Automatic branch detection of the arterial system from abdominal aortic segmentation

Area of Science:

• Medical Imaging
• Computational Anatomy
• Vascular Surgery

Background:

• Accurate identification of abdominal arteries is crucial for diagnosing and treating aortic diseases.
• Manual segmentation and identification of these arteries can be time-consuming and prone to inter-observer variability.

Purpose of the Study:

• To develop and evaluate an automated method for identifying major arteries branching from the abdominal aorta in medical segmentations.
• To improve the efficiency and accuracy of arterial analysis in abdominal aortic imaging.

Main Methods:

• Representing the arterial system as a vascular tree with topological and geometric features.
• Matching vascular tree branches to specific abdominal arteries (celiac, superior mesenteric, renal, common iliac) using a similarity measure.
• Evaluating the method on 239 segmentations from 102 patients.

Main Results:

• The automated method achieved high accuracy in identifying major abdominal arteries.
• Identification errors were below 2.5% for celiac and superior mesenteric arteries.
• Errors were 8.4% for renal arteries and 2.1% for common iliac arteries.

Conclusions:

• The proposed automated method accurately identifies key arteries originating from the abdominal aorta.
• This technique offers a reliable and efficient tool for quantitative analysis in abdominal aortic imaging.
• Potential applications include improved diagnosis, surgical planning, and outcome prediction for aortic pathologies.