A computational approach for intramural length estimation in anomalous aortic origin of a coronary artery

TL;DR

This paper introduces a new computational method to estimate intramural length in a heart condition called anomalous aortic origin of a coronary artery, which could help improve risk assessment and surgical planning.

Contribution

A semi-automatic computational method for estimating intramural length in AAOCA patients, validated against surgical measurements.

Findings

The computational method showed an overall RMSE of 3.4 mm, comparable to radiologic estimates.

For L-AAOCA subjects, the method had lower RMSE than radiologic estimates.

The method is accurate relative to surgical measurements and could aid in risk stratification and surgical planning.

Abstract

Anomalous aortic origin of a coronary artery (AAOCA) is associated with sudden cardiac death. The intramural (IM) length is considered high-risk, yet radiologic measurements by computed tomography angiography (CTA) show variable agreement with measurements at surgery. We aimed to develop a semi-automatic computational method to estimate IM length in a retrospective cohort of surgical AAOCA patients.

In 58 patients [49 right(R), 9 left(L)], CTA images were used to generate 3D segmentations of the aorta and a centerline of the anomalous coronary. The distance from the centerline to the aortic segmentation was calculated. The IM length was estimated from a transition point in the derivative of the distance curve and compared to radiologic and surgical measurements.

Our method demonstrated an overall root-mean-square error (RMSE) of 3.4 mm, comparable to radiologic estimates (3.2 mm). For L-AAOCA subjects, our method showed lower root-mean-square error compared to radiologic estimates (our method: 3.6 mm, radiologic: 4.7 mm). For R-AAOCA subjects, the RMSE was higher in our method compared to radiologic estimates (our method: 3.4 mm, radiologic: 2.8 mm).

This is a pilot study of a computational approach to measure intramural length that is shown to be accurate relative to surgical measurements. Computational methods that represent and quantify morphology, including acute take-off angle, ostial characteristics, minimal luminal area, and intramural length, may be helpful for risk stratification and surgical planning in AAOCA.

Figures