基于特征矢量化的肺结节特征选择算法*

针对肺结节良恶性分类模型中特征选择过程无法避免特征多样性不受破坏的问题,提出一种将肺结节特征矢量化处理的特征选择方法。首先,假设每个肺结节特征都是由数据、类型构成的一个矢量。然后,按照特征类型添加特征到相应的特征子集,并分别利用Relief算法评价特征、特征子集的分类重要性。最后,通过动态阈值的方式筛选得到优化后的特征子集。在150个肺结节样本的分类实验中,采用本文算法所取得的敏感性为94.7%、特异性为93.7%、虚警率为5.2%、受试者工作特性曲线下面积为97.3%。分析表明,本文算法几乎不破坏肺结节特征的多样性,能够显著提高肺结节良恶性分类的准确性。     

人工肺-嗅觉系统集成与混合气体识别方法

针对仿人机器人的嗅觉及多种混合气体识别问题,提出一种人工肺-嗅觉系统(HALOS-I)及基于主动呼吸的气体识别方法.该系统硬件主要集成了微型真空泵、酒精/硫化氢/氨气/烟雾/甲烷5种气体传感器、单片机以及信号采集与处理电路;分别用K-均值聚类分析法、遗传算法结合神经网络(GA+BP)、三级级联神经网络(GA+3BP)进行了5种单一气体及4种混合气体的识别实验,结果表明:GA+BP算法仅对5种单一气体识别率达到90%以上,加入混合气体后识别率较低;GA+3BP算法除对硫化氢和烟雾的混合气体识别率为70%以外,对其余8种气体识别率均在90%以上,表明GA与多级级联BP神经网络相结合方法对多种单一及混合气体具有较高的识别率.     

基于改进LVQ算法的肺结节识别与分类

目前,肺癌的是发病率最高的肿瘤,若能在早期发现癌变并进行相应治疗,将极大的提高患者的生存率。肺癌的症状在早期表现为肺结节。以提高肺结节检测识别率并进行良恶性分类为目的,提出了一种改进的LVQ分类器算法。首先使用C-V算法对原始图像进行肺实质分割,再使用最优阈值法进行感兴趣区域提取,并进行特征提取和特征归一化。使用多次聚类算法检测肺结节。使用基于改进的LVQ分类器进行肺结节的良恶性进行分类。利用改进后的LVQ分类器在LIDC数据集上进行实验,得到了对良性结节的确诊率为87.3%,对恶性结节的确诊率为80.8%。实验结果表明,改进后的算法在良恶性结节分类上具有较高的确诊率,有助于提高医生的工作效率,实现肺结节的辅助发现。     

Spatial and functional relationships between air conduits and blood capillaries in the pulmonary gas exchange tissue of adult and developing chickens

The documented data regarding the three-dimensional structure of the air capillaries (ACs), the ultimate sites of gas exchange in the avian lung is contradictory. Further, the mode of gas exchange, described as cross-current has not been clearly elucidated. We studied the temporal and spatial arrangement of the terminal air conduits of the chicken lung and their relationship with the blood capillaries (BCs) in embryos as well as the definitive architecture in adults. Several visualization techniques that included corrosion casting, light microscopy as well as scanning and transmission electron microscopy were used. Two to six infundibulae extend from each atrium and give rise to numerous ACs that spread centrifugally. Majority of the ACs are tubular structures that give off branches, which anastomose with their neighboring cognates. Some ACs have globular shapes and a few are blind-ending tapering tubes. During inauguration, the luminal aspects of the ACs are characterized by numerous microvillus-like microplicae, which are formed during the complex processes of cell attenuation and canalization of the ACs. The parabronchial exchange BCs, initially inaugurated as disorganized meshworks, are reoriented via pillar formation to lie predominantly orthogonal to the long axes of the ACs. The remodeling of the retiform meshworks by intussusceptive angiogenesis essentially accomplishes a cross-current system at the gas exchange interface in the adults, where BCs form ring-like patterns around the ACs, thus establishing a cross-current system. Our findings clarify the mode of gas exchange in the parabronchial mantle and illuminate the basis for the functional efficiency of the avian lung.     

Construction of human monoclonal single-chain Fv antibodies against small-cell lung cancer by phage display libraries derived from cell-immunized SCID mice engrafted with human peripheral blood lymphocytes

In this study, we describe a phage display strategy to obtain human monoclonal single-chain Fv (scFv) antibodies binding target cancer cell surface proteins. By developing a cancer cell immunization protocol for SCID mice engrafted with human peripheral blood lymphocytes in combination with an antibody phage display method, we have isolated phage antibodies binding small-cell lung cancer cell line H889 by subtractive selection. One of the isolated scFv antibodies, 12EAb, recognized the E2 component of pyruvate dehydrogenase complex (PDC-E2) by immunoprecipitation according to MALDI-TOF MS analysis. Furthermore, we have confirmed the plasma membrane localization of PDC-E2 in small-cell lung cancer cells by immunocytochemistry and cell surface protein biotinylation, although PDC-E2 is usually located in the mitochondrial matrix. These results, including unique localization of identified antigens, were obtained by proteomic approaches. The present methods can be applied to generate human monoclonal scFv antibodies against tumor cells and to identify new molecular targets for immunotherapy and markers for diagnosis.     

Development and spatial organization of the air conduits in the lung of the domestic fowl, Gallus gallus variant domesticus

We employed macroscopic and ultrastructural techniques as well as intratracheal casting methods to investigate the pattern of development, categories, and arrangement of the air conduits in the chicken lung. The secondary bronchi included four medioventral (MVSB), 7-10 laterodorsal (LDSB), 1-3 lateroventral (LVSB), several sacobronchi, and 20-60 posterior secondary bronchi (POSB). The latter category has not been described before and is best discerned from the internal aspect of the mesobronchus. The secondary bronchi emerged directly from the mesobronchus, except for the sacobronchi, which sprouted from the air sacs. Parabronchi from the first MVSB coursed craniodorsally and inosculated their cognates from the first two LDSB. The parabronchi from the rest of the LDSB curved dorsomedially to join those from the rest of the MVSB at the dorsal border. Sprouting, migration, and anastomoses of the paleopulmonic parabronchi resulted in two groups of these air conduits; a cranial group oriented rostrocaudally and a dorsal group oriented dorsoventrally. The neopulmonic parabronchial network formed through profuse branching and anastomoses and occupied the ventrocaudal quarter of the lung. There were no differences in the number of secondary bronchi between the left and right lungs. Notably, a combination of several visualization techniques is requisite to adequately identify and enumerate all the categories of secondary bronchi present. The 3D arrangement of the air conduits ensures a sophisticated system, suitable for efficient gas exchange. Microsc. Res. Tech., 2008. (c) 2008 Wiley-Liss, Inc.     

Deep Belief Network for Lung Nodule Segmentation and Cancer Detection

Cancer disease is a deadliest disease cause more dangerous one. By identifying the disease through Artificial intelligence to getting the mage features directly from patients. This paper presents the lung knob division and disease characterization by proposing an enhancement calculation. Most of the machine learning techniques failed to observe the feature dimensions leads inaccuracy in feature selection and classification. This cause inaccuracy in sensitivity and specificity rate to reduce the identification accuracy. To resolve this problem, to propose a Chicken Sine Cosine Algorithm based Deep Belief Network to identify the disease factor. The general technique of the created approach includes four stages, such as pre-processing, segmentation, highlight extraction, and the order. From the outset, the Computerized Tomography (CT) image of the lung is taken care of to the division. When the division is done, the highlights are extricated through morphological factors for feature observation. By getting the features are analysed and the characterization is done dependent on the Deep Belief Network (DBN) which is prepared by utilizing the proposed Chicken-Sine Cosine Algorithm (CSCA) which distinguish the lung tumour, giving two classes in particular, knob or non-knob. The proposed system produce high performance as well compared to the other system. The presentation assessment of lung knob division and malignant growth grouping dependent on CSCA is figured utilizing three measurements to be specificity, precision, affectability, and the explicitness.     

基于文本与图像的肺疾病研究与预测EI北大核心CSCD

通过对目前现有的肺癌检测技术研究,发现大部分研究人员主要针对肺癌(Computed tomography,CT)影像进行研究,忽略了电子病历所隐藏的肺癌信息,本文提出一种基于图像与文本相结合的肺癌分类方法,从现有的基于深度学习的肺癌图像分类出发,引入了电子病历信息,使用Multi-head attention以及(Bi-directional long short-term memory,Bi-LSTM)对文本建模.实验结果证明,将电子病历信息引入到图像分类模型之后,对模型的性能有进一步的提升.相对仅使用电子病历进行预测,准确率提升了大约14%,精确率大约提升了15%,召回率提升了14%.相对仅使用肺癌CT影像来进行预测,准确率提升了3.2%,精确率提升了4%,召回率提升了4%.     

The Genetic Heterogeneity among Different Mouse Strains Impacts the Lung Injury Potential of Multiwalled Carbon Nanotubes

Genetic variation constitutes an important variable impacting the susceptibility to inhalable toxic substances and air pollutants, as reflected by epidemiological studies in humans and differences among animal strains. While multiwalled carbon nanotubes (MWCNTs) are capable of causing lung fibrosis in rodents, it is unclear to what extent the genetic variation in different mouse strains influence the outcome. Four inbred mouse strains, including C57Bl/6, Balb/c, NOD/ShiLtJ, and A/J, to test the pro‐fibrogenic effects of a library of MWCNTs in vitro and in vivo are chosen. Ex vivo analysis of IL‐1β production in bone marrow‐derived macrophages (BMDMs) as molecular initiating event (MIE) is performed. The order of cytokine production (Balb/c > A/J > C57Bl/6 > NOD/ShiLtJ) in BMDMs is also duplicated during assessment of IL‐1β production in the bronchoalveolar lavage fluid of the same mouse strains 40 h after oropharyngeal instillation of a representative MWCNT. Animal test after 21 d also confirms a similar hierarchy in TGF‐β1 production and collagen deposition in the lung. Statistical analysis confirms a correlation between IL‐1β production in BMDM and the lung fibrosis. All considered, these data demonstrate that genetic background indeed plays a major role in determining the pro‐fibrogenic response to MWCNTs in the lung.