用全局注意力提升MRI图像跨站点一致性，保持肿瘤病理特征。
SA-CycleGAN-2.5D: Self-Attention CycleGAN with Tri-Planar Context for Multi-Site MRI Harmonization
- 引入三平面2.5D结构与自注意力机制，捕捉跨扫描仪的全局强度相关性。
- 在654例胶质瘤患者上将最大均值差异降低99.1%，域分类准确率降至近随机水平。
- 适合需要多中心影像可重复分析的临床研究与放射组学项目。
多中心神经影像分析受扫描仪引起的协变量偏移严重干扰，即体素强度边缘分布 $P($\mathbf{x}$)$ 随采集协议非线性变化，而条件解剖 $P($\mathbf{y}|$\mathbf{x})$ 保持不变。这极大影响放射组学的可重复性，因采集差异常超过生物病理差异。现有统计方法（如ComBat）在特征空间操作，无法支持空间下游任务；标准深度学习方法受限于局部有效感受野（ERF），难以建模场强偏差的全局强度相关性。本文提出SA-CycleGAN-2.5D，基于Ben-David等人的$HΔH$-散度界，集成三项创新：(1) 2.5D三平面流形注入，在$O(HW)$复杂度下保留纵向梯度$ abla_z$；(2) 带密集体素间自注意力的U-ResNet生成器，突破CNN的$O($\sqrt{L}$)$感受野限制，建模全局扫描仪场强偏差；(3) 谱归一化判别器，约束利普希茨常数($K_D \le 1$)，实现稳定对抗优化。在两机构数据集（BraTS和UPenn-GBM）共654例胶质瘤患者上评估，最大均值差异（MMD）从1.729降至0.015（下降99.1%），域分类准确率降至59.7%。消融实验证实全局注意力对更难的异构到同构转换方向至关重要（Cohen's $d = 1.32$, $p < 0.001$）。本框架在2D效率与3D一致性间取得平衡，生成保有肿瘤病理特征的体素级一致图像，支持可重复的多中心放射组学分析。
原文摘要 · Abstract (English)
Multi-site neuroimaging analysis is fundamentally confounded by scanner-induced covariate shifts, where the marginal distribution of voxel intensities $P(\mathbf{x})$ varies non-linearly across acquisition protocols while the conditional anatomy $P(\mathbf{y}|\mathbf{x})$ remains constant. This is particularly detrimental to radiomic reproducibility, where acquisition variance often exceeds biological pathology variance. Existing statistical harmonization methods (e.g., ComBat) operate in feature space, precluding spatial downstream tasks, while standard deep learning approaches are theoretically bounded by local effective receptive fields (ERF), failing to model the global intensity correlations characteristic of field-strength bias. We propose SA-CycleGAN-2.5D, a domain adaptation framework motivated by the $HΔH$-divergence bound of Ben-David et al., integrating three architectural innovations: (1) A 2.5D tri-planar manifold injection preserving through-plane gradients $\nabla_z$ at $O(HW)$ complexity; (2) A U-ResNet generator with dense voxel-to-voxel self-attention, surpassing the $O(\sqrt{L})$ receptive field limit of CNNs to model global scanner field biases; and (3) A spectrally-normalized discriminator constraining the Lipschitz constant ($K_D \le 1$) for stable adversarial optimization. Evaluated on 654 glioma patients across two institutional domains (BraTS and UPenn-GBM), our method reduces Maximum Mean Discrepancy (MMD) by 99.1% ($1.729 \to 0.015$) and degrades domain classifier accuracy to near-chance (59.7%). Ablation confirms that global attention is statistically essential (Cohen's $d = 1.32$, $p < 0.001$) for the harder heterogeneous-to-homogeneous translation direction. By bridging 2D efficiency and 3D consistency, our framework yields voxel-level harmonized images that preserve tumor pathophysiology, enabling reproducible multi-center radiomic analysis.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。