发表机构
University College London; University of Cambridge; Cambridge Institute of Therapeutic Immunology and Infectious Diseases; Wellcome Trust Sanger Institute; Cambridge Stem Cell Institute; Centre for AI in Medicine; Cavendish Laboratory; Medical University of Vienna; European Synchrotron Radiation Facility; CIFAR; Cambridge University Hospitals NHS Foundation Trust; NIHR Cambridge Biomedical Research Centre（伦敦大学学院; 剑桥大学; 剑桥治疗免疫学与传染病研究所; 惠康基金会桑格研究所; 剑桥干细胞研究所; 医学人工智能中心; 卡文迪什实验室; 维也纳医科大学; 欧洲同步辐射装置; 加拿大高级研究所; 剑桥大学医院NHS基金会信托基金; 英国国家健康与研究院剑桥生物医学研究中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出基于DINOv3的LoRCA框架，用于实现无需配对数据的组织学到HiP-CT的结构保留风格转换，其性能优于CycleGAN，为2D组织学切片与3D HiP-CT体积的配准提供了有效途径。
AI 中文摘要
分层相位对比断层成像（HiP-CT）是一种基于同步加速器的X射线成像技术，可对完整器官进行非破坏性体积成像，分辨率跨度达20μm/体素（全器官）至局部区域的近细胞分辨率（约0.8μm/体素），这为将全器官体积信息引入组织学提供了可能。然而，由于不同色彩空间的特征表示存在差异，H&E组织学与HiP-CT体积之间的非线性配准极具挑战性。配准前合成方法在组织学到MRI、组织学到CT的对齐中已取得良好效果，但现有方法要么依赖手动解剖轮廓，要么从零开始训练且无语义约束，限制了其对软组织器官和新模态的通用性。我们提出LoRCA（LoRA循环适配），这是一种基于共享冻结DINOv3构建的循环一致风格转换框架，搭配模态特定的LoRA适配器，学习模态特定的表示并通过解码和对抗训练实现。LoRCA可在无需配对训练数据的情况下实现结构保留转换，冻结的骨干网络作为结构锚点，通过保留预训练的语义提取能力防止内容漂移。我们用Fréchet Inception Distance（FID）评估转换质量，通过互信息和Canny边缘保留评估结构保真度。LoRCA在转换质量和结构一致性上均优于CycleGAN。作为下游配准实用性的初步指标，我们发现风格转换后的图像在MatchAnything工具下，手动对齐的HiP-CT与组织学测试对之间的特征对应关系增加，表明LoRCA风格转换是迈向2D组织学切片到3D HiP-CT体积配准的有前景一步。
英文摘要
Hierarchical Phase-Contrast Tomography (HiP-CT) is a synchrotron based X-ray imaging technique that enables non-destructive, volumetric imaging of intact organs with multi-resolutions bridging 20 $μm$/voxel for whole organs to near-cellular resolution ($\sim$0.8 $μm$/voxel) in local regions. This offers the opportunity to bring volumetric whole-organ context to histology. However, nonlinear registration between H\&E histology and HiP-CT volumes is challenging due to the differences in feature representations of different colour spaces. Synthesis-before-registration methods have shown strong results in histology-to-MRI and histology-to-CT alignment. However, existing approaches either rely on manual anatomical contours or are trained from scratch without semantic constraints, limiting their generalisability to soft tissue organs and novel modalities. We propose LoRCA (LoRA Cycle Adaptation), a cycle consistent style translation framework built on a shared frozen DINOv3 with modality-specific LoRA adapters, learning modality-specific representations that are decoded and adversarially trained. LoRCA enables structure-preserving translation without requiring paired training data. The frozen backbone is intended to be a structural anchor that prevents content drift by preserving pretrained semantic-extraction capability. We evaluate translation quality using Fréchet Inception Distance (FID) and structural fidelity via mutual information and Canny edge preservation. LoRCA outperforms CycleGAN in both translation quality and structural consistency. As a preliminary indicator of downstream registration utility, we find that style-translated images yield increased feature correspondences under MatchAnything on manually aligned HiP-CT and histology test pairs, suggesting that LoRCA-style translation is a promising step towards 2D histological sections to 3D HiP-CT volumes registration.
