基于计算机学习技术整合斑块"生物-物理-化学"多模态超声信息评价斑块稳定性

Unstable atherosclerotic plaque is prone to break and leads to thrombosis, which is the main cause of acute cardiovascular events. Previous study showed that plaque stability was determined by plaque inflammation, vasa-vasorum neovascularization, composition and mechanical environment. However, present imaging techniques could only provide one or two of these plaque characteristics, and thus are less inaccurate in diagnosis. Exploring an diagnostic strategy collecting all of the characteristics of the plaque itself and the force environment simultaneously is of great value for vulnerable plaque identification. Previous findings by others and us have revealed that ultrasound has a significant advantage in evaluating the morphology, angiogenesis and blood flow of the plaque, which implicates that strengthening the diagnostic ability of ultrasound on the inflammation status and composition of the plaque would further improve the accuracy of ultrasound in vulnerable plaque identification. Very recently, we found that strain and strain rate modality of ultrasound could be used to image the plaque deformatoin, and the deforemation ability of the plaque was significantly decreased compared with the normal vascular wall and the stiffened wall with no plaues. This sheds lights on its application in revealing plaque composition. To this end, we firstly would try to acquire the inflammatory status and angiogenesis of the plaque at different stages and after statins therapy by using the ultrasound microbubble targeting VCAM-1. And then test the accuracy of the information from strain and strain rate imaging by comparing it with the pathological findings. Combining these two novel methods with the traditional ultrasound, it is possible to collect all the information reflecting the vulnverability of the plaque including the bio-physio-chemical characteristics, and then we will integrate and analyze the information by machine study technique. With such a study, we will provide new strategies for cardiovascular event prevention.

不稳定动脉硬化斑块容易破损并继发血栓，是急性心血管事件的主要原因。既往研究表明，斑块的炎症、血管新生及斑块成分分布等病理特征及斑块所处的力学环境共同决定斑块的稳定性；而现有影像学检测手段多针对斑块单一特征，准确性差。探讨能够全面检测斑块病理特征和力学环境的方法具有重要意义。我们前期研究发现，超声应变和应变率可用于分析斑块形变能力，斑块处应变及应变率较正常血管壁及无斑块处硬化管壁明显下降，且同步性差，为进一步评价斑块组成成分提供了可能。本项目拟在此基础上，构建不同稳定性斑块模型，观察靶向炎症分子VCAM-1超声造影剂的影像特征，结合病理学探讨其在斑块血管新生、炎症状态方面的检测能力；然后结合传统超声获取斑块结构和局部血流动力学参数；从而全面获取斑块生物-物理-化学信息；最后，将上述信息建成数据库，通过计算机学习技术整合这些数据并探讨其与斑块稳定性的数学关系，进而建立准确评价斑块稳定性的方法。

研究背景.动脉粥样硬化是一种慢性炎症性疾病，在动脉粥样硬化的发生发展过程中，易损斑块的出现通常会毫无症状的潜伏很长时间后突然脱落，引起临床急性心血管事件的发生。动脉粥样硬化的形成主要与脂质蓄积、炎症反应、蛋白酶活性及细胞外基质重构有关。易损斑块容易发生破裂并伴有血栓形成。常规的检查手段只能对疾病相对后期的阶段进行评价，并且并不能对病损的炎症状态做出直接的评判，在识别斑块易损性方面，未能对临床提供指导性意见。.主要研究内容.靶向超声造影技术被认为是评价动脉粥样硬化的一种有前景的无创检查手段。然而，在识别具有破裂风险的斑块方面，还缺乏有效的检查策略，这主要归因于易损斑块标记物的缺乏。有证据表明，连接粘附分子A（JAM-A又称为F11R）参与着白细胞的迁移和募集。然而，JAM-A在不稳定斑块的具体表达情况还未被阐明。考虑到JAM-A在斑块发生和发展中的作用，尤其是和易损斑块的形成相关，我们对JAM-A在动脉粥样硬化斑块中的表达水平进行了分析，并探索了靶向JAM-A微泡造影剂在识别易损斑块中的潜在价值。.重要结果.我们制作了兔颈动脉稳定/易损性斑块模型，应用qPCR、western blotting、免疫组化及免疫荧光等方法分析了斑块处不稳定标记物和JAM-A分子的表达情况。体外合成的靶向JAM-A微泡造影剂（MBJAM-A）及对照造影剂（MBIgG）经静脉注射入模型兔体内，并通过超声造影显像。在体外实验方面，我们研究了JAM-A与动脉粥样硬化发生、发展密切相关的四类细胞（内皮细胞、巨噬细胞、平滑肌细胞及成纤维细胞）中的表达差异。.科学意义.易损斑块分子标记物的鉴定是利用造影增强超声准确评估动脉粥样硬化患者风险的前提。本实验通过分析JAM-A在易损斑块中的表达情况，构建靶向JAM-A微泡造影剂，在兔颈动脉粥样硬化模型中进行成像分析，得出其对易损斑块识别的特性，并且易损斑块处超声造影信号增强的可能原因，提示靶向JAM-A微泡造影剂为斑块易损性的诊断及阻止其发展进程提供了新的方向。