Advanced Deep Learning Techniques for Automated Segmentation of Type B Aortic Dissections

TL;DR

This paper presents four deep learning pipelines using 3D U-Net and Swin-UnetR architectures for automated segmentation of Type B aortic dissection features from CTA images, achieving high accuracy and potential clinical utility.

Contribution

It introduces four novel deep learning-based segmentation pipelines specifically designed for Type B aortic dissection, outperforming previous methods in accuracy.

Findings

Achieved Dice Coefficients of 0.91 for TL, 0.88 for FL, and 0.47 for FLT.

Outperformed previous segmentation accuracy reported in literature.

Demonstrated potential for clinical application in treatment planning.

Abstract

Purpose: Aortic dissections are life-threatening cardiovascular conditions requiring accurate segmentation of true lumen (TL), false lumen (FL), and false lumen thrombosis (FLT) from CTA images for effective management. Manual segmentation is time-consuming and variable, necessitating automated solutions. Materials and Methods: We developed four deep learning-based pipelines for Type B aortic dissection segmentation: a single-step model, a sequential model, a sequential multi-task model, and an ensemble model, utilizing 3D U-Net and Swin-UnetR architectures. A dataset of 100 retrospective CTA images was split into training (n=80), validation (n=10), and testing (n=10). Performance was assessed using the Dice Coefficient and Hausdorff Distance. Results: Our approach achieved superior segmentation accuracy, with Dice Coefficients of 0.91 $\pm$ 0.07 for TL, 0.88 $\pm$ 0.18 for FL, and 0.47 $\pm$ 0.25 for FLT, outperforming Yao et al. (1), who reported 0.78 $\pm$ 0.20, 0.68 $\pm$ 0.18, and 0.25 $\pm$ 0.31, respectively. Conclusion: The proposed pipelines provide accurate segmentation of TBAD features, enabling derivation of morphological parameters for surveillance and treatment planning