MCMA-Net++: Topology-Aware and Graph-Driven Glioma Segmentation in 3D MRI

Jihan Alameddine

Keywords: Glioma segmentation, Topology-aware learning, Graph neural networks, Hybrid CNN-Transformer, Medical image analysis

Abstract:

\begin{abstract}
Glioma segmentation in 3D MRI remains challenging due to tumor heterogeneity, intensity profiles, and hierarchical anatomical structure. We propose MCMA-Net++, introducing two synergistic innovations: (1) Topology-Aware Refinement Loss (TAR-Loss), enforcing topological consistency across nested tumor subregions (ET, TC, WT), and (2) Multi-Scale Anatomical Graph Reasoning (MSAGR), explicitly modeling spatial dependencies through learnable graphs with anatomical priors. Combined with dual-stream CNN-Swin Transformer encoding and Multi-Class Multi-Attention, MCMA-Net++ achieves Dice scores of 0.970 (WT), 0.943 (TC), 0.926 (ET), while reducing HD95 from 5.48mm to 3.21mm compared to MCMA-Net. Graph reasoning contributes +1.3\% Dice for ET and TAR-Loss reduces topology violations by 41\%.
\end{abstract}

Primary Subject Area: Segmentation

Secondary Subject Area: Geometric Deep Learning

Registration Requirement: Yes

Visa & Travel: Yes

Read CFP & Author Instructions: Yes

Originality Policy: Yes

Single-blind & Not Under Review Elsewhere: Yes

LLM Policy: Yes

Submission Number: 97