DB-MSMUNet:Dual Branch Multi-scale Mamba UNet for Pancreatic CT Scans Segmentation

TL;DR

This work tackles the challenge of accurate pancreas and tumor segmentation in CT images, where boundaries are blurred and targets are small. It introduces DB-MSMUNet, a dual-branch UNet that combines a Multi-scale Mamba Module backbone with an Edge Enhancement Path and a Multi-layer Decoder, all guided by Auxiliary Deep Supervision. Across NIH Pancreas, MSD, and a clinical tumor dataset, the method achieves Dice Similarity Coefficients around 89% and consistently outperforms CNN, Transformer, and prior Mamba-based methods, demonstrating robustness to deformation and low-contrast scenarios. These results suggest that the proposed architecture effectively improves boundary precision and small-lesion reconstruction, with potential to enhance downstream clinical workflows in pancreatic cancer management.

Abstract

Accurate segmentation of the pancreas and its lesions in CT scans is crucial for the precise diagnosis and treatment of pancreatic cancer. However, it remains a highly challenging task due to several factors such as low tissue contrast with surrounding organs, blurry anatomical boundaries, irregular organ shapes, and the small size of lesions. To tackle these issues, we propose DB-MSMUNet (Dual-Branch Multi-scale Mamba UNet), a novel encoder-decoder architecture designed specifically for robust pancreatic segmentation. The encoder is constructed using a Multi-scale Mamba Module (MSMM), which combines deformable convolutions and multi-scale state space modeling to enhance both global context modeling and local deformation adaptation. The network employs a dual-decoder design: the edge decoder introduces an Edge Enhancement Path (EEP) to explicitly capture boundary cues and refine fuzzy contours, while the area decoder incorporates a Multi-layer Decoder (MLD) to preserve fine-grained details and accurately reconstruct small lesions by leveraging multi-scale deep semantic features. Furthermore, Auxiliary Deep Supervision (ADS) heads are added at multiple scales to both decoders, providing more accurate gradient feedback and further enhancing the discriminative capability of multi-scale features. We conduct extensive experiments on three datasets: the NIH Pancreas dataset, the MSD dataset, and a clinical pancreatic tumor dataset provided by collaborating hospitals. DB-MSMUNet achieves Dice Similarity Coefficients of 89.47%, 87.59%, and 89.02%, respectively, outperforming most existing state-of-the-art methods in terms of segmentation accuracy, edge preservation, and robustness across different datasets. These results demonstrate the effectiveness and generalizability of the proposed method for real-world pancreatic CT segmentation tasks.

Figures (6)

• Figure 1: The overall framework of DB-MSMUNet.
• Figure 2: Structure diagram of MSMM.
• Figure 3: Overall architecture of Edge Enhancement Path.
• Figure 4: Multi-layer decoder structure diagram.
• Figure 5: The performance comparison across the three datasets.