such a CNN can become very accurate.

Well, imagine a CNN algorithm plugged into your favorite POCUS machine. The CNN is trained on the liver and gallbladder; it has seen millions of example images, both normal and abnormal. It can recognize liver anatomy and point it out for you, the same for every detail around the gallbladder and biliary tree. It’s great at identifying pathology and can make measurements in the correct spots for the wall, common bile duct (CBD), and more. Once again, who really cares? I spent 2 decades scanning the gallbladder, performing research studies, and publishing on it. Well, while it may not have been an issue for me, not everyone invests their free time like that. Yet, many would like to be able to put a probe on the abdomen, have the ultrasound machine tell them where to move it, point out pathology, and come up with a likely diagnosis. Did I mention it could happen in real time, at the patient’s bedside, while you are casually speaking to them? How useful would this be? It could substitute for years of training, maybe even over 2 decades worth. There are other subtle benefits too. Although some studies seem to show that CNN CT algorithms seem to catch so much pathology radiologists can miss, the individual CNN may not be as good at finding something a rare expert might pick up, at least for now. But the CNN never gets tired. It never gets hit with a massive wave of scans to read late at night or overwhelmed with clinicians calling to discuss imaging studies. Thus, even experts can benefit from such algorithms as an aid.

Not happy with the image quality due to patient body habitus or another factor? It turns out another algorithm can actually artificially improve the image clarity and quality, and do so accurately without introducing false data. This has not been introduced into clinical use of POCUS but is likely to be just around the corner. The key is to make sure nothing is invented by the algorithm that is not actually there.

Imagine incredible ultrasound expertise from a short exam that required minimal training to perform. This scenario will come, but not this year or the next. As some speakers and authors have noted, AI coupled with POCUS is a big step toward the fabled and elusive “tricorder” first depicted in the 1960s Star Trek television series. An incredible hand-held device (that does not even require body contact), which diagnoses maladies in a few short sweeps over the patient. The eventual outcome of approaching such a device is greatly increased speed, efficiency, confidence, and accuracy of patient assessment and diagnosis. The benefit of significantly decreased skill/training requirements will also pose some challenges for the workforce, but these are likely to appear gradually and may be hardly noted.

What about combining other data feeds along with the ultrasound images? AI algorithms are great at interpreting EKG tracings and even cardiac and lung auscultation. Studies analyzing digital auscultation signals using deep learning systems are able to diagnose many more abnormalities than humans are. The result could be synergistic and add redundancy in diagnosis, such as for abnormal lung or heart sounds during ultrasound evaluation. Maybe other signals could be incorporated also.

These algorithms just need data, lots of data, and that is the conundrum for people seeking to develop AI apps. What do you think about companies getting de-identified image data without provider and patient awareness? Do you think it would help you to have a smart machine that analyzed the images and made calculations within seconds? What about incorporating other diagnostic signals such as digital auscultation, EKG tracings, or maybe some other signal?

Share your thoughts on AI in ultrasound: comment below, or, AIUM members, continue the conversation on Connect, the AIUM’s online community.

Michael Blaivas, MD, MBA, FACEP, FAIUM, is an Affiliate Professor of Medicine in the Department of Medicine at the University of South Carolina, School of Medicine. He works in the Department of Emergency Medicine at St. Francis Hospital in Columbus, Georgia.

The Scan

The official blog of the AIUM

Tag Archives: POC