Two-Stage Generative Model for Intracranial Aneurysm Meshes with Morphological Marker Conditioning
Wenhao Ding, Choon Hwai Yap, Kangjun Ji, Simão Castro
TL;DR
This work tackles the scarcity of large IA datasets by proposing AneuG, a two-stage conditional VAE that jointly models IA aneurysm complexes and their parent vessels. The aneurysm shape is encoded with Graph Harmonic Deformation ($GHD$) tokens, while vessel centerlines are represented by Fourier-encoded tokens, with conditioning enabled through a differentiable Morphological Marker Calculator ($MMC$) and a Morphing Energy Alignment ($MEA$) constraint to match population statistics. The approach achieves higher fidelity than PCA-based and diffusion baselines on an IA dataset (AneuX) and enables controlled generation of shapes with clinically relevant parameters, improving the utility for flow simulations and morphology–flow studies. This data-efficient, physiology-aware generator advances IA biomechanics research and offers a foundation for future temporal conditioning and image-synthesis integrations.
Abstract
A generative model for the mesh geometry of intracranial aneurysms (IA) is crucial for training networks to predict blood flow forces in real time, which is a key factor affecting disease progression. This need is necessitated by the absence of a large IA image datasets. Existing shape generation methods struggle to capture realistic IA features and ignore the relationship between IA pouches and parent vessels, limiting physiological realism and their generation cannot be controlled to have specific morphological measurements. We propose AneuG, a two-stage Variational Autoencoder (VAE)-based IA mesh generator. In the first stage, AneuG generates low-dimensional Graph Harmonic Deformation (GHD) tokens to encode and reconstruct aneurysm pouch shapes, constrained to morphing energy statistics truths. GHD enables more accurate shape encoding than alternatives. In the second stage, AneuG generates parent vessels conditioned on GHD tokens, by generating vascular centreline and propagating the cross-section. AneuG's IA shape generation can further be conditioned to have specific clinically relevant morphological measurements. This is useful for studies to understand shape variations represented by clinical measurements, and for flow simulation studies to understand effects of specific clinical shape parameters on fluid dynamics. Source code and implementation details are available at https://github.com/anonymousaneug/AneuG.