Prediction of Lung Cancer Stage Using Tumor Gene Expression Data

Lung cancer remains a significant global health challenge and identifying lung cancer at an early stage is essential for enhancing patient outcomes. The study focuses on developing and optimizing gene expression-based models for classifying cancer types using machine learning techniques. By applying Log2 normalization to gene expression data and conducting Wilcoxon rank sum tests, the researchers employed various classifiers and Incremental Feature Selection (IFS) strategies. The study culminated in two optimized models using the XGBoost classifier, comprising 10 and 74 genes respectively. The 10-gene model, due to its simplicity, is proposed for easier clinical implementation, whereas the 74-gene model exhibited superior performance in terms of Specificity, AUC (Area Under the Curve), and Precision. These models were evaluated based on their sensitivity, AUC, and specificity, aiming to achieve high sensitivity and AUC while maintaining reasonable specificity.

Lung-RADS与CT联合诊断恶性SPN的临床意义

目的:探究肺结节肺部影像报告与数据系统(Lung-RADS)分类联合电子计算机断层扫描(CT)诊断恶性孤立性肺结节(SPN)的价值。方法:选择2020年6月—2022年6月萍乡市人民医院收治的120例SPN患者的临床资料,经病理或手术明确24例良性SPN、96例恶性SPN;回顾性分析其肺结节Lung-RADS分类及CT影像检查结果,采用受试者操作特征(ROC)曲线分析Lung-RADS分类、CT征象评分单独及联合诊断恶性SPN的效能。结果:120例患者Lung-RADS分类结果显示,2类16例,3类20例,4A类10例,4B类54例,4C类20例。良性SPN组病灶边界光整、钙化占比均高于恶性SPV组,支气管截断、胸膜牵拉、分叶、肿瘤血管征象占比均低于恶性SPN组,差异均有统计学意义(P<0.05)。良性SPN组CT征象评分2分患者占比高于恶性SPN组,恶性SPN组CT征象4、5、6分患者占比均高于良性SPN组,差异均有统计学意义(P<0.05)。ROC曲线分析显示,Lung-RADS分类、CT征象评分联合诊断恶性SPN的AUC值为0.948,高于二者单独诊断的0.881、0.931,联合诊断恶性SPN的敏感度为97.92%,高于lung-RADS分类、CT征象评分单独诊断的83.33%、89.58%。结论:Lung-RADS分类与CT征象联合诊断恶性SPN较二者单独诊断具备更高诊断效能,可提升恶性SPN诊断敏感度,具有良好的临床应用价值。

The diagnostic rules of peripheral lung cancer preliminary study based on data mining technique

Objective： To discuss the clinical and imaging diagnostic rules of peripheral lung cancer by data mining technique, and to explore new ideas in the diagnosis of peripheral lung cancer, and to obtain early-stage technology and knowledge support of computer-aided detecting （CAD）. Methods： 58 cases of peripheral lung cancer confirmed by clinical pathology were collected. The data were imported into the database after the standardization of the clinical and CT findings attributes were identified. The data was studied comparatively based on Association Rules （AR） of the knowledge discovery process and the Rough Set （RS） reduction algorithm and Genetic Algorithm（GA） of the generic data analysis tool （ROSETTA）, respectively. Results： The genetic classification algorithm of ROSETTA generates 5 000 or so diagnosis rules. The RS reduction algorithm of Johnson＇s Algorithm generates 51 diagnosis rules and the AR algorithm generates 123 diagnosis rules. Three data mining methods basically consider gender, age, cough, location, lobulation sign, shape, ground-glass density attributes as the main basis for the diagnosis of peripheral lung cancer. Conclusion： These diagnosis rules for peripheral lung cancer with three data mining technology is same as clinical diagnostic rules, and these rules also can be used to build the knowledge base of expert system. This study demonstrated the potential values of data mining technology in clinical imaging diagnosis and differential diagnosis.

Survival and comorbidities in lung cancer patients:Evidence from administrative claims data in Germany

Lung cancer is the most common cancer type worldwide and has the highest and second highest mortality rate for men and women respectively in Germany.Yet,the role of comorbid illnesses in lung cancer patient prognosis is still debated.We analyzed administrative claims data from one of the largest statutory health insurance(SHI)funds in Germany,covering close to 9 million people(11%of the national population);observation period was from 2005 to 2019.Lung cancer patients and their concomitant diseases were identified by ICD-10-GM codes.Comorbidities were classified according to the Charlson Comorbidity Index(CCI).Incidence,comorbidity prevalence and survival are estimated considering sex,age at diagnosis,and place of residence.Kaplan Meier curves with 95%confidence intervals were built in relation to common comorbidities.We identified 70,698 lung cancer incident cases in the sample.Incidence and survival figures are comparable to official statistics in Germany.Most prevalent comorbidities are chronic obstructive pulmonary disease(COPD)(36.7%),followed by peripheral vascular disease(PVD)(18.7%),diabetes without chronic complications(17.4%),congestive heart failure(CHF)(16.5%)and renal disease(14.7%).Relative to overall survival,lung cancer patients with CHF,cerebrovascular disease(CEVD)and renal disease are associated with largest drops in survival probabilities(9%or higher),while those with PVD and diabetes without chronic complications with moderate drops(7%or lower).The study showed a negative association between survival and most common comorbidities among lung cancer patients,based on a large sample for Germany.Further research needs to explore the individual effect of comorbidities disentangled from that of other patient characteristics such as cancer stage and histology.

Analyze the rules of Chinese herbal of Professor Xiong Lu in treating metaphase and advanced lung cancer by data mining

Objective： To analyze the experience of chief physician Xiong Lu in treating metaphase and advanced lung cancer through using TCM inheritance support system （V2.5）. Methods： Collecting the prescriptions used for metaphase and advanced lung cancer from November 1, 2014 to February 1, 2015, then the data were entered into the TCM inheritance support system. Based on principle analysis, revised mutual information, complex system entropy cluster and unsupervised hierarchical clustering composing principles were analyzed. Results： Based on the analysis of 228 cases of prescriptions, the frequency of each Chinese medicinal herb and association rules among herbs included in the database were computed. 15 core combinations and 2 new prescriptions were explored from the database. Conclusion： In treating metaphase and advanced lung cancer, chief physician Xiong Lu pay attention to Fuzheng Peiben （Therapy for support Zheng-qi to propup root）, according to the different situation cooperate with Tong Luo （dredging collaterals）, San Jie （Dissipating a mass）, Huo Xue （Activating blood）, Gong Du （Counteracting toxic substance） and so on. Xiong Lu is also good at using toxic drugs and incompatible medicaments.

End-to-End 2D Convolutional Neural Network Architecture for Lung Nodule Identification and Abnormal Detection in Cloud

The extent of the peril associated with cancer can be perceivedfrom the lack of treatment, ineffective early diagnosis techniques, and mostimportantly its fatality rate. Globally, cancer is the second leading cause ofdeath and among over a hundred types of cancer;lung cancer is the secondmost common type of cancer as well as the leading cause of cancer-relateddeaths. Anyhow, an accurate lung cancer diagnosis in a timely manner canelevate the likelihood of survival by a noticeable margin and medical imagingis a prevalent manner of cancer diagnosis since it is easily accessible to peoplearound the globe. Nonetheless, this is not eminently efficacious consideringhuman inspection of medical images can yield a high false positive rate. Ineffectiveand inefficient diagnosis is a crucial reason for such a high mortalityrate for this malady. However, the conspicuous advancements in deep learningand artificial intelligence have stimulated the development of exceedinglyprecise diagnosis systems. The development and performance of these systemsrely prominently on the data that is used to train these systems. A standardproblem witnessed in publicly available medical image datasets is the severeimbalance of data between different classes. This grave imbalance of data canmake a deep learning model biased towards the dominant class and unableto generalize. This study aims to present an end-to-end convolutional neuralnetwork that can accurately differentiate lung nodules from non-nodules andreduce the false positive rate to a bare minimum. To tackle the problem ofdata imbalance, we oversampled the data by transforming available images inthe minority class. The average false positive rate in the proposed method isa mere 1.5 percent. However, the average false negative rate is 31.76 percent.The proposed neural network has 68.66 percent sensitivity and 98.42 percentspecificity.

Identification of key genes and biological pathways in lung adenocarcinoma by integrated bioinformatics analysis

BACKGROUND The objectives of this study were to identify hub genes and biological pathways involved in lung adenocarcinoma(LUAD)via bioinformatics analysis,and investigate potential therapeutic targets.AIM To determine reliable prognostic biomarkers for early diagnosis and treatment of LUAD.METHODS To identify potential therapeutic targets for LUAD,two microarray datasets derived from the Gene Expression Omnibus(GEO)database were analyzed,GSE3116959 and GSE118370.Differentially expressed genes(DEGs)in LUAD and normal tissues were identified using the GEO2R tool.The Hiplot database was then used to generate a volcanic map of the DEGs.Weighted gene co-expression network analysis was conducted to cluster the genes in GSE116959 and GSE-118370 into different modules,and identify immune genes shared between them.A protein-protein interaction network was established using the Search Tool for the Retrieval of Interacting Genes database,then the CytoNCA and CytoHubba components of Cytoscape software were used to visualize the genes.Hub genes with high scores and co-expression were identified,and the Database for Annotation,Visualization and Integrated Discovery was used to perform enrichment analysis of these genes.The diagnostic and prognostic values of the hub genes were calculated using receiver operating characteristic curves and Kaplan-Meier survival analysis,and gene-set enrichment analysis was conducted.The University of Alabama at Birmingham Cancer data analysis portal was used to analyze relationships between the hub genes and normal specimens,as well as their expression during tumor progression.Lastly,validation of protein expression was conducted on the identified hub genes via the Human Protein Atlas database.RESULTS Three hub genes with high connectivity were identified;cellular retinoic acid binding protein 2(CRABP2),matrix metallopeptidase 12(MMP12),and DNA topoisomerase II alpha(TOP2A).High expression of these genes was associated with a poor LUAD prognosis,and the genes exhibited high diagnostic value.CONCLUSION Expression levels of CRABP2,MMP12,and TOP2A in LUAD were higher than those in normal lung tissue.This observation has diagnostic value,and is linked to poor LUAD prognosis.These genes may be biomarkers and therapeutic targets in LUAD,but further research is warranted to investigate their usefulness in these respects.

肺实性结节半自动体积测量提高观察者间Lung-RADS评分一致性

目的:研究3D-Slicer软件实性肺结节体积测量对不同观察者在肺部影像报告数据系统(lung CT screening reporting and data system,Lung-RADS)分类一致性中的影响。方法:纳入76例患者中的76个实性结节。由3位放射科医师分别采用手动和3D-Slicer软件半自动体积测量方法获得结节的直径与体积,并转化为相应的Lung-RADS评分,其中2分为阴性,3分及以上为阳性。采用同类相关系数(intraclass correlation coefficient,ICC)及Bland-Altman指数来评价观察者间直径与体积测量的一致性,Kappa分析评价观察者间Lung-RADS评分及阳性/阴性组间的一致性。结果:ICC分析结果显示手动直径测量的一致性(0.994~0.996)明显低于半自动体积测量的一致性(0.997~0.998),同时Bland-Altman指数分析结果显示手动直径测量的偏倚高于半自动体积测量。采用半自动体积测量,能够比手动直径测量明显提高观察者间Lung-RADS评分及阳性/阴性之间的一致性(0.963~0.975及0.957~0.977 vs.0.833~0.866及0.863~0.892)。结论:3D-Slicer半自动体积测量能够提高实性肺结节观察者间大小测量的一致性,相应的Lung-RADS分类一致性也随之提高。

肠内营养联合加速康复用于非小细胞肺癌手术患者的效果

目的分析基于数据包络分析(DEA)模型评估肠内营养联合加速康复用于非小细胞肺癌手术患者的效果。方法选取2020年10月至2023年10月期间在中山市小榄人民医院治疗的102例非小细胞肺癌手术患者,根据随机数字表法分为两组,对照组51例采用常规干预指导,试验组51例采用肠内营养联合加速康复,比较两组干预效果后,再经DEA模型预测,比较两组技术效益、规模效益、综合效益、规模报酬系数、有效性和规模报酬类型。结果试验组卧床时间、胸管留置时间、抗生素使用时间、住院时间短于对照组(P<0.05),总治疗费用少于对照组(P<0.05),术后7 d的功能状况、情感状况、家庭状况、生理状况、肺癌附加情况评分高于对照组(P<0.05),术后7 d的第1秒用力呼气容积(FEV_1)、用力肺活量(FVC)、每分钟最大通气量(MVV)、血清白蛋白(ALB)、总蛋白(TP)、血红蛋白(Hb)改善情况优于对照组(P<0.05),术后肺部并发症发生率(3.92%比17.65%)低于对照组(χ^(2)=4.993,P=0.025)。同时,建立DEA模型预测,试验组营养状况指标的技术效益值为1的决策单元有32个,综合效益28个,以DEA强有效、规模报酬固定为主。两组营养状况指标的技术效益、综合效益存在差异,肺功能指标的技术效益、综合效益、规模报酬系数存在差异(P<0.05)。结论基于DEA模型评估分析,肠内营养联合加速康复能够提高规模效率和综合效益,利用价值较高,将其用于非小细胞肺癌手术患者中效果显著,可改善术后肺功能和营养状态,降低术后肺部并发症发生率,促进术后康复。

基于数据挖掘及免疫组织化学技术分析EPHA5在肺腺癌中的表达

目的探索促红细胞生成素产生型人肝细胞受体A5(erythropoietin-producing hepatocellular receptor A5,EPHA5)在肺腺癌中的表达。方法通过网络数据库获得EPHA5在不同肿瘤中的变异情况,构建EPHA5蛋白互作网络图,分析EPHA5在肺腺癌中的预后情况。收集肺腺癌患者的手术标本及临床资料,采用免疫组织化学方法检测EPHA5的表达情况,分析肺腺癌中EPHA5表达与患者临床特征的相关性。结果EPHA5在肺腺癌中的突变频率最高,以错义突变为主。除EPH家族的配体之外,EPHA5与RhoA、ADAM10蛋白密切相关。EPHA5的高表达与肺腺癌患者的总生存时间和首次进展时间相关。EPHA5在肺腺癌组织中表达上调,与患者淋巴结转移、分化程度、TNM分期显著相关。结论EPHA5在肺腺癌组织中高表达,可以作为预测肺腺癌预后的一个因素。

肺腺癌中HPRT1基因表达对患者总生存的影响

目的 探索HPRT1基因在肺腺癌中的表达特征、总生存率、功能激活以及免疫浸润中的影响。方法 通过对TCGA肺腺癌数据以及多个GEO数据库中的肺腺癌数据进行挖掘分析,对比并验证HPRT1表达量与预后总生存(overall surival,OS)的关系,通过cluster Profiler分析HPRT1基因高表达组中上调基因的功能富集情况。且通过TIMER以及CIBERSORT算法计算肺腺癌中不同免疫细胞的浸润水平并对比在HPRT1高低表达组间的浸润程度差异。结果 HPRT1在肺腺癌组织中表达量显著上调,且在TCGA中证明HPRT1基因高表达的患者预后OS更差(P <0.01)。经过2个GEO数据集的验证同样发现HPRT1基因高表达表现为预后OS更差(GSE13213,P <0.01;GSE67639,P <0.001)。差异分析显示高表达的患者中有683个基因的表达量显著上调且这些上调基因的功能主要富集与p53以及细胞周期等癌症相关的信号通路。TIMER以及CIBERSORT算法进行的免疫细胞浸润程度分析发现在高表达差预后人群中B细胞以及CD4T细胞含量均更低(P <0.05)。结论 HPRT1基因表达量越高肺腺癌患者的总生存越差,且高表达患者的p53信号通路上调,B细胞以及CD4T细胞浸润程度显著下降。

基于数据挖掘的针灸治疗肺系疫病古代文献整理研究

目的收集古文献中针灸治疗肺系疫病的有关条文,分析针灸治疗肺系疫病的应用特点,为现代提供临床借鉴。方法从国家图书馆古籍文献纸质书籍、国家图书馆电子文献数据库电子书籍、上海中医药大学图书馆纸质书籍、电子书籍中收集秦至清末时期内针灸治疗肺系疫病的古籍文献条文,按照纳入标准与排除标准进行筛选,运用IBM SPSS Statistics 27.0进行频数分析、聚类分析、主成分与因子分析及关联规则分析。结果共纳入226条针灸治疗肺系疫病的文献条文,其中来源于《证治准绳》条文最多,为32条,针灸疗法包括12大类,使用频次最高的为刺血56次,针灸治疗涉及68个穴位,使用频次超过5次的穴位有2个,依次为少商、委中。结论在针灸治疗肺系疫病时可采用针刺与刺血等多种方式相结合,可用“间使、肩井、曲池、膏肓”为治疗核心处方,并辅以辨证取穴。

基于大数据平台的肺部结节随访系统优化探索

Hadoop是公认的行业大数据标准开源软件,因其在分布式环境下具备海量数据处理能力,目前在肺部结节随访系统中应用广泛。然而,Hadoop分布式文件系统(HDFS)在设计之初是为了解决大文件存储与计算问题,对海量数目的小文件存储与检索存在性能低下、主节点NameNode内存占用率高等问题。为此构建一种改进的HDFS数据布局存储方案HFS,通过在NameNode中加入文件处理识别模块实现小文件元数据向SecondnameNode和DataNode集群的迁移;同时设计出DataNode间数据流动的算法,有效降低了NameNode节点的处理压力。分别基于HFS和单一HDFS对肺部结节随访系统进行测试,实验结果表明在NameNode内存占有率和整体数据分析时间等方面,基于HFS的肺部结节随访系统具备明显优势。

基于真实世界数据挖掘中医药治疗不同证型系统性红斑狼疮肺间质病变的用药规律研究

目的 基于真实世界临床数据探讨不同证型系统性红斑狼疮肺间质病变中药组方用药规律。方法 收集2017年1月至2021年12月门诊及住院患者资料,对患者的基本信息、中医证型、药物四气五味归经进行频次频率统计;对前4高频证型的药物进行关联规则分析、聚类分析及证型-药物复杂网络分析。结果 共纳入146例患者、369首处方。经频次统计发现最常见证型是湿热痹阻证、气阴两虚证、肝肾亏虚证、阴虚内热证;共涉及298味中药,出现药物频次5 629次,高频药物有29味,排名前5位分别是茯苓、当归、黄芪、炙甘草、白芍;药物四气以温、寒、平为主,五味以甘、苦、辛为主,归经以肺、脾、肝、肾为主,功效分类以补虚药、清热药、利水渗湿药为主;高频证型关联规则、系统聚类及复杂网络分析得到湿热痹阻证、气阴两虚证、肝肾亏虚证、阴虚内热证核心药物组合分别为四妙丸加减与木防己汤加减、补肺汤加减、独活寄生汤加减、六味地黄丸加减。结论 中药治疗系统性红斑狼疮肺间质病变多以清热祛湿、活血通痹、益气养阴、补肺脾肾为治则,得出的组方规律对临床治疗具有一定的参考价值,同时为新药开发提供思路。

从信号到知识——基于人工智能的医学影像裸数据诊断价值初探

Encouraging and astonishing developments have recently been achieved in image-based diagnostic technology.Modern medical care and imaging technology are becoming increasingly inseparable.However,the current diagnosis pattern of signal to image to knowledge inevitably leads to information distortion and noise introduction in the procedure of image reconstruction(from signal to image).Artificial intelligence(AI)technologies that can mine knowledge from vast amounts of data offer opportunities to disrupt established workflows.In this prospective study,for the first time,we develop an AI-based signal-toknowledge diagnostic scheme for lung nodule classification directly from the computed tomography(CT)raw data(the signal).We find that the raw data achieves almost comparable performance with CT,indicating that it is possible to diagnose diseases without reconstructing images.Moreover,the incorporation of raw data through three common convolutional network structures greatly improves the performance of the CT models in all cohorts(with a gain ranging from 0.01 to 0.12),demonstrating that raw data contains diagnostic information that CT does not possess.Our results break new ground and demonstrate the potential for direct signal-to-knowledge domain analysis.

肺部听诊音数据库建库技术及方法研究

当前无论是物理听诊器亦或是电子听诊器的肺音听诊结果仍然主要依靠医生专业的听诊鉴别能力,尚未能够实现智能诊断判读。当患者在家受到肺部疾病影响时,无法自行发现肺部异常而耽误治疗;当处于呼吸道传染病救治过程中,入耳式的听诊器容易受到污染而造成院内感染。尽管听诊音包含了丰富的健康状态信息,由于缺乏标准化的采集方法、分类标准和分析工具,使得听诊音的客观分析和应用在实践中受到了限制。本研究通过采用统一的听诊音采集设备和流程进行肺部听诊音数据采集、整理、数据库设计,使用软件MatlabR2017a进行数据管理和分析,建立了健康群体和肺部疾病患者群体的肺部听诊音数据库,制订一套标准的听诊音分类、标注规范、音频特征信号参数,构建一个用于存储、管理和分析肺部听诊音数据的系统,为肺部疾病的筛查、监测以及医学人工智能应用转化等相关研究提供重要的数据支持。本研究为肺部听诊音音频数据库建库积累了经验,为音频类数据库管理和分析提供有益的参考和借鉴,为支持后续医学人工智能辅助听诊应用于肺部疾病筛查与监测奠定基础,具有重要的医学价值和实际应用意义。

基于数据挖掘技术分析含薏苡仁方治疗肺癌用药规律

目的:通过数据挖掘技术分析含薏苡仁方治疗肺癌的用药规律。方法:检索、收集中国知网、万方数据知识服务平台、维普网及中国生物医学文献数据库中治疗肺癌的含薏苡仁方的临床研究文献,筛选后建立数据库,统计方中的薏苡仁用量以及高频药物性、味、归经的频次、使用频率等,并对数据进行关联规则分析及聚类分析。结果:共筛选出19个文献,22首中药复方,涉及中药99味。药物的主要药性为寒、温、平,主要药味为甘、苦、辛,主要归经为肺、肝、脾,主要药物类别为清热药、补虚药、化痰止咳药、利水渗湿药、理气药、活血化瘀药。高频药物有黄芪、甘草、白术、半夏、茯苓等。薏苡仁用量范围为10~30 g。通过关联规则分析得出14对药对。通过聚类分析生成1个新组方。结论:在肺癌的治疗中,薏苡仁配伍的总原则是扶正祛邪,攻补兼施,主要配伍甘、苦、辛味药物,主要治法有补益脾肺、健脾祛湿、活血化瘀、燥湿化痰等。薏苡仁的用量范围为10~30 g,用量可随着肺癌病情的加重而增加。

中医药治疗肺癌的用药规律

目的:探究中医药治疗肺癌的用药规律。方法:检索《肿瘤良方大全》《肿瘤方剂大辞典》《卫生部药品标准中药成方制剂》以及《国家药品监督管理局总局国家药品标准(修订)颁布件》中中医治疗肺癌的处方,将数据录入Excel进行数据处理,使用Lantern 5.0、SPSS Modeler 18.0及SPSS Statistics 25.0对肺癌方剂频次统计、系统聚类分析及关联规则分析。结果:共纳入650个处方,572种药,6 120次。药性以寒、温、平为主,药味以苦、甘、辛为主,归经使用最多为胃、脾、肾、肝、肺。使用频次最多的药物为黄芪、白花蛇舌草、甘草、茯苓、沙参、麦冬等。高频药物以清热解毒药和益气养阴药及活血化瘀药主;关联规则分析得到药对关联16条,三联24条,因子内得到13个公因子,系统聚类分析得出7个关联紧密的药组。隐结构得到14个隐变量,每个隐变量有2个隐类,共28个隐类。结论:补气养阴药和清热抗癌解毒、活血化瘀药、健脾利湿药、清热化痰治则,为中医药治疗肺癌提供借鉴,有待进一步结合临床及实验给予验证。

基于文献分析探讨数据挖掘在肺癌中药用药规律的应用

目的:分析数据挖掘在肺癌中药用药规律中的应用,进一步总结中医药治疗肺癌的规律,为中医药治疗肺癌提供新的思路。方法:检索中国知网、万方数据、维普网、Pubmed中采用数据挖掘技术探究中药治疗肺癌用药规律的文献,构建数据库,应用文献分析对文献的数据分析软件、算法、高频中药、药对等信息进行归纳总结。结果:得到88篇有效文献,文献数量呈逐年上升趋势;数据挖掘常用分析软件为MicrosoftExcel、SPSS及其系列软件、中医传承辅助平台;常用分析方法为频次分析、关联规则和聚类分析;97.73%文献统计了高频药物,合计得到151味高频药物,频数排名前三的为黄芪、白花蛇舌草和甘草;涉及的药类主要为清热药和补虚药;高频药物药性以寒性、温性最多,药味主要为甘味、苦味和辛味,归经主要归肝、肺、脾和胃经;80.90%文献分析了治疗肺癌的常用药对,共计得到239组常用药对,对高频药物的关联规则分析得到二项关联5条,两者结果呼应;37.65%文献总结了中医治疗肺癌的常用中药核心组合,57.95%文献总结了常用聚类方或新方,23.86%文献总结了不同中医证型肺癌的用药规律;聚类分析得到5个药物组合。结论:数据挖掘已广泛运用于中医药治疗肺癌的领域,且研究结果符合肺癌“虚、痰、瘀、毒”的病机,间接说明了中医治疗肺癌应以扶正为主,配合活血、化痰、祛湿、解毒等法,为中医药治疗肺癌提供依据。

基于真实世界数据分析中医药非介入治疗肺恶性肿瘤的辨治规律

目的:基于临床病历信息系统,采用数据挖掘方法从真实世界角度分析中医药非介入治疗肺恶性肿瘤的辨治规律。方法:收集2015年1月1日至2021年12月31日北京中医药大学东直门医院病历系统中诊断为肺恶性肿瘤的患者病史、刻下症和用药信息,在剔除使用介入治疗手段的病例后建立数据库。采用Microsoft Office Excel 2019对高频药物的频数、四气五味、归经及功效进行描述性统计,并对病史和刻下症反映的症状进行数据化统计;对统计得到的高频药物和症状信息借助SPSS Modeler18.0软件中Apriori算法对核心数据进行关联规则分析,并用Web节点构建关联网状图;借助SPSS Stastistics 23.0进行聚类分析。结果:共纳入119例未进行介入治疗的肺恶性肿瘤患者,中医常见症状为咳嗽、白痰、喘憋、纳差、眠差、痰黏、便秘、乏力、胸闷、气短;药物四气以温药最多;五味以甘味最多;归经以肺经频率最高。关联规则结果显示,二项支持度最高的组合为半夏→柴胡、半夏→瓜蒌、杏仁→麻黄;三项关联支持度最高的组合为半夏→茯苓、柴胡,杏仁→麻黄、半夏,半夏→杏仁、柴胡,半夏→瓜蒌、贝母。聚类结果聚为3类,依次为肺阴虚证、痰热壅肺证和肺气虚证。结论:中医药非介入治疗肺恶性肿瘤的中医药辨治规律为:聚类证型依次为肺阴虚证、痰热壅肺证和肺气虚证。治疗多用和解少阳、行气止咳平喘和清虚热化痰饮为主,药物以半夏、甘草、黄芩、茯苓、柴胡多见。