节点文献
基于数据挖掘技术的重症肌无力疾病五脏相关性研究
Research on the TCM Five-viscera Correlation Theory of Myasthenia Gravis Based on Data Mining
【作者】 饶媛;
【导师】 刘小斌;
【作者基本信息】 广州中医药大学 , 中医医史文献学, 2010, 博士
【摘要】 邓铁涛教授提出的五脏相关学说是指在人体大系统中,心、肝、脾、肺、肾及其相应的六腑、四肢、皮毛、筋、脉、肉、五官七窍等组织器官分别组成五个脏腑系统,在生理情况下,本脏腑系统内部、脏腑系统与脏腑系统之间、脏腑系统与人体大系统之间、脏腑系统与自然界、社会之间,存在着横向、纵向和交叉的多维联系,相互促进与制约,以发挥不同的功能,协调机体的正常活动;在病理情况下,五脏系统又相互影响。五脏相关学说是介于基础和临床之间的、具有临床指导普遍性的应用理论。它将基础理论推进到辨证论治的实践层面,有效拓展中医病机学学科领域。五脏相关学说解释临床复杂病机时的重视整体、层次分明、主次有别、标本兼顾的特点使其在解决疑难杂症方面具有独特的优势。重症肌无力属临床疑难杂病,具有病情反复、疗程长、治疗棘手等特点。中医治疗重症肌无力疗效显著,但由于中医辨证知识的模糊性和不一致性,缺乏大样本的的临床调查以及客观严谨的的数据分析研究,大大影响了中医药在重症肌无力诊治方面的评价及方法的推广。数据挖掘技术善于从缺乏先验信息的海量数据中发现隐含的有意义的知识,寻找未知的或验证已知的规律性。从这个意义上来说,数据挖掘技术恰好可以解决中医证候研究中的关键技术难题,是中医药现代化研究的重要组成部分。本研究在对五脏相关学说的研究中,以重症肌无力为切入点,建立重症肌无力疾病数据库,运用统计分析和数据挖掘手段,通过对临床症状、脏腑病位、常用中药及归经的分析,从不同角度探讨重症肌无力的五脏相关性,总结重症肌无力中医脏腑病机,构建重症肌无力五脏相关理论框架,为重症肌无力的临床诊治提供参考。研究目的采用数据挖掘技术挖掘和分析重症肌无力疾病中医证候的五脏相关性,总结重症肌无力中医五脏相关病机模式,构建重症肌无力中医五脏相关理论框架。研究方法1.按照西医诊断及分型标准,确定纳入病例,共收集重症肌无力病例447例。2.制定重症肌无力疾病中医五脏相关临床信息采集表,收集重症肌无力患者临床信息资料,填写调查问卷表。3.录入信息,建立重症肌无力疾病数据库。4.进行数据预处理,包括数据的规范化、数据的量化及数据清洗等,为数据挖掘作准备。5.采用SPSS统计软件对重症肌无力性别、年龄、临床症状、合并症、五脏病变、常用药物、药物归经等变量进行频数统计分析。6.采用KNEX. SQL Server等数据挖掘软件及稳健回归、Naive Bayes等算法对重症肌无力症状、五脏病变规律及不同分型重症肌无力关键诊断因素进行数据挖掘,探讨重症肌无力中医脏腑病机,及重症肌无力西医分型与中医辨证的相关性。研究结果1.频数统计结果(1)重症肌无力发病女性患者较男性患者多,年龄以21-40岁为发病高峰。(2)重症肌无力多合并胸腺异常,常伴有甲状腺疾病、类风湿关节炎、系统性红斑狼疮和多发性肌炎等自身免疫性疾病,亦容易合并感冒及感染。(3)重症肌无力常见临床症状有:眼睑下垂,或复视、或斜视、或眼球活动受限,肢体无力,颈软无力,眼睑疲劳,倦怠,咀嚼无力吞咽困难,呼吸气短等,证候特点以虚为本。(4)五脏病变频数统计以脾脏受累贯穿始终,其次为肾、肝、肺、心。多脏同时受累频数统计,四脏同时受累频次最高,其次为三脏同时受累、两脏同时受累、五脏同时受累、脾脏单独受累。(5)各临床分型五脏病变频数统计,Ⅰ型依次为脾、肝、肺、心、肾,Ⅱ-A型依次为脾、肝、肾、肺、心为主,Ⅱ-B型依次为脾、肾、肺、肝、心,Ⅲ型依次为脾、肾、肺、肝、心,Ⅳ型依次为脾、肾、肺、心、肝,Ⅴ型依次为脾、肾、肺、肝、心。重症肌无力危象五脏病变频数统计依次为:脾、肾、肺、心、肝。(6)常用中药以黄芪、甘草、五爪龙、白术、升麻、柴胡、陈皮、当归、党参等补气健脾之品频次最高,山萸肉、杜仲、巴戟天、肉苁蓉、何首乌等补肾温阳药也较为常用。提示重症肌无力与脾、肾关系最为密切。(7)药物归经显示,入脾胃系最多,占76.9%;其次为入肝胆系、入肾膀胱系、入肺大肠系、入心小肠系。2.数据挖掘结果(1)采用KNEX软件稳健回归数算法数据挖掘表明:1)症状重要性分析显示咀嚼无力吞咽困难是重症肌无力中最重要的症状,紧随其次的有:肢体无力、流涎、呼吸无力、构音不清、眼睑下垂、咳嗽、倦怠、饮水反呛、便溏、视物模糊、颈软无力、肢体困重、表情淡漠、肠鸣、胸闷。提示咀嚼无力吞咽困难可以作为判断疾病病情轻重的一个指征。2)除脾外,五脏重要性分析显示重症肌无力中肾重要性最大,其次为肺、心。肝未能在挖掘系统重要性结果中出现。提示重症肌无力与脾、肾、肺关系最为密切。(2)采用SQL Server软件Naive Baves算法数据挖掘表明:1)各临床分型重症肌无力关键诊断因素如下。①Ⅰ型主要因素:年龄小于19岁、未出现肢体无力;次要因素:未出现咀嚼无力吞咽困难、倦怠等症状。②Ⅱ-A型主要因素:未出现咀嚼无力吞咽困难;次要因素:未出现构音不清、呼吸无力、流涎等症状。③Ⅱ-B型主要因素:出现肢体无力;次要因素:出现咀嚼无力吞咽困难、构音不清。④Ⅲ型主要因素:出现呼吸无力;次要因素:出现咀嚼无力吞咽困难、喘促、言语低嘶、膝软无力等症,合并肺部感染。⑤Ⅳ型主要因素:合并重症肌无力危象,出现呼吸无力;次要因素:出现咀嚼无力吞咽困难、咯痰、流涎、构音不清等症状,合并胸腺瘤。2)除脾外,各临床类型重症肌无力主要病变脏腑如下:①Ⅰ型主要因素:未累及肾;次要因素:未累及肺、心。②Ⅱ-A型主要因素:未累及肺;次要因素:未累及肾。③Ⅱ-B型主要因素:累及肾;次要因素:累及肺。④Ⅲ型主要因素:累及肺;次要因素:累及肾、心。⑤Ⅳ型主要因素:累及肺;次要因素:累及肾、心。研究结论1.重症肌无力以脾脏受累为主,与肾、肺密切相关,也可责之于心、肝。这与邓铁涛教授对本病概括的基本病机“脾胃虚损,五脏相关”相符。2.咀嚼无力吞咽困难、肢体无力、呼吸无力、构音不清是鉴别各临床类型重症肌无力的关键因素。其中咀嚼无力吞咽困难较其它症状相比,对于重症肌无力病情判断最为重要。它是轻度全身型(Ⅱ-A型)与中度全身型(Ⅱ-B型)鉴别要点,对出现该症状患者,临床应加强观察,防止病情加重甚至危象发生。3.各临床类型重症肌无力中Ⅰ型以脾受累为主,可累及肝;Ⅱ-A型以脾受累为主,可累及肾;Ⅱ-B型以脾肾受累为主,可累及肺;Ⅲ型以脾肺肾受累为主,可累及心;Ⅳ型以脾肺肾受累为主,可累及心。重症肌无力危象表现为脾肾肺心肝同病。提示病情越轻,涉及病变脏腑越少,而病情越重,涉及病变脏腑越多。4.重症肌无力发病中,以脾脏受损贯穿始终,随着病情的加重,可出现二脏、三脏、四脏、五脏合病。重症肌无力五脏相关病机模式有:脾肝同病、脾肾同病、脾肾肝同病、脾肾肺同病、脾肺肾心同病、脾肾肺心肝同病。5.现行中医内科临床辨证标准往往是基于单脏证的主要症候表现,并不反映疾病往往以多脏证为主时,相关证候的组合排列规律及其辨证诊断价值。邓铁涛教授“脾胃虚损,五脏相关”理论不但在指导重症肌无力诊治有意义,同时对其它危重病疑难病诊治也有普适性。6.本研究从客观临床信息出发,引入数据挖掘技术,对重症肌无力脏腑病机进行探讨,验证了邓老学术经验的正确性,说明数据挖掘技术用于探讨中医五脏相关学说具有一定的应用前景。
【Abstract】 The TCM five-viscera correlation theory is proposed by famous professor Deng Tietao. He thinks the heart, liver, spleen, lungs, kidneys and correlative six hollow organs, four limbs, skin, tendon, blood vessel, muscles, five-sense organs and seven orifices make up human body’s five-viscera system. In physiology conditions, horizontal, vertical, and crossed multidimensional relations exist in zang-fu interior, between zang-fu system and zang-fu system, between zang-fu system and nature, society, which promote and inhibit reciprocally, exert different functions, and harmonize organismic normal activities. In pathology conditions, the five-viscera system influences reciprocally. The theory combines the five elements theory with the zang-fu theory, explains human physiology and correlative pathology, and guides clinic diagnosis and treatment.Myasthenia gravis is a clinical difficult disease, with features of repeated pathogenetic condition, long course of treatment, difficult treatment and poor prognosis. Chinese medicine therapeutic effect of myasthenia gravis is significant. However, TCM differentiation of symptoms knowledge is ambiguity and inconsistency, and lack of large sample clinical investigation as well as objective and rigorous analysis of data which greatly affects the evaluation of the Chinese medicine diagnosis and treatment of myasthenia gravis. Data mining is good at discovering hidden and meaning knowledge from the mass data. So data mining can be applied for study on TCM syndromes.1. Objective:Study the TCM five-viscera correlation theory of myasthenia gravis based on data mining, and conclude TCM Zang-fu pathogenesis and progress model of myasthenia gravis viscera.2. Methods:(1)Collected 447 cases of patients according to diagnosis and classification standards of myasthenia gravis(2)Designed five-viscera correlation clinical information collection questionnaire of myasthenia gravis, collected clinical information on patients with myasthenia gravis, filled out the questionnaire form.(3)Inputed information and established a database of myasthenia gravis disease.(4)Data preprocessing, including data standardization, quantification and data cleaning, preparation for data mining.(5)Frequency statistical analysis of sex, age, clinical symptoms, complications, five internal organs diseases, frequently used traditional Chinese medicine, herbals meridian distribution on myasthenia gravis by using SPSS statistical software.(6) Approached TCM Zang-fu pathogenesis of myasthenia gravis and the relevance of different types myasthenia gravis and TCM Zang-fu pathogenesis by using KNEX, SQL Server data mining software and robust regression, Naive Bayes.3. Results:(1) Frequency statistical analysis1)Myasthenia gravis incidence of women is higher than men, aged 21-40 years is morbility peak.2)Myasthenia gravis is often complicating with thymic abnormalities, and thyroid disease, rheumatoid arthritis, systemic lupus erythematosus, polymyositis and other autoimmune diseases.3)Common clinical symptoms of myasthenia gravis include:ptosis, general fatigue, eyelid fatigue, lassitude, weakness of chewing swallowing, by deficiency syndromes. The characteristic of symptoms is weak.4)About five viscera lesions, the most common four viscera simultaneously is involved, followed by three viscera simultaneously involved, the two viscera simultaneously involved, the five viscera simultaneously involved, the spleen alone involved. Cumulative frequency statistics for the five organs of myasthenia gravis indicate the spleen is involved all along, followed by kidney, liver, lung, heart.5)Ⅰtype damages spleen, liver mainly,Ⅱ-A type damages spleen, liver, kidney mainly,Ⅱ-B type damages spleen, kidney, liver mainly,Ⅲtype damages spleen, kidney, lung, liver mainly,Ⅳtype damages spleen, kidney, lung, heart mainly.6)Frequently traditional Chinese medicine include herbals of invigorating vital energy and spleen such as astragalus, licorice, Wu Zhao Long, Atractylodes, Cimicifuga, Bupleurum, tangerine peel, Chinese angelica, Codonopsis, and herbals of invigorating kidney and warming yang such as cornus, Eucommia, Morinda officinalis, Cistanche, Polygonummultiflorum, etc. The results indicate that morbility of myasthenia gravis is most closely with spleen and kidney.7)Meridian distribution of herbals show up the percentage of the spleen system is maximum, accounting for 73.7%, followed by the liver system, kidney system, lung system, heart system.(2) Data mining1)Results of KNEX software by using robust regression algorithms:①Weakness of chewing swallowing is the most important symptom of myasthenia gravis symptoms, followed by general fatigue, salivation, respiratory weakness, unclear articulation, ptosis, cough, fatigue, bucking when drinking water, loose stools, blurred vision, neck weakness, limbs weary and weight, apthy, rugitus, chest distress. It indicates that weakness of chewing swallowing may serve as an indication of disease severity.②In addition to the spleen, the importance analysis of other viscera on myasthenia gravis shows the kidney is the most important, followed by lung, heart. The liver don’t appear in the results of data mining system. The results indicates that morbility of myasthenia gravis is most closely with spleen, kidney and lung.2)Results of SQL Server software by using Naive Bayes algorithms about important influencing factor of various clinical type of myasthenia gravis:①Ⅰtype primary factors:age<19 years old, not appearing general fatigue; secondary factors:not appearing weakness of chewing and swallowing, lassitude.②Ⅱ-A type primary factors:not appearing weakness of chewing and swallowing; secondary factors:not appearing salivation, respiratory weakness, unclear articulation.③Ⅱ-B type primary factors:appearing general fatigue; secondary factors: appearing weakness of chewing and swallowing, unclear articulation.④Ⅲtype primary factors:appearing respiratory weakness; secondary factors: appearing weakness of chewing and swallowing, dyspnea with rapid and short breath, low speech, weak knee, complicating with pulmonary infection.⑤Ⅳtype primary factors:complicating with myasthenic crisis, appearing respiratory weakness; secondary factors:appearing weakness of chewing and swallowing, stethocatharsis, salivation, unclear articulation, complicating with thymoma①In addition to spleen, involved other viscera about various clinical types of myasthenia gravis as follows:Ⅰtype primary factors:not involving renal; secondary factors:not involving lung and heart.Ⅱ-A type primary factors:not involving lung; secondary factors:not involving renal.Ⅱ-B type primary factors:involving renal; secondary factors:involving lung and heart.Ⅲtype primary factors:involving lung; secondary factors:involving heart and lung.Ⅳtype primary factors:involving lung; secondary factors:involving heart and lung.3. Conclusion:(1) Myasthenia gravis is involved with spleen mainly, related with kidney, lungs closely, and also related with heart and liver secondly. This is consistent with pathogenesis "spleen deficiency and five-viscera correlation" which is summed up by Professor Deng Tietao.(2)Weakness of chewing swallowing, general fatigue, respiratory weakness, unclear articulation are key symptoms to identify the clinical type of myasthenia gravis. Weakness of chewing swallowing is the most important symptom for determining disease severity, comparing with other symptoms. It is not onlyⅡ-A type andⅡ-B type differentiation point but also the clinical indication of using corticosteroids such as prednisone.(3)Ⅰtype damages spleen, liver mainly;Ⅱ-A type damages spleen, kidney mainly and liver;Ⅱ-B type damages spleen, kidney, lung mainly,Ⅲtype damages spleen, kidney, lung mainly and heart;Ⅳtype damages spleen, kidney, lung mainly and heart. (4) During the period of myasthenia gravis, spleen is diseased all along. With aggravation of pathogenetic condition, two viscera diseased simultaneously happen, then three viscera diseased simultaneously, four viscera diseased simultaneously, five viscera diseased simultaneously. Five-viscera correlation Pathogenesis mode of myasthenia gravis include: spleen and liver diseased simultaneously, spleen and kidney diseased simultaneously, liver and spleen and kidney diseased simultaneously, spleen and kidney and lung diseased simultaneously, spleen and kidney and lung and heart diseased simultaneously, five viscera diseased simultaneously.(5) "Spleen deficiency and five-viscera correlation " theory of professor Deng Tietao is not only significant in guiding diagnosis and treatment of myasthenia gravis, but also in other critical illness diagnosis and treatment of difficult diseases.(6)This study proves the correctness of professor Deng Tietao’s theory, by starting from objective clinical information, the introduction of data mining technology, concluding the TCM Zang-Fu pathogenesis of myasthenia gravis. It also indicates that data mining technology is effective in studying TCM five-viscera correlation theory.
【Key words】 Five-viscera correlation theory of TCM; Myasthenia gravis; Data mining; Zang-fu pathogenesis;