MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
2510.17529v1
cs.CV, cs.AI, cs.LG
2025-10-22
Авторы:
Yovin Yahathugoda, Davide Prezzi, Piyalitt Ittichaiwong, Vicky Goh, Sebastien Ourselin, Michela Antonelli
Abstract
Active Surveillance (AS) is a treatment option for managing low and
intermediate-risk prostate cancer (PCa), aiming to avoid overtreatment while
monitoring disease progression through serial MRI and clinical follow-up.
Accurate prostate segmentation is an important preliminary step for automating
this process, enabling automated detection and diagnosis of PCa. However,
existing deep-learning segmentation models are often trained on
single-time-point and expertly annotated datasets, making them unsuitable for
longitudinal AS analysis, where multiple time points and a scarcity of expert
labels hinder their effective fine-tuning. To address these challenges, we
propose MambaX-Net, a novel semi-supervised, dual-scan 3D segmentation
architecture that computes the segmentation for time point t by leveraging the
MRI and the corresponding segmentation mask from the previous time point. We
introduce two new components: (i) a Mamba-enhanced Cross-Attention Module,
which integrates the Mamba block into cross attention to efficiently capture
temporal evolution and long-range spatial dependencies, and (ii) a Shape
Extractor Module that encodes the previous segmentation mask into a latent
anatomical representation for refined zone delination. Moreover, we introduce a
semi-supervised self-training strategy that leverages pseudo-labels generated
from a pre-trained nnU-Net, enabling effective learning without expert
annotations. MambaX-Net was evaluated on a longitudinal AS dataset, and results
showed that it significantly outperforms state-of-the-art U-Net and
Transformer-based models, achieving superior prostate zone segmentation even
when trained on limited and noisy data.
Ссылки и действия
Дополнительные ресурсы: