5506
Article
Aura Hagerty

AI for Abdominal Aortic Aneurysm: Creating a Tech-Enabled Practice

In a short period of time, vascular care has undergone a substantial transformation. What may have seemed impossible 60 years ago is now an indispensable component of any high-performing surgical program. The reduction of mortality rates, amongst many other benefits, can be directly attributed to new discoveries and procedures in vascular care. The development of novel technologies has further accelerated the ability of vascular proceduralists and the teams surrounding them to provide the right care for the right patient at the right time.

One such layer of intelligence that supports these new technologies is artificial intelligence (AI). The technology has garnered more attention amongst vascular proceduralists in some of the world’s most renowned hospitals and academic institutions, specifically AI for abdominal aortic aneurysm (AAA) care at Yale New Haven Health (YNHH).

Challenges in AAA Care

AAA care can be uniquely complex. Given the often asymptomatic nature of the disease state, specific imaging is rarely pursued and thus, AAA is an inherently incidental finding. On top of that, failure to detect a patient with AAA may result in aortic rupture, which is associated with a mortality rate of up to 45% for patients who present to hospitals. Further complicating AAA care is:

  1. There are conflicting guidelines when it comes to which populations are ideal for screening.
  2. Even those recommended for screening often don’t receive it (roughly 39% of patients recommended actually get imaging)
  3. There are outliers beyond those within the screening guidelines that will also have AAA

How Yale New Haven Health Applies AI to AAA Workflows

AI has proven to be a powerful technology in hospitals, showcasing its ability to help with incidental findings and relieving shortcomings in screening guidelines by helping to ensure that patients don’t fall through the cracks. AI has enhanced workflows at YNHH in significant ways. By applying AI to torso images, the abdominal aortic measurement algorithm* automatically engages and analyzes CT scans slice-by-slice and is agnostic to the technique of the scan, the scanner itself and protocols. Once min/max diameters are measured, the AI generates a report and the patient’s information is stored in an application.

     
The flowcharts below illustrate the difference between a traditional workflow for an AAA case versus a workflow with AI by the physician’s side. Note that some health systems do not have protocols in place for sub-acute AAA (those 3-5 cm in diameter) but these patients still need to be seen by the vascular surgery team.

Traditional AAA Workflows

“If a patient is in the ED, then the CT scan can be interpreted by the system before the radiologist even has a chance to look at it, and the entire team will be notified about the finding and will be able to see the patient before the read is even available.”

Edouard Aboian, MD, Assistant Professor of Surgery (Vascular) at Yale Medicine

Example of AI-Empowered AAA Workflows at YNHH (>5cm)

The Power of Report-Based AI in AAA Care

Working in conjunction with image-based AI, report-based AI takes the findings from both radiology reports and patient data (such as EMRs) and turns them into actionable insights for vascular teams to monitor patients with abdominal aortic aneurysms.

aortic aneurysm AI patient management follow-up board
A snapshot of the Aidoc patient management board, helping hospitals ensure follow-up. Powered by natural language processing AI to collect vital AAA data from reports including: size, location and type.

Example of AI-Empowered AAA Workflows at YNHH (3-5 cm)

Acting as a layer of actionable intelligence, this workflow illustrates the benefits of AI for abdominal aortic aneurysm for less urgent cases at YNHH, ensuring that patients are not lost to follow-up.

The Future of AI in AAA Care

Because AI can comb through mountains of data that human beings simply cannot, we are optimistic that AI will soon provide physicians with insights such as growth patterns and trends surrounding AAA, thus providing the necessary information to determine AAA healthcare guidelines. Seeing how AI has already provided such a pivotal improvement to AAA patient workflows in YNHH, we’re optimistic about the future of AI in vascular care.

*Aidoc’s Abdominal Aortic Measurement solution is currently pending 510(k) clearance by FDA), and is not available for sale within the United States.

Aura Hagerty