A region-specific spatiotemporal framework for predicting abdominal aortic aneurysm morphological evolution

Rupture of abdominal aortic aneurysm (AAA) is highly lethal, and its unpredictability and suddenness pose a major clinical challenge. Although aneurysm morphology serves as an important indicator of rupture risk, the growth and morphological evolution of the aneurysm over time remain difficult to accurately predict. Based on Computed Tomography Angiography (CTA) sequences from 6 patients (29 CTA scans) with multiple follow-ups (over approximately 13.2 years), this study proposes RSS-AAA (Region-Specific Spatiotemporal), a longitudinal framework capable of tracking and predicting the dynamic progression of aneurysm morphology. We innovatively integrate velocity/acceleration priors with an aneurysm specific bending energy term, enabling the reconstruction of past aneurysm morphology and the prediction of its future progression. Compared to the real AAA geometry, the morphology reconstructed by our method achieved that the average geometric error is 4.6 mm, the range of geometric error is 3–5 mm, the average bending energy error is 7.09%, and the range of the bending energy error is 3.26–14.62%. As examined by univariate sensitivity analysis, this framework demonstrates high stability. Furthermore, computational fluid dynamics simulations confirmed the hemodynamic accuracy of the RSS-AAA reconstruction model, demonstrating highly consistent flow fields and wall shear stress distributions. This method addresses the limitations of the current practice that relies solely on the maximum diameter for risk assessment, providing clinicians with an effective tool to predict future AAA progression and conduct early risk assessment for patients.