动脉粥样硬化引起的易损斑块破裂已经严重危害到人类的健康，而血管内光学相干断层成像（IVOCT）凭借其高分辨率已经成为识别冠脉易损斑块的主要工具，但图像判读费时费力，通常还依赖于医生的经验。目前已有基于传统机器学习的研究实现了对单帧图像的分类，但这些信息不足以辅助医生确定治疗方案，仍然需要医生二次判读。基于Faster R-CNN（R-CNN，区域卷积神经网络），针对IVOCT图像中易损斑块的特点，在数据增强、预测框（BBox）编码、网络结构等方面进行了改进和优化，实现了对易损斑块的自动识别，并选取易损斑块的病变累积角度、纤维帽厚度、巨噬细胞浸润情况、浅表微钙化情况和血管狭窄程度作为指标，对易损斑块的破裂风险进行多方面评估。在公开数据集CCCV2017 IVOCT中进行训练，测试后取得了较好结果，该方法可推广应用于同类图像。

动脉粥样硬化是一种慢性炎症疾病, 是由于机体内脂质代谢异常从而导致富含胆固醇的脂蛋白在动脉壁中大量积累, 进而造成动脉壁增厚、变硬、弹性减退和管腔狭窄的病理改变。据报道, 低剂量弱激光疗法(LDLT)具有保护心血管的作用, 但其发挥作用的机制尚未明确。本研究旨在探索LDLT是否可以延缓动脉粥样硬化疾病的发展。用高脂饮食(HFD)诱导载脂蛋白E缺失(ApoE–/–)的小鼠18周, 以形成血管动脉粥样硬化斑块病变, 在HFD诱导的后10周里每日对小鼠进行LDLT处理。处理结束后, 取小鼠的主动脉、心脏和肝脏等组织进行组织学染色和聚合酶链式反应(PCR)试验, 以判断小鼠动脉粥样硬化病情的发展变化。试验结果显示: 相比对照组小鼠, LDLT处理后的小鼠主动脉斑块面积更小, 同时肝脏内脂质沉积也更少; 另外, 实时荧光定量PCR(RT-qPCR)试验结果显示, LDLT处理显著下调肝脏中参与胆固醇合成的甾醇调节元件结合转录因子2(Srebp2)和3-羟基-3-甲基戊二酰辅酶A还原酶(Hmgcr)的基因的表达。这些试验结果说明LDLT可以改善机体脂质代谢从而对动脉粥样硬化疾病具有改善作用。

动脉粥样硬化与各类心血管疾病密切相关,动脉粥样硬化会直接导致动脉血管壁变硬,产生一系列血管生理功能异常,诱发各类心血管疾病。检测动脉粥样硬化特征在心血管医学方面具有十分重要的意义。为了检验光声平台检测活体组织动脉粥样硬化的能力以及探索光声频谱分析在外周血管检测中的应用前景,本课题组开展了仿真和在体动物实验。采用K-wave仿真从理论上证明了光声频谱分析的可行性。本课题组建立了高脂饲料喂养的动脉粥样硬化小鼠模型,搭建了一套光声检测系统,收集了相应的信号数据进行光声频谱分析。动脉粥样硬化模型小鼠和正常小鼠光声信号频谱数据经方差分析后显著分离,实验数据的频谱分析结果与K-wave理论分析结果具有较好的一致性。频谱分析技术在检测小动脉血管粥样硬化方面具有一定的潜力,光声频谱分析技术有望成为小动脉血管硬化检测的有效手段之一。

皮肤胆固醇含量可以作为评价动脉粥样硬化的重要指标之一,现有的皮肤胆固醇含量检测主要基于实验室活检进行,缺少快速无创的检测技术和装备。针对以皮肤胆固醇含量为评价指标的动脉粥样硬化的早期快速筛查需求,本文提出了基于荧光光谱法的皮肤胆固醇快速无创检测方法,研发了一种皮肤胆固醇无创检测系统。为了提高测量的准确性和稳定性,该系统对温度引起的检测试剂荧光效率的波动进行了修正。本文结合气相色谱法对测量结果的准确性进行了验证,并通过检测正常人群和动脉粥样硬化高风险人群的皮肤胆固醇含量,明确了该系统的临床应用价值。本文的研究结果表明,462~520 nm波段内的平均荧光强度与温度的相关系数为-0.995(p<0.0001),可据此建立温度校准曲线对由温度差异引起的荧光波动进行修正。校正后,系统测量的皮肤胆固醇含量与气相色谱测量值的相关性显著,相关系数为0.905(p<0.0001)。在动脉粥样硬化高风险人群的筛查实验中,动脉粥样硬化高风险人群和正常人群的皮肤胆固醇检测结果具有显著差异(P=0.0004)。与现有技术相比,基于荧光光谱法的皮肤胆固醇检测技术具有测量快速无创等优势,为大规模开展动脉粥样硬化的早期风险筛查提供了先进技术手段。

Early detection of vulnerable plaques is the critical step in the prevention of acute coronary events. Morphology, composition, and mechanical property of a coronary artery have been demonstrated to be the key characteristics for the identification of vulnerable plaques. Several intravascular multimodal imaging technologies providing co-registered simultaneous images have been developed and applied in clinical studies to improve the characterization of atherosclerosis. In this paper, the authors review the present system and probe designs of representative intravascular multimodal techniques. In addition, the scientific innovations, potential limitations, and future directions of these technologies are also discussed.

Since changes in mechanical properties of biological tissues are often closely related to pathology, the viscoelastic properties are important physical parameters for medical diagnosis. A photoacoustic (PA) phase-resolved method for noninvasively characterizing the biological tissue viscoelasticity has been proposed by Gao et al. [G. Gao, S. Yang, D. Xing, “Viscoelasticity imaging of biological tissues with phase-resolved photoacoustic measurement,” Opt. Lett. 36, 3341–3343 (2011)]. The mathematical relationship between the PA phase delay and the viscosity–elasticity ratio has been theoretically deduced. Moreover, systems of PA viscoelasticity (PAVE) imaging including PAVE microscopy and PAVE endoscopy were developed, and high-PA-phase contrast images reflecting the tissue viscoelasticity information have been successfully achieved. The PAVE method has been developed in tumor detection, atherosclerosis characterization and related vascular endoscopy. We reviewed the development of the PAVE technique and its applications in biomedical fields. It is believed that PAVE imaging is of great potential in both biomedical applications and clinical studies.

Atherosclerosis has been recognized as a chronic inflammation disease, in which many types of cells participate in this process, including lymphocytes, macrophages, dendritic cells (DCs), mast cells, vascular smooth muscle cells (SMCs). Developments in imaging technology provide the capability to observe cellular and tissue components and their interactions. The knowledge of the functions of immune cells and their interactions with other cell and tissue components will facilitate our discovery of biomarkers in atherosclerosis and prediction of the risk factor of rupture- prone plaques. Nonlinear optical microscopy based on two-photon excited autofluorescence and second harmonic generation (SHG) were developed to image mast cells, SMCs and collagen in plaque ex vivo using endogenous optical signals. Mast cells were imaged with two-photon tryptophan autofluorescence, SMCs were imaged with two-photon NADH autofluorescence, and collagen were imaged with SHG. This development paves the way for further study of mast cell degranulation, and the effects of mast cell derived mediators such as induced synthesis and activation of matrix metalloproteinases (MMPs) which participate in the degradation of collagen.

皮肤胆固醇是评估动脉粥样硬化性疾病风险的新型生物标志物。为实现皮肤胆固醇的无创、快速检测，设计了一套基于漫反射光谱技术的检测系统。采用毛地黄皂苷-辣根过氧化物酶共聚物溶液模拟梯度浓度皮肤胆固醇。通过检测该共聚物溶液和受试者在体皮肤胆固醇，验证了系统的可行性。结果表明，在440~550 nm特征波段内，由相对漫反射率引入的参数S值与共聚物溶液浓度呈线性关系（相关系数r=0.992，P<0.01）；控制年龄、性别等因素后，受试者在体皮肤胆固醇无创检测结果与血液中总胆固醇和低密度脂蛋白胆固醇呈显著正相关，相关系数r分别为0.837（P<0.01）和0.778（P<0.01）。漫反射光谱技术为皮肤胆固醇的检测提供了一种无创、便捷的方法。在体皮肤胆固醇的无创检测有助于动脉粥样硬化性疾病的早期发现。

针对皮肤胆固醇无创检测的迫切需求, 设计了一种基于微型光谱仪的便携式和智能化的皮肤漫反射光谱实时测量系统。 配制毛地黄皂苷—辣根过氧化物酶共聚物溶液, 利用毛地黄皂苷可与皮肤胆固醇分子中的羟基特异性结合和辣根过氧化物酶能与TMB试剂(主要成分为3,3’,5,5’-四甲基联苯胺)反应显色的特性, 实现皮肤胆固醇的高灵敏和高特异的识别与指示, 并通过测量反应后颜色的变化程度来定量皮肤胆固醇的浓度。 以与人体皮肤结构相近的猪皮肤为实验对象, 采用萃取法获得梯度浓度的胆固醇样本, 利用上述光谱测量系统检测样本内的胆固醇浓度来验证该方法的可行性。 实验结果显示, 相对漫反射率能够区分不同浓度的胆固醇样本, 在特征单波长(442, 450和463 nm)处和特征波段442~500 nm内, 漫反射率强度因子均可以定量反映样本内的胆固醇浓度, 经线性拟合, 判定系数R2分别为0.960, 0.959, 0.958和0.958。 研究结果表明, 使用漫反射光谱技术可以实现皮肤胆固醇的快速无创检测, 将其应用于动脉粥样硬化性疾病风险评估, 对于该类疾病的防控具有重要意义。