Increasing the feasibility of minimally invasive procedures in type A aortic dissections: a framework for...

Area of Science:

• Medical Imaging
• Computational Anatomy
• Cardiovascular Surgery

Background:

• Accurate aortic dimension quantification is vital for treating aortic dissections, particularly Type A.
• Minimally invasive treatments require precise measurements of aortic dimensions and dissection membranes.

Purpose of the Study:

• To develop a precise method for measuring aortic dimensions in dissection pathologies.
• To enable minimally invasive treatment of Type A dissections through accurate quantification.

Main Methods:

• A 2D graph-based approach using phase congruency for aortic outer wall segmentation.
• Extension to 3D using a novel conic directional filter and Log-Gabor filterbank for dissection membrane extraction.
• Development of a complete pipeline for segmentation and quantification of aortic dissections.

Main Results:

• Subvoxel accuracy achieved in 2D/3D membrane extraction algorithms across 10 datasets.
• Outer wall segmentation compared favorably with manual annotations on 11 CTA datasets.
• Aortic inner and outer surface lengths validated against manual measurements in cadaveric CT datasets.

Conclusions:

• A complete pipeline for aortic dissection segmentation and quantification has been proposed.
• Validation of 3D contour length quantification is a significant step towards custom stent-graft design.
• The developed methods facilitate precise measurements for improved patient treatment outcomes.