In just two years, artificial intelligence has gone from a big idea in a small space to global adoption and integration
Not so long ago, a small stand stood in the corner of the floor at the Radiological Society of North America (RSNA) Scientific Assembly and Annual Meeting. The stand was showcasing the latest in artificial intelligence (AI) innovation and development within the radiology field and it garnered plenty of speculation and discussion. Did this small spot on an impressively large floor represent a fly-by-night idea that would be here one year and gone the next?
As the radiology community prepares for the 105th Scientific Assembly and Annual Meeting 2019, this question have been answered. AI has proven itself to be a solid and reliable partner in the radiology environment, providing the radiologist with additional tools and resources that support their roles and extremely busy caseloads. It has long since left behind the idea that it is merely an overhyped fad or brief splutter of emergent technology with little chance of success.
Today, it is seen as the team member that never sleeps, constantly sifting through the images to help the radiologist prioritize urgent cases and adjust workflows more effectively. AI has become embedded in medical practices and institutions, supporting the radiologist through its ability to prioritize, allocate and alert in high-pressure environments. It’s on a clear path towards a standard of care. And despite all of this – echoes of the AI hype still surround the AI showcase at RSNA. In fact, it’s becoming increasingly challenging to differentiate between those that are doing AI and those who are simply just saying it. This post will serve as a guide for navigating the AI Showcase at RSNA, offering you the essential criteria you need and the most important questions you should be asking any AI Vendor you meet.
Finding the right AI pathways
The AI showcase at the 2019 RSNA event has grown enormously since that first small stand. The annual meeting has recently announced that it will host an expanded AI space for attendees, one that is dedicated to AI solutions and services. It’s set to be an impressive showcase of AI that includes exhibitors, educators and researchers and that will allow for radiology professionals to immerse themselves in the vast and complex world of AI capabilities. Already more than 100 AI companies are set to become RSNA 2019 exhibitors on the 40, 000 sq. ft of exhibit space and this will be bolstered by vendor presentations, deep learning courses, emerging data technologies and networking opportunities.
The event mirrors the changing views of the radiology environment. Many professionals are interested in finding out more about how the technology works and how it can be used to leverage their skills and their growing workloads. This is reflected further in the addition of training courses and classrooms dedicated to explaining how AI can be embedded into workflows and how it can be used to supplement stressed teams.
The RSNA has also announced the launch of a new journal entitled Radiology: Artificial Intelligence that publishes AI-related scientific papers. Considering how AI paper submissions have increased since 2015, this move is hardly surprising – from none in 2015 to around 10% of all submissions in 2018.
Getting the most from the RSNA event
01: Navigate the noise
The sheer volume of exhibitors and solutions will be enough to make most people feel overwhelmed by choice and limited by time. To bypass this, establish exactly what value you would like to derive from any given AI platform and use this to determine which solutions most closely fit your parameters. This will whittle down the list of exhibitors and showcases significantly, and allow you to further refine your engagements so you can get actual value from the event. At the very least, you’ll need the RSNA 2019 floorplan and RSNA 2019 program before you start the event.
02. Establish the value
There are three stages or layers to establishing the value of a given AI solution. Each one introduces specific questions that allow you to fine-tune your approach and the vendors you engage with at the RSNA conference 2019. These layers are algorithmic, product and solution, and they will allow you to evaluate the AI RSNA 2019 exhibitors from an informed perspective, allowing you to assess them at each level of development.
03: The algorithmic layer
This layer asks that you examine the solution with the following questions in mind:
- What is the source of the data and what is the validation quality?
- Who annotated it – or where was it annotated – and can it do what it promises?
- What are the average accuracy levels across algorithms?
04: The product layer
At this layer, you should be asking the following questions:
- How do I use it? Does it ask that I change the way I work today?
- Will it work with existing systems?
- How long will it take to integrate into existing systems?
- Should there be concerns about patient data?
- Has anyone else used the system and what were their experiences?
- Regulatory clearance? FDA? CE?
The solution should integrate with existing workflows with minimal impact on work during implementation. It should also come with a proven track record so that you know that its track record is relevant.
05. The solution layer
At this point, the critical questions should be:
- What are the clinical outcomes and what metrics can I expect to evaluate?
- What value does this solution offer to me and the patient?
- Can the return on investment be actualized and how can the costs be justified?
06. Find partners that have solid reputations
Over the past few years, AI solutions have achieved essential accreditations and approvals from important healthcare organizations. These include the Therapeutic Goods of Australia (TGA), the US Food and Drug Administration (FDA), and the European Union CE. Aidoc is one of the market-leading providers of radiology solutions and has already obtained the CE mark for four algorithms and has FDA clearance in triage for cervical spine fractures, pulmonary embolism, and intracranial-hemorrhage. The company currently covers around 80% of CT scans and has moved into non-acute pathologies for CT such as oncology and the measurement and comparison of lesions.
Aidoc’s booth will be located in the North Hall Level 02, #10305 in the AI Showcase. Aidoc’s CEO, Elad Walach, will be speaking in the AI Theater on Monday, December 2nd at 2pm about how to find the right AI partner and ask the right integration questions to ensure that any AI implementation is the one that will deliver return on investment, solid patient care and value, and support the radiologist.
Seeing possibilities together
The situation has changed significantly as the 2019 event rolls around. It’s a glowing indictment of a technology that is moving beyond the hype, but it also highlights the need for caution. It is a complex environment that asks for meticulous attention to detail, absolute surety of diagnosis and rigorous standards of care. AI is a powerful tool that can become indispensable in helping to alleviate many of the issues that arise in data- and patient- overloaded healthcare environments, but it isn’t perfect. As more and more healthcare institutions adopt AI, it is building a foundation for AI as a standard of care and this standard has to be upheld and developed within absolute parameters.
As the RSNA event approaches, excitement about the potential of AI is immensely refreshing. It is paving the way for a dynamic community of physicians and innovators that will always advocate for the highest standards of patient care. The future is no longer tomorrow, it is already here and saving lives.