AI Talking Points: Radiologists and Vendors at RSNA 2019

I feel like this is the first RSNA where radiologists and teams can have useful conversations with AI vendors. The collective background AI knowledge of radiologists, technologists, department heads and informaticists is mature enough to get specific with proposals, sales cycles, installations and integrations. 

On the ACR's recent webinar titled: Evaluating and Deploying AI: Secrets to Success, over 50% of the participants either use AI at this time or intend to use AI in the next 12 months.

At the RSNA Artificial Intelligence spotlight conference in San Francisco, Elliot Siegel asked, “My institution is preparing for next year's acquisitions, what platforms should I look at it?”

Cleveland Clinic outlined their steps for AI adoption in the ACR Webinar as well. BAA, NDA, DUA, RFI, RFP are a few of the acronyms explained in the video replay starting at 41:13.

Vendors should expect specific questions from our radiology community, “How do you organize your BAA?” and “Are you prepared to respond to an RFP?” If a radiologist and team would like to see your platform in action, do you have a demonstration site? Better yet, do you have multiple demonstration avenues with different PACS vendors to accurately simulate our current workflow?

In addition, we as a group are getting much more sensitive to inherent biases in AI. This Aunt Minnie article summarizes some common problems when AI platforms scale and how explainable AI can be helpful. AI needs robust clinical evaluation in healthcare

Fellow radiologists, get your specific questions ready and I'll see you in the AI Pavilion.