Q&A: vRad’s CIO on AI Beyond Worklist Prioritization, New Solutions for Radiologists, COVID-19 and More

For the latest information on vRad’s Artificial Intelligence program please visit vrad.com/radiology-services/radiology-ai/

Originally published by Michael Walter as Q&A: vRad’s CIO on AI beyond worklist prioritization, new solutions for radiologists, COVID-19 and more on Radiology Business 

As 2020 comes to a close, radiologists find their profession at a major crossroads. AI and other game-changing technologies are rapidly evolving, government policies are forcing practices to rethink their business models, and a once-in-a-lifetime pandemic continues to cause chaos for the entire healthcare industry. 

With that wealth of opportunities and challenges in mind, Radiology Business spoke with Imad B. Nijim, the chief information officer at vRad, about what the future may hold for both his own company and the imaging industry as a whole. Nijim is a veteran of the healthcare technology space, spending considerable time focused on radiology, and has seen the industry undergo countless changes over the years.

Our full conversation can be read below:

AI technology continues to improve, impacting nearly every aspect of patient care. What is vRad working on right now when it comes to AI?

vRad is unique in that we’re a clinical practice with our own development team of software engineers. That helps us in two ways: we can build, implement and iterate in a very short amount of time; and we can focus our energies on the most clinically relevant technology. We’ve been working with AI for a few years now and I am proud to say that we have more than 10 different AI models in production supporting clinical and non-clinical workflows.

This year, a big area of focus has been clinical quality assurance (QA). This means leveraging AI to identify potential discrepancies between the radiology report and the images themselves. When I think about AI in radiology, I think about two things: the natural language processing (NLP) side of things, which is what the computer “reads” on the radiologist’s report, and image AI, which is what the computer “sees” on the images. We are bringing those two things together in a very clinical, actionable way. 

For example, right now we’re running an AI-enabled QA pilot program for aortic pathology cases that automatically compares the report and images to see if they agree or disagree. The results so far are very exciting, and our plan is to introduce models into production in 2021 that will warn the radiologist of a potential discrepancy on critical findings in time to impact a positive clinical outcome. 

What about the future? How might the company’s focus on AI change in the years ahead?

There is a future for radiology AI that I like to call “augmented radiology.” Everything we’re doing right now—the NLP, the image AI, the QA—is part of the march toward that point. Augmented radiology will work alongside the radiologist in areas of operations, governance, clinical oversight, FDA policies, ethical issues and more. Augmented radiology is not about building models, it is about building an AI-enabled organization.

Outside of AI, what is vRad the most excited to share with the world?

We have been developing a new mobile app that connects our radiologists to the broader care team, and we’re all very excited about that. The app enhances communication by putting a critical result alert or a signed report right in the ordering or referring physician’s pocket.

As you know, radiologists operate in a distributed reading environment, often supporting multiple care facilities using multiple systems. The referring physician environment is equally complicated in that it is often an entire care team.  Our goal is to connect those two groups in the most natural and common way: through their cell phone.

The mobile app is still in the pre-release stage at this point, but everyone who has used it is really happy with it. One surgeon actually said it was a game-changer, and I completely agree.

We’re looking forward to opening the app up for more widespread utilization in Q1, which should be a tremendous gain for both radiologists and referring physicians.

Let’s talk COVID-19. What kind of research has vRad done in that area?