Automated Aortic Quantification Based on Artificial Intelligence: Validation Using Contrast-Enhanced and Non-Contrast...

Area of Science:

• Radiology and Medical Imaging
• Artificial Intelligence in Healthcare
• Cardiovascular Imaging

Background:

• Early detection of aortic dilatation is crucial for preventing severe aortic disease.
• Timely intervention based on accurate aortic measurements can improve patient outcomes.
• Current methods for aortic quantification may be time-consuming or lack standardization.

Purpose of the Study:

• To develop and validate an artificial intelligence (AI) method for automated aortic quantification.
• To assess the AI's performance on both contrast-enhanced and non-contrast CT scans.
• To support early detection of aortic dilatation and facilitate timely clinical intervention.

Main Methods:

• Utilized open-source tools to develop an AI algorithm for aortic measurement.
• Analyzed 380 paired contrast-enhanced and non-contrast CT scans from 190 patients.
• Validated AI-derived measurements against manual annotations and clinical records.

Main Results:

• The AI demonstrated strong agreement with manual annotations (correlation coefficients up to 0.987).
• High sensitivity for AI detection of aortic dilatation was observed: 97% for contrast-enhanced CT and 94% for non-contrast CT.
• The AI workflow showed excellent performance across different imaging modalities.

Conclusions:

• The developed AI-based workflow enables highly sensitive and automated aortic quantification.
• This method is effective for both contrast-enhanced and non-contrast CT scans, increasing clinical applicability.
• The AI tool supports early detection of aortic dilatation, potentially improving patient management.