人工智能应用于青光眼临床筛查及卫生效益分析

目的:探讨应用人工智能系统对青光眼患者的检出率、筛查的时间成本和经济成本,总结分析其应用于青光眼筛查的优越性与可行性。方法:回顾性研究。收集2019-2020年使用腾讯觅影眼底照片辅助诊断系统进行青光眼筛查的53700例受检者,2012-2018年使用北京同仁医院眼科研究所远程眼科会诊中心的493221例患者,邯郸眼病研究中接受基础眼病筛查的6716名受检者资料。评价人工智能筛查系统对青光眼的检出率,并采用Pearson x~2检验,比较人工智能筛查系统与远程阅片平台、基础眼病筛查对于青光眼检出率的差异,并对三种筛查方式的时间成本和经济成本进行比较分析。结果:临床应用表明,其成本效益的可靠数据及主要指标:检出率、x~2值、时间成本、经济成本等效益显著。人工智能筛查系统对青光眼的检出率为6.56%,远程眼科会诊平台的检出率为2.78%,邯郸眼病研究的检出率为1.70%,人工智能筛查系统与另外两种模式对青光眼诊断检出率的差异有统计学意义(P0.05)。而且使用人工智能筛查系统的患者其医疗时间成本和经济成本花费较低。结论:可行性分析表明,相较于远程眼科会诊平台和基础眼病筛查,人工智能筛查系统对青光眼的检出率较高,同时患者的时间成本和经济成本都得到了大幅度降低,这为青光眼筛查提供了一种有效而快捷的方式,值得临床推广。     

Changes in the anatomic structures and cellular composition of lymph nodes after prenatal exposure to oleandomycin

Tracheobronchial and mesenteric lymph nodes have been investigated in 120 embryos, fetuses and newborn (5-, 14-, 30-day-old) rats by means of anatomical, histological and electron microscopical methods. Prenatal influence of oleandomycin phosphate on development of lymph nodes has been analysed. Effect of the drug on the mother--fetus system produces certain reactive changes in the anatomical structures responsible for the barrier-filtration and drainage functions of the lymph nodes. The immunocytopoietic function is inhibited, as demonstrates certain decrease of lymphatic line cells.     

Artificial neural network-aided image analysis system for cell counting.

BACKGROUND: In histological preparations containing debris and synthetic materials, it is difficult to automate cell counting using standard image analysis tools, i.e., systems that rely on boundary contours, histogram thresholding, etc. In an attempt to mimic manual cell recognition, an automated cell counter was constructed using a combination of artificial intelligence and standard image analysis methods. METHODS: Artificial neural network (ANN) methods were applied on digitized microscopy fields without pre-ANN feature extraction. A three-layer feed-forward network with extensive weight sharing in the first hidden layer was employed and trained on 1,830 examples using the error back-propagation algorithm on a Power Macintosh 7300/180 desktop computer. The optimal number of hidden neurons was determined and the trained system was validated by comparison with blinded human counts. System performance at 50x and lO0x magnification was evaluated. RESULTS: The correlation index at 100x magnification neared person-to-person variability, while 50x magnification was not useful. The system was approximately six times faster than an experienced human. CONCLUSIONS: ANN-based automated cell counting in noisy histological preparations is feasible. Consistent histology and computer power are crucial for system performance. The system provides several benefits, such as speed of analysis and consistency, and frees up personnel for other tasks.     

A diagnostic expert system

The introduction of new technologies in the field of electronics has influenced the development of technical equipment over the last few years. The progressive miniaturization of integrated circuits makes possible an expansion of the spectrum of functions offered by this equipment. This also applies to medical technology. These more complex units call for new methods of fault detection and diagnosis. In addition to analytical redundancy, tools developed by the artificial intelligence research community, such as expert systems, are becoming more and more important for fault diagnosis. On the basis of a realized diagnosis expert system the possibilities as well as the limits of such system are discussed. Also, possible future developments of artificial intelligence, like machine learning, are considered.     

组学时代下机器学习方法在临床决策支持中的应用

随着组学技术的不断发展,对于不同层次和类型的生物数据的获取方法日益成熟。在疾病诊治过程中会产生大量数据,通过机器学习等人工智能方法解析复杂、多维、多尺度的疾病大数据,构建临床决策支持工具,辅助医生寻找快速且有效的疾病诊疗方案是非常必要的。在此过程中,机器学习等人工智能方法的选择显得尤为重要。基于此,本文首先从类型和算法角度对临床决策支持领域中常用的机器学习等方法进行简要综述,分别介绍了支持向量机、逻辑回归、聚类算法、Bagging、随机森林和深度学习,对机器学习等方法在临床决策支持中的应用做了相应总结和分类,并对它们的优势和不足分别进行讨论和阐述,为临床决策支持中机器学习等人工智能方法的选择提供有效参考。     

Morphological characteristics of the thymus gland of albino rats after prenatal administration of indomethacin

By means of histological, histochemical and electron microscopical methods the thymus of white rat fetuses and offspring has been investigated during various age periods after indomethacin++ influence (2.5 mg/kg). In the fetuses retardation in separation of the gland parenchyma into lobules has been revealed. During the first two weeks of life the section area of the medulla decreases. Amount of lymphoid cells decreases; small and degenerating lymphocytes decrease in their number, while the part of lymphoblasts, middle lymphocytes and figures of mitosis increases. Enzymatic activity in the nervous structures of the organ is inhibited. Some essential disturbances of the intracellular structures are revealed; they demonstrate certain destructive changes. The data obtained show a decreasing function of the thymus after the prenatal influence of indomethacin++ during the first month of life, which is especially manifested during the first two weeks.     

Artificial senses for characterization of food quality

1 Introduction There are five primary senses: vision, hearing, smeltaste and touch. They are the functions of eyes, ears, nosetongue and skin respectively. These functions are coordinated by the brain. In the characterization of foodpeople sense the samples and differentiate “good”from“bad”along a continuum. However, the human sensorysystem is l subjective, with mental and physical inconsistencies, and time consuming. Artificial senses such asmachine vision, the electronic ear, elect…     

Catecholaminergic and serotoninergic brain structures in European green frogs: an autoradiographical study

Aminergic brain structures have been investigated by means of light microscopical autoradiography after injection of the tritiated catecholamines noradrenaline and dopamine and the indoleamine (or tryptamine) serotonin into the brain cavity of frogs of the Rana esculenta complex. These amines are fairly specifically taken up by catecholaminergic and serotoninergic neurons, respectively, which are located in structures like the catecholaminergic preoptic recess organ; the mixed catecholaminergic-serotoninergic paraventricular organ/nucleus infundibularis-complex and nucleus reticularis mesencephali; the telencephalic septal and striatal areas and the tectum opticum, which contain many catecholaminergic axon terminals; the habenular area, which contains serotoninergic axon terminals. The autoradiographical data on the location and the nature of these aminergic brain structures agree well with the mainly fluorescence microscopical and immunocytochemical data from the literature. The autoradiographical detection method can be combined at the light and the electron microscopical level with other histological, histochemical, or immunohistochemical techniques in one and the same preparation, and the results of the different treatments may eventually be made visible simultaneously.     

A user-friendly biological workstation

Learning methods developed by artificial intelligence research teams are very efficient for biological sequences analysis but they need running on large computers accessed by terminals. These computers are interfaced with standard displays involving long and unpleasant alphanumerical data handling. The "biological work station" is a personal computer with a color graphic screen providing a user-friendly interface for the artificial intelligence learning programs running on large computers. It provides to biologist a graphical convenient tool for sequence analysis built with efficient man-machine communication methods such as multiwindows, icons and mouse selection. It allows the biologist to edit and display sequences in an efficient and natural way, showing off directly on color pictures the data and the results of learning programs.     

Blood morphology of the European eel (Anguilla anguilla). IV. Thrombocytes and their developmental stages]

By the aid of histological special stainings, cytochemical proofs, phase contrast and electron microscopical methods the thrombocytes (spindle cells) of the European ell (Anguilla anguilla) and their stages of development were described as an independent series of cells. After an artificial blood loss thrombocytoblasts, prothrombocytes and mature thrombocytes could distinctly be demonstrates. Besides the kidney the spleen was found to be the main place of the thrombopoiesis. Phase contrast investigations provided evidence for the relatively strong locomotion of these cells. By means of cytochemical proofs a further differentiation of the thrombocytes from other white blood cells can be made. Electron microscopical investigations of the thrombocytes yielded a great similarity to the thrombocytes of the amphibians. The endoplasmatic ring and particular inclusions could be shown to be characteritstic features of the fine structure.     

基于机器学习的药物-靶标相互作用预测*

近年来,随着计算机硬件、软件工具和数据丰度的不断突破,以机器学习为代表的人工智能技术在生物、基础医学和药学等领域的应用不断拓展和融合,极大地推动了这些领域的发展,尤其是药物研发领域的变革。其中,药物-靶标相互作用(drug-target interactions, DTI)的识别是药物研发领域中的重要难题和人工智能技术交叉融合的热门方向,研究人员在DTI预测方面做了大量的工作,构建了许多重要的数据库,开发或拓展了各类机器学习算法和工具软件。对基于机器学习的DTI预测的基本流程进行了介绍,并对利用机器学习预测DTI的研究进行了回顾,同时对不同的机器学习方法运用于DTI预测的优缺点进行了简单总结,以期对开发更加有效的预测算法和DTI预测的发展提供帮助。     

Synthesizing cellular intelligence and artificial intelligence for bioprocesses

Microbial processes operated under realistic conditions are difficult to describe by mechanistic models, thereby limiting their optimization and control. Responses of living cells to their environment suggest that they possess some "innate intelligence". Such responses have been modeled by a cybernetic approach. Furthermore, the overall behavior of a bioreactor containing a population of cells may be described and controlled through artificial intelligence methods. Therefore, it seems logical to combine cybernetic models with artificial intelligence to evolve an integrated intelligence-based strategy that is physiologically more faithful than the current approaches. This possibility is discussed, together with practical considerations favoring a hybrid approach that includes some mathematical modeling.     

Intelligent software for laboratory automation

The automation of laboratory techniques has greatly increased the number of experiments that can be carried out in the chemical and biological sciences. Until recently, this automation has focused primarily on improving hardware. Here we argue that future advances will concentrate on intelligent software to integrate physical experimentation and results analysis with hypothesis formulation and experiment planning. To illustrate our thesis, we describe the 'Robot Scientist' - the first physically implemented example of such a closed loop system. In the Robot Scientist, experimentation is performed by a laboratory robot, hypotheses concerning the results are generated by machine learning and experiments are allocated and selected by a combination of techniques derived from artificial intelligence research. The performance of the Robot Scientist has been evaluated by a rediscovery task based on yeast functional genomics. The Robot Scientist is proof that the integration of programmable laboratory hardware and intelligent software can be used to develop increasingly automated laboratories.     

The Lateral Line System of Hagfishes (Craniata: Myxinoidea)

The morphology of the peripheral lateral line system in Eptatretus stoutii is described using a combination of electron microscopical, histological and immunocytochemical techniques. The epidermis and cranial nerves of Myxine glutinosa were also examined for evidence of a lateral line system. Together with results previously published by other researchers, we conclude that most or all eptatretid hagfish possess a lateral line system composed of shallow trenches, or grooves, lined with a single class of flask-shaped receptive cells. No species of myxinid hagfish has been shown to possess any component of this system. Three groups of lateral line grooves are present in eptatretids: preoptic, dorsal postoptic, and ventral postoptic groups. The preoptic grooves are innervated by a single pair of ganglionated cranial nerves and the two postoptic series of grooves are innervated by a pair of cranial nerves which each possess two ganglia, possibly as a result of fusion of two pairs of ganglionated nerves. Morphometric analysis indicates that these grooves continue to grow throughout life and that the variability in the distribution of grooves between animals may result from developmental instability in lateral line ontogeny. The morphology of the receptive organs is compared to neuromasts of the vertebrate lateral line system and ciliated receptor cells of other chordates. The morphology of hagfish lateral line receptors may represent the primitive condition for craniate lateral line organs, or more likely, they are highly derived as a result of regressive evolution. © 1997 The Royal Swedish Academy of Sciences. Published by Elsevier Science Ltd.     

Imaging and artificial intelligence for progression of age-related macular degeneration

Age-related macular degeneration (AMD) is a leading cause of severe vision loss. With our aging population, it may affect 288 million people globally by the year 2040. AMD progresses from an early and intermediate dry form to an advanced one, which manifests as choroidal neovascularization and geographic atrophy. Conversion to AMD-related exudation is known as progression to neovascular AMD, and presence of geographic atrophy is known as progression to advanced dry AMD. AMD progression predictions could enable timely monitoring, earlier detection and treatment, improving vision outcomes. Machine learning approaches, a subset of artificial intelligence applications, applied on imaging data are showing promising results in predicting progression. Extracted biomarkers, specifically from optical coherence tomography scans, are informative in predicting progression events. The purpose of this mini review is to provide an overview about current machine learning applications in artificial intelligence for predicting AMD progression, and describe the various methods, data-input types, and imaging modalities used to identify high-risk patients. With advances in computational capabilities, artificial intelligence applications are likely to transform patient care and management in AMD. External validation studies that improve generalizability to populations and devices, as well as evaluating systems in real-world clinical settings are needed to improve the clinical translations of artificial intelligence AMD applications.     

一种非酒精性脂肪性肝病的超声影像人工智能辅助诊断系统开发

摘要 目的：结合人工智能方法设计针对肝脏超声影像的辅助诊断系统，辅助医生对大样本肝脏超声影像数据的标准化和高效化诊断，实现基于肝脏超声图像的非酒精性脂肪性肝病的精准诊断。方法：通过开发肝脏超声影像的识别与分类、脂肪肝分级分析和肝脏脂肪含量定量分析三个模块，建立一套非酒精性脂肪性肝病的超声影像人工智能辅助诊断系统，该系统能够自动区分输入到系统中不同采样视野的超声影像类型，并对肝脏超声图像进行数字化分析，给出待测超声图像是否呈现脂肪肝以及其肝脏脂肪含量的百分比值。结果：本研究中的超声图像识别分类模块可高通量区分出肝肾比图像和衰减率图像的两类超声影像，其分类的准确率达100%。脂肪肝分级分析模块在测试集数据的准确率达到84%，展现出可胜任辅助医生诊断的能力。基于人工肝脏脂肪含量定量方法开发的肝脏脂肪含量定量分析模块的准确率达到67.74%。结论：本研究已开发出一套基于肝脏超声影像的智能辅助诊断系统，可以辅助医生快速、简单、无创地筛选出潜在患有脂肪肝的患者，虽然现阶段实现肝脏脂肪定量分析仍有难度，但已展现出较大的临床应用潜力。     

Development of blood vessels in the human eyeball

Blood vessels of 167 eyeballs of human embryos and fetuses have been studied, using macro- microscopical, histological and morphometrical methods. Time of appearance and differentiation of blood vessels in the eye anlage has been determined. Topography and architecture of the vitreous body artery are described in detail, as well as the vascular tunic of the lens and the pupil membrane. Regularities and time of reversal development of the vascular formations mentioned are followed during the whole intrauterine development. The data obtained add to extend our knowledge on the role of disturbances of the normal morphogenesis of the eyeball blood bed in pathogenesis of its congenital diseases.     

Non-aqueous permanent mounting for immunofluorescence microscopy

It is generally assumed that an aqueous mounting medium is necessary for the preservation of immunofluorescent-labelled microscopical preparations and polyvinyl alcohol-based solutions (e.g. Mowiol) being the most frequently used mounting media; however, both the quality and intensity of the fluorescence signal in most immunolabelled preparations after aqueous mounting slowly diminish with time, and finally, samples become unsuitable for examination. In the present work, we describe a very simple and rapid non-aqueous mounting procedure for cultured cells and tissue sections, which preserves the fluorescent signal in an excellent way after immunodetection or use of other specific labelling methods. It is based on the current histological protocol in which, after fluorescence labelling, preparations are dehydrated in ethanol, cleared in xylene and mounted in DePeX. Using this non-aqueous mounting medium, the fluorescent signal remains high and stable, allowing a suitable and permanent preservation of labelled and counterstained microscopical preparations.     

Evolutionary combinatorial chemistry, a novel tool for SAR studies on peptide transport across the blood-brain barrier. Part 2. Design, synthesis and evaluation of a first generation of peptides.

The use of high-throughput methods in drug discovery allows the generation and testing of a large number of compounds, but at the price of providing redundant information. Evolutionary combinatorial chemistry combines the selection and synthesis of biologically active compounds with artificial intelligence optimization methods, such as genetic algorithms (GA). Drug candidates for the treatment of central nervous system (CNS) disorders must overcome the blood-brain barrier (BBB). This paper reports a new genetic algorithm that searches for the optimal physicochemical properties for peptide transport across the blood-brain barrier. A first generation of peptides has been generated and synthesized. Due to the high content of N-methyl amino acids present in most of these peptides, their syntheses were especially challenging due to over-incorporations, deletions and DKP formations. Distinct fragmentation patterns during peptide cleavage have been identified. The first generation of peptides has been studied by evaluation techniques such as immobilized artificial membrane chromatography (IAMC), a cell-based assay, log Poctanol/water calculations, etc. Finally, a second generation has been proposed.     

Reasoning methods in medical consultation systems: artificial intelligence approaches

It has been argued that the problem of medical diagnosis is fundamentally ill-structured, particularly during the early stages when the number of possible explanations for presenting complaints can be immense. This paper discusses the process of clinical hypothesis evocation, contrasts it with the structured decision making approaches used in traditional computer-based diagnostic systems, and briefly surveys the more open-ended reasoning methods that have been used in medical artificial intelligence (AI) programs. The additional complexity introduced when an advice system is designed to suggest management instead of (or in addition to) diagnosis is also emphasized. Example systems are discussed to illustrate the key concepts.