Advanced X-Ray Micro-CT Methods for The Delineation of Tumour Heterogeneity in Lung Cancer

1)简要描述研究背景，包括潜在影响使用机器学习技术对ct图像进行肿瘤质地异质性分析，显示了预测癌症恶性程度和侵袭性[1]的潜力，这可能为患者的治疗计划提供信息。相衬x射线技术，从传播x射线[2]的相移中获得图像对比度，提高了对样品纹理和微观结构的灵敏度，而传统的基于衰减的x射线成像似乎大多是均匀的。本项目拟研究x射线相衬显微ct作为离体肺样本成像和三维表征的方法。新方法将与当前的组织病理学标准进行比较，这将是基本的事实。此外，机器学习和其他先进的计算技术将应用于图像的配准，分割和分类。这项工作将有助于扩大这些技术向新的体内临床应用的努力。2)目标和目标该项目将需要自下而上的方法，从成像设置的建模，优化和微调开始。随后将开发图像检索算法，以及用于临床组织样本试点研究的成像协议。从模型和离体组织样本的试验数据中，可以推断出该技术在体内实施的综合系统设计和规范。3)研究方法的新颖性边缘照明技术已应用于离体组织样品[3][4]的x射线相衬成像，并将其能力扩展到层析成像[5]。在这些里程碑的基础上，该项目将利用相衬图像的高分辨率和纹理来对肺癌的肿瘤异质性进行分类。4)与EPSRC的战略和研究领域保持一致。这是一个跨学科的项目，与EPSRC在工程、物理科学和医疗技术领域的主题保持一致。重点研究领域是医学成像、传感器和仪器仪表以及人工智能技术。5)任何参与的公司或合作者。该项目的基础设施包括通过UCL-Nikon繁荣合作伙伴关系的第一个商业原型x射线相衬扫描仪，以及两个独特的定制x射线

1) Brief description of the context of the research including potential impactAnalysis of tumour texture heterogeneity using machine learning techniques with CT-images shows potential for the prediction of cancer malignancy and aggressiveness [1], which may inform patient treatment plans. Phase-contrast x-ray techniques, which derive image contrast from phase-shifts in propagating x-rays [2], provide improved sensitivity to texture and microstructure in samples that would appear mostly homogenous to conventional, attenuation-based x-ray imaging. This project proposes to investigate x-ray phase-contrast micro-CT as a method to image ex-vivo lung samples and produce three dimensional representations. The new approach will be compared to the current histopathological standard, which will be the ground truth. Additionally, machine learning and other advanced computational techniques will be applied for the registration, segmentation, and classification of the images. This work will contribute to the efforts in scaling up these techniques towards new in-vivo clinical applications.2) Aims and ObjectivesThe project will entail a bottom-up approach, starting with the modelling, optimisation and fine tuning of the imaging setups. Image retrieval algorithm development will follow, along with an imaging protocol for a pilot study with clinical tissue samples. From both modelling and the pilot data on ex-vivo tissue samples, it will be possible to extrapolate a comprehensive system design and specification for the prospective in-vivo implementation of the technique.3) Novelty of Research MethodologyThe edge-illumination technique has been applied to the phase-contrast x-ray imaging of ex-vivo tissue samples [3][4], with its capabilities extended to tomographic imaging [5]. Building on these milestones, this project will exploit the high-resolution and texture of phase-contrast images to categorise tumour heterogeneity in lung cancer.4) Alignment to EPSRC's strategies and research areasThis is a strongly interdisciplinary project, aligning with EPSRC themes across the domains of engineering, physical sciences, and healthcare technologies. Key areas of research are medical imaging, sensors and instrumentation, and artificial intelligence technologies.5) Any companies or collaborators involvedThe infrastructure available to this project includes a first commercial prototype X-ray phase-contrast scanner through the UCL-Nikon Prosperity Partnership, and two unique custom-built X-ray