深层聚合残差密集网络的超声图像左心室分割

目的 超声图像是临床医学中应用最广泛的医学图像之一,但左心室超声图像一般具有强噪声、弱边缘和组织结构复杂等问题,其图像分割难度较大。临床上需要一种效率高、质量好的超声图像左心室分割算法。本文提出一种基于深层聚合残差密集网络（deep layer aggregation for residual dense network,DLA-RDNet）的超声图像左心室分割算法。方法 对获取的超声图像进行形态学操作,定位目标区域,得到目标图像。构建残差密集网络（residual dense network,RDNet）用于提取图像特征,并将RDNet得到的层次信息通过深层聚合（deep layer aggregation,DLA）的方式紧密融合到一起,得到分割网络DLA-RDNet,用于实现对超声图像左心室的精确分割。通过深监督（deep supervision,DS）方式为网络剪枝,简化网络结构,提升网络运行速度。结果 数据测试集的实验结果表明,所提算法平均准确率为95.68%,平均交并比为97.13%,平均相似性系数为97.15%,平均垂直距离为0.31 mm,分割轮廓合格率为99.32%。与6种分割算法相比,所提算法的分割精度更高。在测试阶段,每幅图像仅需不到1 s的时间即可完成分割,远远超出了专业医生的分割速度。结论 提出了一种深层聚合残差密集神经网络对超声图像左心室进行分割,通过主、客观对比实验表明本文算法的有效性,能够较对比方法更实时准确地对超声图像左心室进行分割,符合临床医学中超声图像左心室分割的需求。     

基于数学形态学及支持向量机的心率失常识别

为实现对不同类型的心电图自动分析,研究并提出了一种顺序筛选极大值的R波定位算法,并采用支持向量机(SVM)进行最后的心律失常心拍识别。定位算法以数学形态学为基础,结合心电图自身特点,定义R波筛选区间,避免了传统算法中的阈值选择;定位R波峰后以R波峰为中心提取不同类型的心率失常的心拍,选择径向基(RBF)支持向量机进行识别分类。使用MIT-BIH心率失常数据库文件进行实验仿真,结果表明,算法对含不同类型心拍的心电图R波峰正确检测率较高(99.36%),学习后的SVM能有效识别早搏、房颤、束支传导阻滞、正常等不用类型心拍,总体识别率达到99.75%。     

Alterations of the Sialylation Machinery in Brugada Syndrome

Brugada Syndrome (BrS) is an inherited arrhythmogenic disorder with an increased risk of sudden cardiac death. Recent evidence suggests that BrS should be considered as an oligogenic or polygenic condition. Mutations in genes associated with BrS are found in about one-third of patients and they mainly disrupt the cardiac sodium channel NaV1.5, which is considered the main cause of the disease. However, voltage-gated channel’s activity could be impacted by post-translational modifications such as sialylation, but their role in BrS remains unknown. Thus, we analyzed high risk BrS patients (n = 42) and healthy controls (n = 42) to assess an involvement of sialylation in BrS. Significant alterations in gene expression and protein sialylation were detected in Peripheral Blood Mononuclear Cells (PBMCs) from BrS patients. These changes were significantly associated with the phenotypic expression of the disease, as the size of the arrhythmogenic substrate and the duration of epicardial electrical abnormalities. Moreover, protein desialylation caused a reduction in the sodium current in an in vitro NaV1.5-overexpressing model. Dysregulation of the sialylation machinery provides definitive evidence that BrS affects extracardiac tissues, suggesting an underlying cause of the disease. Moreover, detection of these changes at the systemic level and their correlation with the clinical phenotype hint at the existence of a biomarker signature for BrS.     

Promising Therapies for Atrial Fibrillation and Ventricular Tachycardia

Sudden cardiac death due to arrhythmias, such as atrial fibrillation or ventricular tachycardia, account for 15–20% of all deaths. Myocardial infarction increases the burden of atrial fibrillation and ventricular tachycardia by structural and electrical remodeling of the heart. The current management of new-onset atrial fibrillation includes electric cardioversion with very high conversion rates and pharmacologic cardioversion, with less a than 50% conversion rate. If atrial fibrillation cannot be converted, the focus becomes the control of the symptoms ensuring a constant rhythm and rate control, without considering other contributory factors such as autonomic imbalance. Recently, a huge success was obtained by developing ablation techniques or addressing the vagal nerve stimulation. On the other hand, ventricular tachycardia is more sensitive to drug therapies. However, in cases of non-responsiveness to drugs, the usual therapeutic choice is represented by stereotactic ablative therapy or catheter ablation. This review focuses on these newly developed strategies for treatment of arrhythmias in clinical practice, specifically on vernakalant and low-level tragus stimulation for atrial fibrillation and stereotactic ablative therapy for drug-refractory ventricular tachycardia. These therapies are important for the significant improvement of the management of atrial fibrillation and ventricular tachycardia, providing: (1) a safer profile than current therapies, (2) higher success rate than current solutions, (3) low cost of delivery.     

基于STM32的便携式家用心电检测仪的设计

设计了一种基于STM32的便携式家用心电检测仪。心电电极采集体表单导联心电信号,经预处理电路对心电信号进行放大、滤波和电平抬升后,送至STM32中进行模/数转换和数字处理,在液晶屏上实时显示心电波形、心率和分析结果。实验表明,该心电仪能有效提取心电信号的特征点,准确测得心率,分析出4种常见心率失常症状,并可测得HRV的时域参数。     

基于支持向量机的心律失常识别方法

心律失常判别是心电数据处理的重要研究方向.提出一种基于支持向量机的心律失常判别方法来实现心律失常识别.根据临床诊断标准、结合计算机图像处理方法,提取心律参数特征和QRS复波形状特征,然后使用支持向量机对其进行分类,用MIT-BIH数据库对识别结果进行判断,并根据统计结果对该方法做出评价.     

Sarcomeres Morphology and Z-Line Arrangement Disarray Induced by Ventricular Premature Contractions through the Rac2/Cofilin Pathway

The most common ventricular premature contractions (VPCs) originate from the right ventricular outflow tract (RVOT), but the molecular mechanisms of altered cytoskeletons of VPC-induced cardiomyopathy remain unexplored. We created a RVOT bigeminy VPC pig model (n = 6 in each group). Echocardiography was performed. The histopathological alternations in the LV myocardium were analyzed, and next generation sequencing (NGS) and functional enrichment analyses were employed to identify the differentially expressed genes (DEGs) responsible for the histopathological alternations. Finally, a cell silencing model was used to confirm the key regulatory gene and pathway. VPC pigs had increased LV diameters in the 6-month follow-up period. A histological study showed more actin cytoskeleton disorganization and actin accumulation over intercalated disc, Z-line arrangement disarray, increased β-catenin expression, and cardiomyocyte enlargement in the LV myocardium of the VPC pigs compared to the control pigs. The NGS study showed actin cytoskeleton signaling, RhoGDI signaling, and signaling by Rho Family GTPases and ILK Signaling presented z-scores with same activation states. The expressions of Rac family small GTPase 2 (Rac2), the p-cofilin/cofilin ratio, and the F-actin/G-actin ratio were downregulated in the VPC group compared to the control group. Moreover, the intensity and number of actin filaments per cardiomyocyte were significantly decreased by Rac2 siRNA in the cell silencing model. Therefore, the Rac2/cofilin pathway was found to play a crucial role in the sarcomere morphology and Z-line arrangement disarray induced by RVOT bigeminy VPCs.     

心室纤颤和心动过速的小波非广度熵分析

为了对心动过速和心室纤颤进行准确而可靠的识别,提出了基于小波的多分辨率分析和熵相结合的分析方法.利用传统的Shannon熵信号分析方法获得的室颤和室速的识别率分别为96.4%和98.2%,而用非广度框架进行分析时获得的室颤和室速的识别率分别为100%和98.2%.表明作为检测室颤与室速的一个判据,Tsallis多分辨率熵(MRET)比Shannon多分辨率熵(MRE)具有更强的识别能力.该方法是一种稳定的、有效的特征提取方法,为其他非平稳生理信号的分析提供了新的手段.     

Time Needed to Improve Clinical Parameters By Daily Hemodialysis

Daily hemodialysis therapy (DHD), 2 hours, 6 times per week, is able to cure complications that persist on standard hemodialysis (SHD), 4 hours, 3 times per week. Cardiovascular manifestations (high blood pressure, left ventricular hypertrophy), nutritional deficient states, and postdialysis asthenia are improved during the first month of DHD therapy and are usually cured at 3 months. Daily hemodialysis may be considered as a rescue therapy. The next step will be to select which patients can return to the classical SHD therapy without recurrence of their complications.