Prompt-mamba filtering networks for accurate hepatocellular carcinoma lesion segmentation in abdominal CT

1

Prompt-Mamba-AF enhances segmentation of hepatocellular carcinoma (HCC) lesions in abdominal CT scans, addressing challenges like morphological heterogeneity.
2

The framework integrates anatomy-aware prompts for feature extraction and employs Mamba-based modeling to capture long-range volumetric dependencies.
3

Structure-aware filtering is introduced to ensure topological consistency along lesion boundaries, improving segmentation accuracy.
4

Validation on LiTS, 3DIRCADb, and CHAOS benchmarks shows Prompt-Mamba-AF outperforms existing CNN and Transformer architectures in segmentation tasks.
5

The model achieves high Dice similarity and boundary accuracy while maintaining a compact parameter footprint, making it suitable for clinical workflows.

Npj Digital Medicine

by Long Xia, Hai-Yang Chen, Ya-Wen Cao, Chen-Quan Gan, Jun-Zhang Zhao, Wei-Hua Zheng, Haiwen Jia, Shuai Jiang, Xuwang Li, Hua Li, Yi-Nuo Tu, Jun-Jing Zhang
January 27, 2026

1