AI in Neuroradiology: Keys to Workflow Integration, Governance, and Facilitating Improved Outcomes
In a recent interview, Vivek Singh, MD, discussed the practical realities of integrating AI into neuroradiology workflows, addressing challenges and concerns with implementation and potential automation bias, and emphasizing the importance of appropriate governance with AI tools and measurable outcomes in stroke imaging.
Many early AI tools failed not because they lacked accuracy, but because they didn't fit into existing radiology workflows, posited neuroradiologist Vivek Singh, MD, in a recent interview with Diagnostic Imaging.
"For neuroradiology specifically, you need something that is not only going to integrate with the workflow but is also going to actually change your outcomes or do something measurable that is actually going to provide some value,” maintained Dr. Singh, who is affiliated with the Medical University of South Carolina in Charleston, S.C.
For successful deployment and integration of AI software, Dr. Singh emphasized reliability across diverse patient populations as well as vendor support and ongoing data monitoring of the software. He also emphasized the importance of multidisciplinary buy-in for AI software adoption in the neuroradiology space, citing his institution’s experience with utilizing RapidAI software in improving care coordination efficiency and reducing door-to-treatment time.
“We use Rapid AI at our institution, and they've really been great about improving our door-to-treatment time because the neurosurgery team is aware of a stroke coming in already. … They're already prepped and ready for the patient before they even get there,” pointed out Dr. Singh.
In addition to enhanced efficiency, Dr. Singh notes the importance of using AI adjunctively within the appropriate clinical context. For example, he cites emerging technology with AI that can provide 3D modeling insights into aneurysm morphology and enhanced visualization beyond what may be apparent to the neuroradiologist’s eye. By providing measurements of shear and stress forces of an aneurysm on a vessel wall, the AI software can help discern subtle shifts in direction that may make the aneurysm more prone to rupture, according to Dr. Singh.
“That's how you actually use AI to save patients right there by intervening early on things that are really critical and using it for what it's most helpful for, which is doing things that we can't do, and calculating things we can't calculate, but using that information in the correct context to take care of people,” emphasized Dr. Singh.
Related Content
