Yale New Haven radiologists leverage CT scan reviews with AI

The growing volume of imaging in the Yale New Haven health system has placed increasing demand on radiologists and staff. Many institutions are facing a similar challenge.

The problem
The health system started looking at artificial intelligence technology in the context of emergency room workflows. Part of this need was figuring the best way to prioritize patients who came in with acute findings discovered during head CT scans.

“We largely entered for operational and research interests to start,” said Dr Melissa A. Davis, associate professor of radiology and biomedical imaging and vice chair for imaging informatics, radiology and biomedical imaging at Yale New Haven. “These technologies are largely still new, so their impacts are not completely known. We wanted to be at the leading edge of that conversation.”

Proposal
The initial proposal was that the AI would evaluate CTs of the head without contrast across the main hospital and the health system would validate the findings with dedicated neuroradiology-trained radiologists.

“We would also evaluate if there were any gains in turnaround time for these images and report discrepancy rates for a retrospective cohort,” Davis noted. “The plan was to deploy solutions that would flag and prioritize intracranial hemorrhages on non-contrast CT examinations.

“The biggest expectation was around the user experience,” she added. “It couldn’t be another step in the workflow. The expectation was that everything would be seamless and integrated into the radiologist’s current workflow. Without that approach, adoption of technology would be challenging and burdensome.”

Meeting the challenge
The AI program is led by the department of radiology and biomedical imaging and aligned with IT to ensure a smooth integration and transition, Davis explained.

“At the outset, the AI was deployed for the emergency radiologists and neuroradiologists,” she said. “It was used to evaluate head CT scans for potential. When a head CT scan was performed, the AI tool would analyze the images and flag any cases with findings suspected of blood.

“Integration was a key aspect of this implementation,” she continued. “The solution had to seamlessly integrate with existing radiology workflow systems. Embedding an icon in the worklist to indicate when AI had flagged a case enabled easy adoption and use of the technology.”

“The adoption of AI technology in healthcare can lead to significant improvements in efficiency, patient care and outcomes.” – Dr Melissa A. Davis, Yale New Haven.
Additionally, there was a notification that would pop up when an acute finding was detected, facilitating immediate awareness among radiologists and other clinicians.

“There was a significant reduction in turnaround time for our level 1 trauma center ER, but not in other locations,” Davis reported. “It did flag an outpatient with a head bleed that otherwise would have been sent home. This was the biggest initial success. We also noticed radiologists began to see a bit of comfort in having a second set of ‘eyes’ on these cases.

“As the implementation evolved, the use of AI expanded beyond head CT scans to other radiology applications, including the detection of pulmonary embolisms and coronary artery calcification,” she continued. “This also expanded the conversation of AI’s place within radiology from one focused on taking over to one focused on augmenting our current workflows.”

While staff often talk about gaining time, Davis said sometimes the AI actually slows her down a bit, forcing her to look at an image or area much more closely – especially if she doesn’t agree with the AI.

“You always have to maintain critical thinking,” she stated.

Results
“For Yale, there have been three potential areas of downstream value we have looked at: increased sensitivity and specificity leading to improved accuracy; discovery of incidental findings leading to more clinically appropriate interventions; and improved efficiency aiding in patient length of stay reduction,” Davis said. “Studies have highlighted those impacts.”

In one study on improved accuracy that Davis pointed to, an AI algorithm was applied to a retrospective cohort of 1,387 consecutive CT pulmonary angiograms. The prevalence of pulmonary embolism (PE) was 13.6% (189 cases). The algorithm was 93% sensitive and 96% specific in the detection of PE. The positive predictive value was 77% and the negative predictive value was 99%.

Davis said the conclusion to be drawn from the study is that the high negative predictive value shows success at screening.

Advice for others
For healthcare provider organizations considering the adoption of similar AI technology in their clinical workflows, there are several crucial pieces of advice to consider, Davis offered.

“First, it’s essential to thoroughly assess your organization’s specific clinical needs and challenges,” she said. “AI technology can be a powerful tool, but its effectiveness depends on how well it aligns with your objectives.

“Identify the areas where AI could potentially make a meaningful impact, such as improving workflow efficiency, enhancing diagnostic accuracy or expediting critical results communication,” she continued. “Prioritize the adoption of AI in these areas, as it can be challenging to implement across the board simultaneously.”

Second, engage key stakeholders, including healthcare professionals, IT teams and administrators, in the decision-making process, she advised.

“Ensure there is buy-in and support from all relevant parties,” she insisted. “Healthcare providers who will directly use the technology should be involved in the selection process, as their feedback and insights are invaluable for successful implementation.

“Third, carefully evaluate AI technology vendors,” she continued. “Consider factors like the accuracy of the AI algorithms, ease of integration with existing systems, vendor reputation, and the level of ongoing support and updates provided. Seek out references and case studies from organizations that have successfully implemented similar AI solutions to gain insights into their experiences.”

Further, invest in robust IT infrastructure and ensure the organization’s systems can support the integration of AI seamlessly, she said.

“Collaboration between your IT team and the AI vendor is crucial to address any technical challenges that may arise during implementation,” she added.

Finally, focus on education and training, she said.

“Ensure healthcare professionals, particularly radiologists and clinicians, are adequately trained to use the AI technology effectively,” she advised. “Emphasize the importance of maintaining a critical mindset when working with AI and encourage continuous learning to stay updated with the latest advancements.

“The adoption of AI technology in healthcare can lead to significant improvements in efficiency, patient care and outcomes,” she concluded. “However, successful implementation requires careful planning, stakeholder engagement, vendor evaluation, IT infrastructure readiness and ongoing education. By taking these steps, healthcare provider organizations can harness the potential of AI to enhance their clinical workflows and ultimately provide better care to their patients.” Healthcare IT News