A vascular synthetic model for improved aneurysm segmentation and detection via Deep Neural Networks

TL;DR

This work tackles the scarcity of annotated data for intracranial aneurysm detection in MRA-TOF images by introducing VaMos, a full synthetic 3D vascular model that reproduces cerebral arteries, bifurcations, and aneurysms. The authors build a pipeline combining geometry synthesis, realistic background brain matter and noise, and aneurysm insertion to generate hundreds of synthetic patches around targeted bifurcations. They train a 3D U-Net on real data and augmented real data with 998 synthetic patches, demonstrating substantial gains in lesion-level and patient-level detection when synthetic data are included, while maintaining a reasonable false-positive rate. The results support using synthetic augmentation to enhance DL-based ICA detection and segmentation, with implications for faster dataset generation and potential extension to other imaging modalities; code for VaMos is made available for reproducibility and further development.

Abstract

We hereby present a full synthetic model, able to mimic the various constituents of the cerebral vascular tree: the cerebral arteries, the bifurcations and the intracranial aneurysms. By building this model, our goal was to provide a substantial dataset of brain arteries which could be used by a 3D Convolutional Neural Network (CNN) to either segment or detect/recognize various vascular diseases (such as artery dissection/thrombosis) or even some portions of the cerebral vasculature, such as the bifurcations or aneurysms. In this study, we will particularly focus on Intra-Cranial Aneurysm (ICA) detection and segmentation. The cerebral aneurysms most often occur on a particular structure of the vascular tree named the Circle of Willis. Various studies have been conducted to detect and monitor the ICAs and those based on Deep Learning (DL) achieve the best performances. Specifically, in this work, we propose a full synthetic 3D model able to mimic the brain vasculature as acquired by Magnetic Resonance Angiography (MRA), and more particularly the Time Of Flight (TOF) principle. Among the various MRI modalities, the MRA-TOF allows to have a relatively good rendering of the blood vessels and is non-invasive (no contrast liquid injection). Our model has been designed to simultaneously mimic the arteries geometry, the ICA shape and the background noise. The geometry of the vascular tree is modeled thanks to an interpolation with 3D Spline functions, and the statistical properties of the background MRI noise is collected from MRA acquisitions and reproduced within the model. In this work, we thoroughly describe the synthetic vasculature model, we build up a neural network designed for ICA segmentation and detection, and finally, we carry out an in-depth evaluation of the performance gap gained thanks to the synthetic model data augmentation.

Figures (9)

• Figure 1: Schematic representation of the Circle of Willis on a human vasculature.
• Figure 2: Schematic representation of the whole bifurcation model
• Figure 3: Examples of modified bifurcation centerlines. The solid lines represent the actual coordinates of the bifurcations' branches, the dashed lines stand for the best spline fit functions, and the dotted lines show the effect of modified spline coefficients.
• Figure 4: Comparison between the modeled bifurcations and the Ground Truth crop from a MRA-TOF. We show the comparison on terms of both gray level voxels patches (leftmost panels) and 3D bifurcation layout (rightmost panels).
• Figure 5: Comparison between the modeled bifurcations (bearing an aneurysm). In the upper sub-figures (3D representations), the aneurysm is represented in blue, the mother artery in green.