COMMA: Coordinate-aware Modulated Mamba Network for 3D Dispersed Vessel Segmentation

TL;DR

This work tackles the challenge of accurately segmenting dispersed 3D vascular structures, where patch-based training often loses essential spatial context. It introduces COMMA, a dual-branch network that concurrently processes full 3D images and cropped patches, augmented by a Coordinate-aware Modulated (CaM) block and a channel-compressed Mamba (ccMamba) design to preserve spatial cues while maintaining efficiency. Key contributions include the CaM block for coordinate-guided interaction, the ccMamba-based global branch, and a large manually labeled 570-case vascular dataset, with comprehensive evaluation across six datasets and ablation studies confirming the value of each component. The approach yields state-of-the-art segmentation performance, particularly for small vessels, and demonstrates favorable computational efficiency, making it practical for clinical and research use and paving the way for semi-/self-supervised extensions using unlabeled data.

Abstract

Accurate segmentation of 3D vascular structures is essential for various medical imaging applications. The dispersed nature of vascular structures leads to inherent spatial uncertainty and necessitates location awareness, yet most current 3D medical segmentation models rely on the patch-wise training strategy that usually loses this spatial context. In this study, we introduce the Coordinate-aware Modulated Mamba Network (COMMA) and contribute a manually labeled dataset of 570 cases, the largest publicly available 3D vessel dataset to date. COMMA leverages both entire and cropped patch data through global and local branches, ensuring robust and efficient spatial location awareness. Specifically, COMMA employs a channel-compressed Mamba (ccMamba) block to encode entire image data, capturing long-range dependencies while optimizing computational costs. Additionally, we propose a coordinate-aware modulated (CaM) block to enhance interactions between the global and local branches, allowing the local branch to better perceive spatial information. We evaluate COMMA on six datasets, covering two imaging modalities and five types of vascular tissues. The results demonstrate COMMA's superior performance compared to state-of-the-art methods with computational efficiency, especially in segmenting small vessels. Ablation studies further highlight the importance of our proposed modules and spatial information. The code and data will be open source at https://github.com/shigen-StoneRoot/COMMA.

Figures (9)

• Figure 1: Illustration of the dispersed nature of vascular structures. (a) Qualitative analysis: visualizations of dispersed vascular structures compared with other compact organs and tissues. (b) Quantitative analysis: dispersion and sparsity indices of various organs. The legend indicates 'dataset-organ' pairs.
• Figure 2: Illustration of the (a) patch-wise-based method compared to (b) our proposed dual-branch method.
• Figure 3: The overall framework of our proposed COMMA.
• Figure 4: The 3D visualization results from (a) KiPA and (b) PARSE datasets.
• Figure 5: The slice visualization results to illustrate the intuitive evaluation.