What do Radiology heads and leaders have to say about AI?
As consultants to some of the largest MedTech players, we always think in the shoes of the Radiology leaders. What do they think? What do they need? Are they currently getting it? Is there something they want to see out there but is currently unavailable? Is AI truly as useful and practical as companies say? Are there any hurdles? Do they feel threatened by it? To answer these and many more questions, we had long talks with some of the key note speakers during the event.
TABLE OF CONTENTS
It is no surprise to anyone working in MedTech today that Artificial intelligence (AI) and Machine learning (ML) solutions are by no question sweeping the industry. At this year’s Radiological Society of North America convention, the topics Artificial intelligence and Machine learning, the capabilities and applications, and how these can be integrated into Radiologists’ daily operations were a major buzz. Hundreds of booths, both by many start-ups and the established players, showcase events, educational sessions, even one-to-one conversation around the halls – AI was a topic of high interest, raising many questions, opinions and concerns.

All companies, regardless of whether they were just entering the market, or they had been in the industry for decades, wanted to showcase their AI and ML solutions. The conversation spanned across different stages of the Radiologist journey and how different applications help Radiologists shave off valuable seconds – be it from scheduling, reading and diagnosis, reporting or follow-ups.
While all this excitement, buzz and flow of information is useful, one burning question lingered in our heads. Namely, what do Radiology heads and leaders have to say about AI?
As consultants to some of the largest MedTech players, we always think in the shoes of the Radiology leaders. What do they think? What do they need? Are they currently getting it? Is there something they want to see out there but is currently unavailable? Is AI truly as useful and practical as companies say? Are there any hurdles? Do they feel threatened by it? To answer these and many more questions, we had long talks with some of the key note speakers during the event.
The perspective of Radiology heads and leaders on AI and ML revolves around five main pillars
- Bright future
Radiologists are aware and appreciate all the progress being made, especially in the past year. In a single year alone, the competition and consequential innovation in AI and ML has grown significantly. This is especially fuelled by smaller start-ups, which provide niche solutions to specific problems. Having that said Radiologists think Creative destruction might not be enough, and point out they would like to see more collaboration between players, especially the big medical device giants. The IT non-medical giants entering the field might especially benefit by doing so, as some physicians perceive them as “not medical and fake knowledge”.
- A tool in their kit
Radiologists appreciate AI as one of their next tools, not as a substitute. It helps them focus their time, prioritize and get rid of the burden of monotonous, simple tasks which may actually take away their concentration. However, there is still much to be done, like learning how to use it and work with it. Most importantly, Radiologists need to know how they can shape it themselves and not get overwhelmed by fad promises. “Embrace it. If you don’t use it, it will use you.”
- Personalised precision for the mass cases
The computational power of AI is indisputable. The improvements it brings in workflows, workloads and detection/ diagnosis can help deliver more personalised and precise treatment. However, while this notion holds true for mass cases, Radiologists fear AI may not be up to the task for the rare occasions. Such cases might present themselves three or four times in a lifetime, therefore lacking the big data required for AI to read them successfully.
- Feeding AI with the right cases
Many practices, especially smaller ones, are struggling with the time resource to label and report their Radiology data, which is a must for AI to come into play. This presents a struggle for a practical widespread application of AI. The number of cases one needs to go through to become an expert Radiologist is in the hundreds, while for a secure, working AI model – it is in the hundred thousand. The company that unites multiple medical facilities into one consortium that supplies data is a leader. “The guy who owns the data, rules the game.”
- Matching the flexibility of the brain
Even more so, transferring knowledge from AI to people is still seen as challenging. As while medicine is an exact science, learning is not. Will AI ever be able to fill the room and the people, make appropriate jokes, know exactly how to portray the material so that people understand and obtain it? Radiologists believe this is unlikely.
How can vendors resonate better with Radiology leaders and support the commercialization of their solutions?
By adopting the three principles outlined below, we believe vendors can be more powerful in their quest for recognition, adoption and ultimately mass penetration of their solutions, starting with convincing the Radiology leaders and then moving along the entire Radiology ecosystem.
- Transparent & clinically proven solutions
Vendors should take advantage of the significant growth in Radiology leaders’ willingness to test and deploy AI applications. Solutions that transparently and convincingly address the issues of rare cases, data labelling and knowledge transfer will be powerful.
- Seamless integration and interoperability
Radiology heads also expect vendors’ AI solutions to be seamlessly integrated in their workflow. Positioning oneself as a new, hassle-free, interoperable tool in the Radiologists kit will allow widespread adoption of vendors’ solution and commercial success.
- Collaborations with facilities & across the industry
Vendors should initiative and establish broader collaborations with medical facilities of all sizes, start-ups and even competitors in particular application areas (e.g. lung cancer AI cluster). For one thing, this move will help gain more data and provide the clinical proof Radiologists need, especially when it comes to rarer cases. But ultimately, it will also raise much needed trustworthiness, conviction and an image of serving a greater purpose beyond profit – currently lacking such vendors in the eyes of Radiology leaders.
Healthcare & Medical Device Manufacturing
The Customer Experience Firestarter
The e-guide how to make sure everyone in your organization is committed to great customer experience. Adapted for the healthcare industry based on our experience with the world’s leaders in imaging equipment and respiratory care.
