Towards Full Automation of Geometry Extraction for Biomechanical Analysis of Abdominal Aortic Aneurysm; Neural Network-Based versus Classical Methodologies
Farah Alkhatib,Mostafa Jamshidian,Donatien Le Liepvre,Florian Bernard,Ludovic Minvielle,Antoine Fondanèche,Elke Gizewski,Eva Gassner,Alexander Loizides,Maximilian Lutz,Florian Enzmann,Hozan Mufty,Inge Fourneau,Adam Wittek,Karol Miller
2024-10-30
Abstract:In this study, we investigated the impact of image segmentation methods on the results of stress computation in the wall of abdominal aortic aneurysms (AAAs). We compared wall stress distributions and magnitudes calculated from geometry models obtained from classical semi-automated segmentation versus automated neural network-based segmentation. 16 different AAA contrast-enhanced computed tomography (CT) images were semi-automatically segmented by an analyst, taking between 15 and 40 minutes of human effort per patient, depending on image quality. The same images were automatically segmented using PRAEVAorta commercial software by NUREA (<a class="link-external link-https" href="https://www.nurea-soft.com/" rel="external noopener nofollow">this https URL</a>), developed based on artificial intelligence (AI) algorithms, and automatically post-processed with an in-house MATLAB code, requiring only 1-2 minutes of computer time per patient. Aneurysm wall stress calculations, automatically performed using the BioPARR software (<a class="link-external link-https" href="https://bioparr.mech.uwa.edu.au/" rel="external noopener nofollow">this https URL</a>), revealed that, compared to the classical semi-automated segmentation, the automatic neural network-based segmentation leads to equivalent stress distributions, and slightly higher peak and 99th percentile maximum principal stress values. This difference is due to consistently larger lumen surface areas in automatically segmented models as compared to classical semi-automated segmentations, resulting in greater total pressure load on the wall. However, our statistical analysis indicated that the differences in AAA wall stress obtained using the two segmentation methods are not statistically significant and fall within the typical range of inter-analyst and intra-analyst variability, justifying the use of AI-based automatic segmentation in a fully automated AAA stress computation pipeline.
Computational Engineering, Finance, and Science