LASSNet: Level-Aware Availability-Conditioned Spatial-Semantic Fusion for Brain Tumor Segmentation with Missing MRI Modalities

Haobin Chen; Ao Chang; Rundong Wang; Zhicheng Li; Zhihao Tang; Heqin Zhu

cs.CVSep 6, 2026

LASSNet: Level-Aware Availability-Conditioned Spatial-Semantic Fusion for Brain Tumor Segmentation with Missing MRI Modalities

Authors: Haobin Chen, Ao Chang, Rundong Wang, Zhicheng Li, Zhihao Tang, Heqin Zhu

Organizations: School of Biomedical Engineering, Division of Life Sciences and Medicine, University of Science and Technology of China (USTC), Hefei, Anhui, 230026, China · Medical Imaging, Robotics, Analytic Computing & Learning (MIRACLE) Lab, YRD-RIGHT, Suzhou Institute for Advanced Research, USTC, Suzhou, Jiangsu, 215123, China · College of Information Science and Technology, Donghua University

Abstract

Brain tumor segmentation from multimodal MRI relies on complementary evidence across four imaging sequences, yet one or more modalities may be unavailable because of acquisition cost, protocol variation, scan failure, or patient condition. Existing work has explored reconstruction, knowledge transfer, and direct feature fusion, but leaves open whether missing-modality fusion should change with representation level. High-resolution lateral features retain spatial detail, whereas compressed bottleneck features encode semantic and inter-modality context. We therefore hypothesize that fusion should be conditioned jointly on modality availability and feature hierarchy. We propose the Level-Aware Availability-Conditioned Spatial-Semantic Fusion Network (LASSNet), which contains two level-specialized modules. Hierarchical Availability-Conditioned Fusion (HACF) constructs four lateral representations using count-normalized aggregation of available modalities, mask-conditioned channel modulation, and local 3D refinement. Tri-Scale Relational-Spatial Fusion (TriRSF) models relations among available modality descriptors and spatial context across multiple bottleneck resolutions, followed by cross-scale aggregation and availability-conditioned global spatial attention. A shared coarse-to-fine decoder starts from TriRSF semantics and progressively injects HACF features, without reconstructing missing inputs. Across all 15 non-empty modality configurations, LASSNet obtains mean Dice scores of 76.7% and 83.2% over WT, TC, and ET on BraTS2019 and BraTS2023, respectively.

Figures & tables

Figure 1: Overview of the proposed framework. HACF constructs availability-conditioned lateral representations, while TriRSF models modality relations and spatial context at the bottleneck. They are parallel paths coupled by coarse-to-fine decoding; TriRSF does not take HACF as its input. The final global Transformer operates over spatial tokens after modality aggregation.

Panel of Table 1 — Table 1: DSC% ( $\uparrow$ ) comparison across different missing modalities on BraTS2019. Filled and dashed circles denote present and missing modalities, respectively. Our proposal is identified with $\dagger$ . Best and second best results in bold and underline , respectively. ∗ denotes a statistically significant improvement over the corresponding baseline across the 15 missing-modality configurations using a paired Wilcoxon signed-rank test with Holm correction ( $p<0.05$ ).

Panel of Table 2 — Table 2: DSC% ( $\uparrow$ ) comparison across different missing modalities on BraTS2023. Filled and dashed circles denote present and missing modalities, respectively. Our proposal is identified with $\dagger$ . Best and second best results in bold and underline , respectively. ∗ denotes a statistically significant improvement over the corresponding baseline across the 15 missing-modality configurations using a paired Wilcoxon signed-rank test with Holm correction ( $p<0.05$ ).

Figure 2: Visualization of LASSNet segmentation results under missing modality scenarios. Enhancing tumor (yellow), necrotic tissue (red), and edema (green).

Variant	Lateral fusion	Bottleneck fusion	WT $\uparrow$	TC $\uparrow$	ET $\uparrow$	Mean DSC $\uparrow$
Base	Masked mean	Masked mean	86.21	78.29	57.01	73.84
Uniform ACF	Uniform ACF	Uniform ACF	85.02	77.54	56.63	73.06
w/o HACF	Masked mean	TriRSF	85.89	78.12	57.27	73.76
w/o TriRSF	HACF	Masked mean	85.77	77.82	56.70	73.43
Full	HACF	TriRSF	87.89	80.38	61.88	76.72

Table 3: Ablation study of the proposed hierarchical fusion framework under all missing-modality scenarios on BraTS2019. Different variants evaluate the effectiveness of HACF and TriRSF at lateral and bottleneck feature levels. Results are reported using the DSC (%) metric for WT, TC, ET, and Mean DSC.

Explore similar work

A Multimodal Feature Distillation with Mamba-Transformer Network for Brain Tumor Segmentation with Incomplete Modalities
Apr 22, 2024Ming Kang, Fung Fung Ting, Shier Nee Saw +3Cross-Modal Knowledge Distillation Cross-Modality Medical Image Segmentation
School of Information Technology, Monash University, Malaysia Campus, Subang Jaya 47500, Malaysia · Augmented Intelligence and Multimodal analytics (AIM) for Health Lab, Faculty of Information Technology, Monash University, Clayton VIC 3800, Australia · Monash Medical AI Group, Monash University, Clayton VIC 3800, Australia +1
Uni-Encoder Meets Multi-Encoders: Representation Before Fusion for Brain Tumor Segmentation with Missing Modalities
Apr 24, 2026Peibo Song, Xiaotian Xue, Jinshuo Zhang +5Cross-Modality Medical Image Segmentation Hybrid CNN-Transformer Architectures
School of Mathematics, Shandong University · 2GSFS, The University of Tokyo · 3MedVisAI Lab +2
D3Seg: Dependency-Aware Diffusion for Brain Tumor Segmentation with Missing Modalities
May 21, 2026Danish Ali, Ajmal Mian, Naveed Akhtar +1Missing-Modality Learning BraTS
The University of Western Australia, Perth, WA 6009, Australia · The University of Melbourne, Melbourne, Parkville VIC 3010, Australia

Apr 22, 2024J/K move · Enter open · S save

A Multimodal Feature Distillation with Mamba-Transformer Network for Brain Tumor Segmentation with Incomplete Modalities

Ming Kang, Fung Fung Ting, Shier Nee Saw, Raphaël C. -W. Phan+2

School of Information Technology, Monash University, Malaysia Campus, Subang Jaya 47500, Malaysia · Augmented Intelligence and Multimodal analytics (AIM) for Health Lab, Faculty of Information Technology, Monash University, Clayton VIC 3800, Australia · Monash Medical AI Group, Monash University, Clayton VIC 3800, Australia+1

Cross-Modal Knowledge Distillation Cross-Modality Medical Image Segmentation Medical Image Segmentation Multimodal Fusion Brain Tumor Segmentation

Open paper AskPDF arXiv

Existing brain tumor segmentation methods usually utilize multiple Magnetic Resonance Imaging (MRI) modalities in brain tumor images for segmentation, which can achieve better segmentation performance. However, in clinical applications, some modalities are often missing due to resource constraints, resulting in significant performance degradation for methods that rely on complete modality segmentation. In this paper, we propose a Multimodal feature distillation with Mamba-Transformer hybrid network (MMTSeg) for accurate brain tumor segmentation with missing modalities. We first employ a Multimodal Feature Distillation (MFD) module to distill feature-level multimodal knowledge into different unimodalities to extract complete modality information. We further develop an Unimodal Feature Enhancement (UFE) module to model the semantic relationship between global and local information. Finally, we built a Cross-Modal Fusion (CMF) module to explicitly align the global correlations across modalities, even when some modalities are missing. Complementary features within and across modalities are refined by the Mamba-Transformer hybrid architectures in both the UFE and CMF modules, dynamically capturing long-range dependencies and global semantic information for complex spatial contexts. A boundary-wise loss function is employed as the segmentation loss of the proposed MMTSeg to minimize boundary discrepancies for a distance-based metric. Our ablation study demonstrates the importance of the proposed feature enhancement and fusion modules in the proposed network and the Transformer with Mamba block for improving the performance of brain tumor segmentation with missing modalities. Extensive experiments on the BraTS 2018 and BraTS 2020 datasets demonstrate that the proposed MMTSeg framework outperforms state-of-the-art methods when modalities are missing.

Abstract

Explore similar work

A Multimodal Feature Distillation with Mamba-Transformer Network for Brain Tumor Segmentation with Incomplete Modalities

Uni-Encoder Meets Multi-Encoders: Representation Before Fusion for Brain Tumor Segmentation with Missing Modalities

D3Seg: Dependency-Aware Diffusion for Brain Tumor Segmentation with Missing Modalities