From Anatomy to Aneurysm: Morphological-Hemodynamic Coupling and Predictive Modeling in Aberrant Splenic Artery.

Background: Aberrant splenic artery (SA), defined as the SA arising from the superior mesenteric artery (SMA) rather than the celiac axis, is a rare anatomical variant with a prevalence of less than 1% in the general population. Although clinical observations suggest that individuals with this variant are highly susceptible to developing SA aneurysms (SAAs), the underlying pathophysiological mechanisms remain poorly understood. Objectives: This study aimed to systematically investigate the morphological and hemodynamic features of aberrant SA anatomy and their contributions to SAA formation, and to develop a predictive model for individualized risk stratification. Methods: This multi-center retrospective cohort study enrolled 195 patients with aberrant SA anatomy from 16 centers across China between January 2008 and June 2026, including 94 patients with concomitant SAAs and 101 without. In addition, 100 patients with common SAAs and 100 non-aneurysmal controls with normal vascular anatomy were enrolled for comparative analysis. Three-dimensional vascular models were reconstructed from computed tomography angiography? (CTA) images, and twelve morphological parameters were measured. Computational fluid dynamics simulations were performed to analyze fourteen hemodynamic parameters. A two-stage machine learning model based on five morphological parameters was developed to predict SAA occurrence and location. Results: Aberrant SAAs were significantly larger and predominantly located in the proximal SA, whereas common SAAs were mainly located in the distal splenic hilum. Morphological analysis revealed that the aberrant SA exhibited a smaller branching angle with its parent vessel, an increased SA/SMA cross-sectional area ratio, and a higher tortuosity index. Hemodynamic analysis demonstrated significantly lower time-averaged wall shear stress (TAWSS) at the proximal SA in aberrant anatomy, whereas the SA-parent vessel junction showed elevated oscillatory shear index (OSI), which was inversely correlated with the branching angle. Longitudinal follow-up of non-aneurysmal patients (n=14; mean follow-up, 36.8 months) showed stable vessel dimensions over time. The KNN-based predictive model achieved an AUROC of 0.828 for predicting SAA occurrence and was deployed as an interactive web application (the zs_abeSA model) for clinical use. Conclusions: Aberrant SA anatomy is an independent risk factor for SAA formation. The pathogenic mechanism involves a cascade from morphological remodeling to hemodynamic derangement, ultimately leading to aneurysm formation at the proximal SA. The zs_abeSA model provides a practical tool for individualized risk assessment and clinical decision-making, with direct implications for screening, surveillance, and therapeutic strategies.