At RSNA 2020, a Look at Artificial Intelligence—In a Virtual Context

Dec. 3, 2020
During the RSNA Annual Conference, being held virtually this year because of the COVID-19 pandemic, industry expert Joe Marion shared his perspectives on some of the solutions being offered on the exhibit floor

Joe Marion, Principal in the Waukesha, Wisconsin-based Healthcare Integration Strategies, LLC consulting firm, has attended 45 annual RSNA Conferences. The RSNA Annual Conference, sponsored by the Oak Brook, Illinois-based Radiological Society of North America, is normally held every year the week after Thanksgiving in the sprawling McCormick Place Convention Center in downtown Chicago.

This year, because of the COVID-19 pandemic, RSNA20 is being held virtually on the organization’s website. Still, the association managed to offer hundreds of sessions, as well as a virtual exhibit floor with hundreds of vendors presenting thousands of solutions.

Marion spoke on Thursday, day four of RSNA20, with Healthcare Innovation Editor-in-Chief Mark Hagland, on some of the key technological trends he’s been seeing at the virtual conference.

So, this year obviously is different, given that it’s a virtual conference. Has that felt different to you?

It definitely is feeling different! The stress level is much lower, because if you have different meetings, you’re sitting in front of your computer, and you don’t have to walk from the South Building to the North Building. It’s much less stressful.

What strikes you the most about this RSNA?

The most obvious is the fact that artificial intelligence is for real. I think that most of the major vendors are all speaking to some sort of AI platform. And from what I’ve heard thus far, the one that really strikes me the most is GE’s Edison platform, primarily because they’ve done a better job in the context of soliciting and securing third-party applications. They have something like 250 applications in their app store. So it’s not like this is one or two things that people are playing with. It’s becoming mainstream now. And I think that COVID is one of the drivers or accelerators for making this happen.

Because people are needing more tools for working remotely?

Yes, and the long assessment modules and applications that have come to light, people are starting to use those. And they saw COVID as providing an opportunity for using those tools for diagnostics. So that’s probably the biggest area. And in the past, it’s been more smoke and mirrors, or future stuff. But the clever thing that GE is doing is that they’re offering this on a subscription basis. That’s smart, because it provides a vehicle. It’s like the Android or Apple play stores. You can choose what you want.

Are you primarily talking about diagnostics?

That’s the diagnostic side of the application. The other side of it where AI is greatly expanding is in the area of what’s being called workflow orchestration. Siemens, which I’m going to get a demo on tomorrow, has taken leadership in that area, probably based on an outgrowth of their previous Medicalis work. They’re referring to it as Syngo Carbon. And how they describe it, is as a mechanism or vehicle with which to bring everything together—images and supporting information. I don’t think it’s necessarily unique, it’s the kind of things that McKesson had been working on before they became Change—where they were mining information out of the EHR and making that available to the diagnostician. So this appears to be similar. And Rik Primo describes it as what he and I were talking about several years ago: the deficiency of the EHR, in that the EHR does not present information in the way that the clinician needs to see it. For example, in Epic, there’s a radiology tab, a cardiology tab, a pathology tab, etc. The problem is that that’s not the way a physician works. There’s a disease or symptom process that they’re following that involves a timed sequence of events that have occurred that they want to review. And I think there’s a huge opportunity for the imaging companies to in essence fill that gap in the EHR. And other companies—Fuji comes to mind—they can launch the EHR from their PACS workstations. But Siemens is now in essence using AI to intelligently review information, so that’s a step up. And I’m not sure that Siemens is the only one doing that, but they call it the connecting element that gives rise to knowledge. So again, I think it’s a trend that we’ll see in the context of what people are doing to bring more than just the images to the user interface.

So prioritization of studies is a part of that workflow orchestration phenomenon, correct?

Yes. Siemens was calling it something like unusual or critical findings. So you’d be presented with a worklist, and offering order of prioritization. And it provides the benefit of the integration of information; so instead of the clinician themselves having to invest the time and effort to explore and find those elements to decide what to do, it’s the system bringing that together intelligently, based on rule sets.

That will be a significant trend. The AI diagnostics tools will be important, but I still feel that the AI workflow aspect has just as much credibility and value as does the image guidance aspect. The third element, I think, is the whole question of analytics. And they all seem to be working in the context of analytics or the use of that data. And Philips, Siemens, and Ge are all going down a path where the integration of analytics both for the administer and the clinician, are being made smarter, in terms of the integration of operations. One demo I saw highlighted, related to how quickly you could break down which radiologist was the most productive, or which machine or area, or whatever, is not experiencing the throughput you’d expect; and it’s all graphically presented. So this is the convergence of the visualization with the AI rule sets to intelligently mine the data for management and clinical purposes.

Is all of this moving forward at about the pace you’d expected it to? Has it been moving more quickly than you’d expected? More slowly?

Compared to last year, there appear to have been significant strides made. It’s hard to assess whether these are concrete applications or vaporware. The stuff is coming. And part of what’s happening during this virtual RSNA is allowing people to focus. If I walk into an exhibit hall, how do I know where to look? So in this virtual platform, they can be more guided in how where they ask you to look. And when you click on the “Virtual Exhibits” button, nearly all the vendors have a virtual display of their exhibit. And they present it as, here’s the layout, and which area do you want to explore. But back to your original question, the impression I get is that the vendors are really starting to make some headway and are focusing on all these things: the AI from an application perspective, the workflow orchestration, and the analytics, are all moving forward. And it’s pretty obvious that they are all continuing to make strides in terms of emphasizing an enterprise level of PACS. And so even though it’s a radiology meeting, many are emphasizing that it really is an enterprise focus.

The fourth element that’s really big in this area is the notion of cloud-based applications. And Change, I think, has been the most obvious about this, in terms of moving from what one could call a dedicated, standalone type of application, to a cloud-based environment. And they speak specifically to cloud-native applications. So it’s one thing to say, I’ve got fewer applications, and I can host through Amazon Web Services or Azure or whoever, but it’s another to say, I’ve actually written an application to be a cloud-based application. A lot of people are making that distinction; and McKesson s really touting that as the future. Philips is also pretty adamant that that’s the direction. The hard part is that there are a lot of old faces that are no longer part of the Philips organization. So, is it now that the Philips people are still in charge? Or have some of the Carestream people dug into that, and are directing strategy?

What does the business landscape look like for vendors, in the next year or two?

The most obvious aspect is that they see the consolidation of healthcare providers. And it’s not like you walk into St. Clare’s Hospital and sell them a CT machine. Now, you’re dealing with an entire health system, and you’re trying to sell them on the notion of managed use and beyond. Clearly, Philips has taken the lead in this, but Siemens and GE are not far behind, in the context of selling a system-wide solution, and a long-term contract. And that’s very much in the vein of the move towards cloud-based architecture; so there are a lot more things they can support in a cloud-based environment. And in that regard, it’s much easier to offer solutions that are ongoing for a fee, as opposed to buying a capital-based product that I buy this year, and I’m not going to change for three years.

One more item that caught my attention is that a couple of my vendors: a couple of them are speaking to the transarterial valve replacement—so instead of cracking the chest and going in and cutting into the heart and replacing a valve, instead, you go in through the groin and push a catheter into the exact spot over the valve, and expand a new valve into place. So the majority of valve replacement being done today, is being done via that process as opposed to actual heart surgery. Someone in their 70s is not a strong heart surgery candidate, for obvious reasons; and this can provide better functionality, and better quality of life, at potentially a lower cost. So that’s the cardiology side of it; but what’s interesting, and I need to ask a few vendors why they’re highlighting that at a radiology meeting. But in the context of whether they’re deciding whether a patient is a good candidate—they’re saying, here’s a way of planning that valve replacement, using imaging. So it strikes me as interesting that it’s being highlighted at a radiology meeting. But then again, people at Agfa are really describing their platform as an enterprise platform that does radiology, cardiology, pathology, or whatever.

What do you think will happen in the next twelve months?

The biggest factor from my perspective will be how the virtual meeting will have been received. Did people find it more useful than trekking to Chicago at the end of November and beginning of December? It could represent a negative for the city of Chicago; but for a lot of these meetings, are we entering a new, virtual age? Otherwise, we’ll see more of the same, and more demonstration of the AI-type technology, both for diagnostics and for workflow.

Sponsored Recommendations

Care Access Made Easy: A Guide to Digital Self-Service for MEDITECH Hospitals

Today’s consumers expect access to digital self-service capabilities at multiple points during their journey to accessing care. While oftentimes organizations view digital transformatio...

Going Beyond the Smart Room: Empowering Nursing & Clinical Staff with Ambient Technology, Observation, and Documentation

Discover how ambient AI technology is revolutionizing nursing workflows and empowering clinical staff at scale. Learn about how Orlando Health implemented innovative strategies...

Enabling efficiencies in patient care and healthcare operations

Labor shortages. Burnout. Gaps in access to care. The healthcare industry has rising patient, caregiver and stakeholder expectations around customer experiences, increasing the...

Findings on the Healthcare Industry’s Lag to Adopt Technologies to Improve Data Management and Patient Care

Join us for this April 30th webinar to learn about 2024's State of the Market Report: New Challenges in Health Data Management.