AdaSurvMamba：用于多模态生存分析的动态融合与语义扫描
AdaSurvMamba: Dynamic Fusion and Semantic Scanning for Multimodal Survival Analysis
浏览论文内容
中文总结 AI 辅助
研究针对多模态生存分析中传统方法的局限，提出AdaSurvMamba框架，含DSIR模块动态调制跨模态交互强度、SAS模块重组令牌成连续序列，实验证明该框架在五个TCGA队列上比现有方法有优势。
中文摘要 AI 辅助
利用全切片图像（WSIs）和基因组图谱进行多模态生存分析对癌症预后至关重要。近期，像Mamba这样的状态空间模型成为序列建模的强大工具。然而，将其成功应用于复杂多模态任务存在两个关键局限。一是传统融合策略假设多模态交互强度静态，忽略各模态在不同患者和局部区域的波动诊断重要性；二是标准Mamba架构沿预定义物理路径处理令牌，破坏空间分散医学特征的语义连续性并加剧长程衰减。为应对这些挑战，我们引入AdaSurvMamba作为多模态生存分析的新型自适应框架。该框架有双尺度重要性感知重建（DSIR）模块动态调制跨模态交互强度，在序列和令牌级别评估诊断重要性以重建输入输入表示表示。还提出语义聚合扫描（SAS）模块克服上下文碎片化，通过共享原型池将离散令牌动态重组为语义连续序列，利用全局模态上下文和语义先验明确调制状态转换步长以自适应控制信息吸收率。在五个TCGA队列上的实验表明比现有方法有持续优势。代码可在指定网址获取。
英文摘要
Multimodal survival analysis utilizing whole slide images (WSIs) and genomic profiles is fundamental for cancer prognosis. Recently, state-space models like Mamba have emerged as powerful tools for sequence modeling. However, translating this success to complex multimodal tasks is hindered by two critical limitations. First, conventional fusion strategies assume a static multimodal interaction strength, ignoring the fluctuating diagnostic importance of each modality across different patients and local regions. Second, the standard Mamba architecture processes tokens along predefined physical paths. This rigid scanning disrupts the semantic continuity of spatially scattered medical features and exacerbates long-range decay. To address these challenges, we introduce AdaSurvMamba as a novel adaptive framework for multimodal survival analysis. The framework features a Dual-Scale Importance-Aware Reconstruction (DSIR) module to dynamically modulate cross-modal interaction strength. It evaluates diagnostic importance at both the sequence and token levels to reconstruct the input representations. Furthermore, we propose a Semantic Aggregation Scanning (SAS) module to overcome contextual fragmentation. The SAS module dynamically reorganizes discrete tokens into semantically continuous sequences via a shared prototype pool. It explicitly modulates the state transition step size using global modality context and semantic priors to adaptively control the information absorption rate. Experiments across five TCGA cohorts demonstrate consistent gains over existing methods. Code is available at https://github.com/zjlGO/AdaSurvMamba.
发表机构
- Dalian University of Technology（大连理工大学）
- University of Alberta（阿尔伯塔大学）
- Yale University（耶鲁大学）
机构由 AI 辅助整理，请以论文原文为准。
