Reliable and Accurate Registration and Review of Elemental and Molecular Imaging Data

该项目与国家物理实验室在个性化医学、早期诊断、多模态成像和复杂系统研究方面的战略重点直接一致。该项目与EPSRC的优先事项“成像模式、新型分析技术和创新，以改善疾病的预测、诊断和治疗”直接相关。结合和分析来自多种模式的数据对于充分理解发生在不同长度和分子尺度上的复杂生物过程至关重要。例如，质谱分析代谢物与组织化学和免疫组织化学信息之间的相关性可以将代谢活性与病变组织的结构和功能信息联系起来。空间分辨率的差异会使相关性和重叠分析变得困难，因为在一种模态中分辨率良好的小特征在另一种模态中可能会显得大而模糊。此外，在注册过程中，将变量之间相关性的绝对度量与准确性分开是否具有挑战性。我们将分别利用粒子诱导x射线发射（PIXE）和质谱成像（MSI）获取匹配的元素和分子信息数据集，以及来自血红素和伊红（H&E）的组织学信息。在配准后，我们将开发校正两个不匹配空间分辨率之间图像模糊的方法，以及评估配准质量的方法和指标，如结构相似性指数。NPL目前正在领导一个由CRUK资助的主要联盟，建立多模态质谱成像管道，以提供大范围长度尺度上的空间分辨率肿瘤代谢集成图。萨里大学也正在建立多模式方法，其不同的重点是将分子（通常是代谢）特征与元素映射方法相结合（EPSRC奖学金授予M. Bailey教授）。确定新的数据整合程序是这两个方案持续需要的领域。建立新的合作研究来解决这个问题是及时的。

This project is directly aligned with NPL strategic priorities in personalised medicine, early diagnosis, multi-modal imaging, complex systems research. The project is directly aligned to EPSRC priorities 'imaging modalities, novel analytical techniques and innovation needed to improve prediction, diagnosis and treatment of disease'. Combining and analysing data from multiple modalities is vital to fully understand complex biological processes that occur on different length and molecular scales. For examples the analysis of correlation between metabolites from mass spectrometry and information from histochemistry and immunohistochemistry can relate metabolic activity to structural and functional information in diseased tissues. Disparity in spatial resolution can make analysis of correlations and overlap difficult as a small well resolved feature in one modality may appear large and blurred in another. Additionally, is it challenging to separate absolute measures of correlation between variables from accuracy in the registration process. We will seek to acquire matched datasets of elemental and molecular information using particle-induced X-ray emission (PIXE) and mass spectrometry imaging (MSI) respectively, as well as histological information from haematoxylin and eosin (H&E). After registration, we will develop methods to correct for image blur between two mismatched spatial resolutions, and methods and metrics to assess the quality of the registration such as structural similarity index.NPL is currently leading a major consortium, funded by CRUK to establish multimodal mass spectrometry imaging pipelines to deliver spatially resolved integrated maps of tumour metabolism over wide ranging length scales. Multimodal approaches are also being established at The University of Surrey, with a different focus of combining molecular (typically metabolic) features with elemental mapping methods (EPSRC Fellowship awarded to Prof. M. Bailey). New data integration routines are identified as areas of ongoing need in both programmes. It is timely to establish new collaborative research to address this.