Robust Centerline Prediction for Accurate Vessel Wall Visualization of Intracranial Vessels in Multi-Contrast 3D MRI Data

TL;DR

This paper introduces a novel algorithm for fast, robust, and accurate centerline prediction in multi-contrast 3D MRI data, significantly improving vessel wall visualization especially for small intracranial vessels.

Contribution

The proposed method enhances centerline accuracy by incorporating vessel surface data, leading to better curved planar reformation visualization in low-resolution MRI datasets.

Findings

Improved accuracy in small vessel visualization.

Faster and more robust centerline computation.

Enhanced vessel wall assessment in clinical MRI data.

Abstract

An accurate planar visualization (curved planar reformation - CPR) of intracranial arteries is mandatory for an improved luminal and mural assessment especially in low resolution Magnetic Resonance Imaging (MRI) data sets acquired in standard clinical settings. CPR visualization methods based on the centerline of the desired structure are fast and easy to implement but the accuracy strongly depends on the spatial resolution of the 3D data set and the size of the desired vessel. In the manuscript, a novel algorithm for fast and robust centerline calculation in multi-contrast 3D MRI data is presented. It considers the extracted surface of the vessel structure for a more accurate centerline prediction resulting in an enhanced CPR visualization of small vessels.