A machine learning algorithm developed based on computed tomography angiography imaging (CTA) was found to be useful in predicting expansion of small abdominal aortic aneurysms (AAAs), according to a letter describing 2 studies to the editor of Angiology.

One of the advantages of analysis using machine learning lies in the ability of the technology to detect patterns and identify pertinent variables that might otherwise be missed by human observers. The choice of treatment to manage AAAs is determined based on the risk for growth/rupture. Although several risk factors for AAA growth and/or rupture have been identified (eg, wall stiffness, AAA expansion rate), the evaluation of such a risk needs to be improved.

In one study, a machine learning algorithm was developed based on CTA scans of 50 patients with small aneurysms and ≥2 CTA scans conducted at least 6 months apart. A set of 9 anatomic features were selected to develop this algorithm for the prediction of AAA growth. These features were the major and minor axes of AAA, major and minor axes of lumen without intraluminal thrombus (ILT), ILT area, lumen area without ILT, AAA area, and maximum ILT area and thickness. Significant AAA expansion (ie, >4 mm/year) could be predicted with this algorithm.


Continue Reading

In the other study, the machine algorithm that was developed was found to predict AAA diameter at 12 and 24 months in 85% and 71% of patients, respectively, with a margin of error of ±2 mm.

Related Articles

The generalizability of these findings must first be confirmed and validated through larger multicenter trials.

“Although major challenges remain, we believe that [machine learning] will become a powerful tool that will help surgeons to better assess the risk of AAA rupture and to decide the most appropriate timing for surgical repair,” noted the authors.

Reference

Lareyre F, Adam C, Carrier M, Raffort J. Prediction of abdominal aortic aneurysm growth and risk of rupture in the era of machine learning. Angiology. April 2020. doi:10.1177/0003319720916300