Automating aortic cross-sectional measurement of 3D aorta models

Area of Science:

• Cardiovascular Imaging
• Medical Engineering
• Radiology

Background:

• Aortic dissection has high mortality rates, and surgical interventions carry significant risks.
• Accurate longitudinal aortic diameter measurement is crucial for clinical decision-making.
• Current methods rely on 2D measurements, which may be less efficient or accurate than 3D approaches.

Purpose of the Study:

• To develop and validate an algorithm for automated measurement of longitudinal aortic diameters from 3D models.
• To assess the diagnostic validity of 3D modeling measurements as a substitute for traditional 2D measurements.
• To establish a foundation for transitioning aortic measurement methodologies from 2D to 3D.

Main Methods:

• Sixty-six 3D non-contrast steady-state free precession magnetic resonance images of aortic pathology were retrospectively identified.
• 3D aorta models were manually segmented.
• A novel mathematical algorithm was applied to analyze segmented 3D models and generate maximal aortic diameters, which were correlated with clinical measurements.

Main Results:

• The automated measurement tool was successfully applied to 50 (76% success rate) of the 3D aortic models.
• Excellent correlation was observed between automated 3D measurements and clinical measurements across multiple aortic locations (ICC > 0.89 for most sites, p < 0.001).
• High intra-class correlation coefficients (ICC) and statistical significance (p < 0.001) were achieved for measurements at the sinus of Valsalva, ascending aorta, brachiocephalic artery, and descending aorta.

Conclusions:

• Automated measurement of aortic diameters from 3D models demonstrates excellent correlation with clinical measurements.
• This validated tool is a critical first step toward a fully automated process for aortic analysis.
• The findings support the transition of aortic measurement methodologies from 2D to 3D imaging for improved diagnostic confidence and efficiency.