Early Prediction of Abdominal Aortic Aneurysm Rupture Risk Using Numerical Biomechanical Analysis.

: We aimed to predict patient-specific rupture risks and growth behaviors in abdominal aortic aneurysm (AAA) patients using biomechanical evaluation with finite element analysis to establish an additional AAA repair threshold besides diameter and sex. : A total of 1219 patients treated between 2005 and 2024 (conservative and repaired AAAs) were screened for a pseudo-prospective single-center study. A total of 15 ruptured (rAAA) vs. 15 non-ruptured AAAs (control group) were matched for pre-rupture imaging (first rAAA) and the initial post-rupture imaging (second rAAA) with two images in the asymptomatic control group (first and second control). The matching criteria were as follows: aneurysm diameter, sex, and time period between imagings. The biomechanical properties were analyzed with the finite element method (A4clinicsRE, Vascops GmbH, Graz, Austria). : Both groups had the same median aortic diameter of 5.5 cm in the first imaging but had significantly different aneurysm progressions with 6.9 cm (5.5-9.4 cm) in the second rAAA vs. 6.0 cm (5.1-7.3 cm) in the second control group ( = 0.006). The first rAAA, compared to the first control, showed significantly a higher peak wall stress (PWS) (211.8 kPa vs. 180.5 kPa, = 0.029) and luminal diameter (43.5 mm vs. 35.3 mm; = 0.016). The second rAAA, compared to the matched second control, showed a significantly higher PWS (281.9 kPa vs. 187.4 kPa, = 0.002), luminal diameter (58.3 mm vs. 39.7 mm; = 0.007), PWRR (0.78 vs. 0.49, = 0.014) and RRED (79.8 vs. 56.5, = 0.014). The rAAA group showed over-proportional averages, over the observation time, and an increase in PWS (nearly 10× faster in rAAA) and luminal diameter (nearly 4× faster in rAAA) per month. : The finite element analysis of biomechanical properties could be used for the early prediction of an increased rupture risk in AAA patients. This was confirmed by matched imaging analyses before and after AAA rupture. Further multicenter data are needed to support these findings.