Learning Wall Segmentation in 3D Vessel Trees using Sparse Annotations

TL;DR

This paper introduces a method for 3D carotid artery wall segmentation using sparse clinical annotations and adversarial 2D networks to generate pseudo-labels, enabling efficient training without manual 3D masks.

Contribution

The novel approach combines sparse annotations with adversarial 2D segmentation to produce 3D labels, improving segmentation efficiency and accuracy in clinical settings.

Findings

Enhanced segmentation performance with bifurcation cross-sections.

Sampling distance had minimal impact on results.

Potential for improved carotid artery disease assessment.

Abstract

We propose a novel approach that uses sparse annotations from clinical studies to train a 3D segmentation of the carotid artery wall. We use a centerline annotation to sample perpendicular cross-sections of the carotid artery and use an adversarial 2D network to segment them. These annotations are then transformed into 3D pseudo-labels for training of a 3D convolutional neural network, circumventing the creation of manual 3D masks. For pseudo-label creation in the bifurcation area we propose the use of cross-sections perpendicular to the bifurcation axis and show that this enhances segmentation performance. Different sampling distances had a lesser impact. The proposed method allows for efficient training of 3D segmentation, offering potential improvements in the assessment of carotid artery stenosis and allowing the extraction of 3D biomarkers such as plaque volume.