https://doi.org/10.1140/epjp/s13360-025-07224-w
Regular Article
Hybrid POD–Transformer model for physical analysis of real-time blood flow in cerebral aneurysms based on computational data
1
Advanced Technical College, University of Warith Al-Anbiyaa, Karbala, Iraq
2
Department of Chemical Engineering, College of Engineering, Imam Mohammad Ibn Saud Islamic University (IMSIU), 11432, Riyadh, Saudi Arabia
3
Department of Civil Engineering, College of Engineering, Cihan University-Erbil, Erbil, Iraq
4
Faculty of Data Science and Information Technology, INTI International University, Persiaran Perdana BBN, 71800, Putra Nilai, Nilai, Malaysia
5
College of Engineering, Department of Mechanical Engineering, Najran University, P.O Box 1988, King Abdulaziz Road, Najran, Saudi Arabia
6
Deanship of Scientific Research, Imam Mohammad Ibn Saud Islamic University (IMSIU), Riyadh, Saudi Arabia
7
Independent Researcher, Mechanical Engineering, Doha, Qatar
8
Department of Pathological Analyzes, Al Manara College for Medical Sciences, Maysan, Iraq
a
This email address is being protected from spambots. You need JavaScript enabled to view it.
Received:
24
November
2025
Accepted:
15
December
2025
Published online:
5
January
2026
Accurate assessment of rupture risk in cerebral aneurysms, particularly in the middle cerebral artery (MCA), requires detailed hemodynamic analysis of blood flow parameters such as wall shear stress (WSS), oscillatory shear index (OSI), and pressure. However, full-order computational fluid dynamics (CFD) simulations are time-consuming and computationally expensive, limiting their routine clinical use. In this study, we propose a reduced-order modeling framework that combines proper orthogonal decomposition (POD) with a transformer neural network to enable fast and accurate prediction of pulsatile blood flow and key hemodynamic indices within patient-specific aneurysm geometries. Blood flow was modeled using the incompressible Navier–Stokes equations with the non-Newtonian Casson model under laminar conditions, and simulations were performed over three cardiac cycles. The dominant flow features were extracted using POD, and the temporal evolution of modal coefficients was learned using a transformer architecture trained on time-windowed data. Results show that the proposed POD + transformer approach accurately predicts velocity and pressure fields with reconstruction errors below 2% and captures WSS distributions with acceptable fidelity. OSI, due to its inherently complex and oscillatory nature, showed higher prediction errors, highlighting the need for more refined modeling strategies. Overall, the framework provides a promising step toward real-time, data-driven aneurysm hemodynamic analysis, offering significant potential for clinical risk stratification and decision support in neurovascular care.
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to Società Italiana di Fisica and Springer-Verlag GmbH Germany, part of Springer Nature 2026
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
