An autonomous design assistant for aneurysm repair

Cardiovascular disease is becoming increasingly common in Western societies, and abdominal aortic aneurysm (AAA) is one of the most deadly. When an AAA ruptures, the mortality rate is approximately 90%. Historically, open surgery was the standard of care, but in the early 1990s, a less invasive method called endovascular aneurysm repair (EVAR) was introduced. EVAR uses a stent graft to prevent the aneurysm from rupturing. While this is a much less invasive form of treatment, there are still challenges, such as the stent not sealing properly (causing leakage) or shifting from its intended position. In order to improve the safety and long-term efficacy of EVAR, each EVAR must be tailored to the individual, patient-specific conditions to ensure the best possible treatment. Currently, not only is the selection of the correct stent graft design largely based on the surgeon`s experience, but the surgeon also has a limited number of off- the-shelf solutions to choose from. This is actually an outdated approach, as advances in additive manufacturing, such as electrospinning and selective laser melting, now make it possible to create next-generation, customized stent grafts for improved patient outcomes. But how do we find the best possible design for each patient? In this project, we are developing an artificial intelligence (AI)-based design assistant to create customized stent grafts for each patient. The AI assistant will be able to create new designs from scratch. It will be able to evaluate the performance of these designs based on physical models and will be trained to adapt the designs to individual patients.