Current state-of-the-art and utilities of machine learning for detection, monitoring, growth prediction, rupture risk assessment, and post-surgical management of abdominal aortic aneurysms

Abstract

Ultrasound imaging has long been playing a central role in detecting abdominal aortic aneurysms (AAAs). With a recent trend of reducing prevalence of AAAs, ultrasound screening is only recommended for men aged 65 to 75 years with previous smoking history, and a national level of a screening program for women is currently not recommended in the US. In the 2000s, several research groups demonstrated the utility of finite element stress analysis using patient-specific images, which was promising for an accurate assessment of the rupture risk, but physical models remain to be enhanced by considering patient variability and multi-physical characteristics. This review aims to provide a survey of emerging and alternative technologies and new methodologies, such as personalized medicine and data-driven approaches, that may make potential breakthroughs on detection of small AAAs, monitoring of patients during the follow-ups, prediction of AAA growth, assessment of the rupture risk, and post-surgical prognosis for AAA patient management.

Keywords

• Data-driven approaches
• EVAR
• Circulating biomarkers
• Pulse wave imaging
• Physics-based machine learning
• Digital twin