Automated Aortic Anatomy Analysis: from Image to Clinical Indicators

Area of Science:

• Medical Imaging
• Interventional Neuroradiology
• Computational Anatomy

Background:

• Cerebrovascular diseases are often treated using endovascular catheter navigation from the groin to the brain.
• Complex aortic arch and supra-aortic vessel anatomy pose significant challenges for catheter selection and navigation.
• These anatomical complexities can lead to prolonged procedures, increased complications, and treatment failures.

Purpose of the Study:

• To develop and validate a computational framework for analyzing the anatomy of the aortic arch and supra-aortic trunks.
• To automatically extract key anatomical and geometrical features relevant to endovascular navigation.
• To enhance pre-procedural planning for interventional neuroradiology.

Main Methods:

• A framework was developed to analyze patient-specific meshes of the aortic arch and supra-aortic vessels.
• A Convolutional Neural Network (CNN)-based pipeline was used for prior mesh segmentation.
• The framework automatically computes features like arch type, tortuosity, and angulations.
• Quantitative and qualitative validation was performed by experienced neuroradiologists.

Main Results:

• The framework reliably characterizes the anatomy of the aortic arch and supra-aortic trunks.
• Computed features accurately describe navigational difficulties encountered during catheterization.
• Validation confirmed the reliability and clinical relevance of the vessel characterization.

Conclusions:

• This automated framework provides clinicians with crucial pre-procedural anatomical insights.
• It facilitates optimal catheter selection and navigation planning for endovascular procedures in the aortic arch region.
• The method has the potential to improve outcomes and reduce complications in treating cerebrovascular diseases.