Validation Study Comparing Artificial Intelligence for Fully Automatic Aortic Aneurysms Segmentation and Diameter Measurements on Contrast and Noncontrast Enhanced Computed Tomography: A la mémoire de Lucie Salomon du Mont (In memory of Lucie Salomon du Mont).

BACKGROUND

Accurate aortic diameter measurements are essential for diagnosis, surveillance, and procedural planning in aortic disease. Semiautomatic methods remain widely used but require manual corrections, which can be time-consuming and operator-dependent. Artificial intelligence-driven fully automatic methods may offer improved efficiency and measurement accuracy. This study aims to validate a fully automatic method against a semiautomatic approach using computed tomography angiography (CTA) and noncontrast CT scans.

METHODS

A monocentric retrospective comparative study was conducted on patients who underwent endovascular aneurysm repair (EVAR) for infrarenal, juxta-renal, or thoracic aneurysms and a control group. Maximum aortic wall-to-wall diameters were measured before and after repair using the fully automatic software (PRAEVAorta2, Nurea, Bordeaux, France) and compared to measurements performed by 2 vascular surgeons using a semiautomatic approach on CTA and noncontrast CT scans. Correlation coefficients (Pearson's R) and absolute differences were calculated to assess agreement.

RESULTS

A total of 120 CT scans (60 CTA and 60 noncontrast CT) were included, comprising 23 EVAR, 4 thoracic EVAR, 1 fenestrated EVAR, and 4 control cases. Strong correlations were observed between the fully automatic and semiautomatic measurements in both CTA and noncontrast CT. For CTA, correlation coefficients ranged from 0.94 to 0.96 (R = 0.88-0.92), whereas for noncontrast CT, they ranged from 0.87 to 0.89 (R = 0.76-0.79). Median absolute differences in aortic diameter measurements varied between 1.1 mm and 4.2 mm across the different anatomical locations. The fully automatic method demonstrated a significantly faster processing time, with a median execution time of 73 sec (interquartile range: 57-91) compared to 700 (interquartile range: 613-800) for the semiautomatic method (P < 0.001).

CONCLUSION

The fully automatic method demonstrated strong agreement with semiautomatic measurements for both CTA and noncontrast CT, before and after endovascular repair in different aortic locations, with significantly reduced analysis time. This method could improve workflow efficiency in clinical practice and research applications.