RSNA 2019 Takeaway #1: The Benefits & Limitations Of AI Are Becoming More Clear

The past few weeks have been a whirlwind, with the excitement of the Thanksgiving holiday followed by the even more exciting hustle and bustle of one of our favorite meetings: RSNA. From the vendors to the speakers to the broad variety of topics that touch on every aspect of radiology, there’s always a lot to love about the RSNA meeting, and inevitably a lot to learn. This year was no exception, and we couldn’t help but walk away from RSNA 2019 excited about the future of radiology.

Changes are happening across every facet of the industry, but unsurprisingly it was the discussions related to innovation within Computed Tomography that we were the most excited to be a part of. It would take countless hours and likely hundreds of pages to detail everything from RSNA19, but there were a few key takeaways we want to highlight from the conference over the next few days. First and foremost is an update on AI.  

It wouldn’t be a 2019 conference of note if deep learning and AI weren’t top points of conversation! On Monday Dr. Soonmee Cha, a veteran neuroradiologist and program director at the University of California San Francisco, gave an inspired presentation about the clear benefits and limitations of AI within radiology. 

She noted that there are a number of key tasks radiologists and trainees perform that the technology simply cannot replace. For instance, AI can’t sign final reports or attend meetings, two things that are crucial in today’s radiology landscape. “AI will not replace radiologists or replace radiology trainees,” Cha noted, “And it won’t cut costs. Maybe in the long run, but initially it will actually drive up costs.”

So how can AI improve radiology and aid radiologists in their daily lives?

Cha argued that perhaps the most important way AI has the potential to help radiologists is by battling one of their biggest issues: increasing workloads. Having the flexibility to leave the reading room and speak with the very patients they are tasked with treating could be crucial to decreasing burnout and exhaustion in radiologists, who suffer at some of the highest rates out of any specialty in the medical sphere, and it’s very likely that AI could make that possible by managing other parts of the workload. In addition to that, AI has huge potential to improve image quality, decrease acquisition times, eliminate artifacts, improve patient communication, and even decrease radiation dose. 

The 2019 conference made it clear that the fear-mongering stage of AI discussions is over. The radiology industry is starting to get excited about the potential benefits AI can offer and is slowly but surely beginning to embrace the technology as it is further honed and refined for clinical use. There’s still a lot of work to be done, but if there was one resounding point of emphasis at RSNA this year, it’s that the future looks bright for AI.