Qi Xiang-Jun, Chen Xin-Rong, Mo Jia-Hao, Li Pei-Xin, Cai Meng-Yi, Lan Wan-Ning, Chen Han-Rui, Chen Zhuang-Zhong, Chen Guo-Ming, Lin Li-Zhu
the First Clinical College of Guangzhou University of Chinese Medicine Guangzhou 510405, China.
the Second Clinical College of Guangzhou University of Chinese Medicine Guangzhou 510405, China.
Zhongguo Zhong Yao Za Zhi. 2021 Aug;46(15):4016-4022. doi: 10.19540/j.cnki.cjcmm.20201127.501.
The tumor prescriptions contained in Dictionary of Tumor Formulas, Compendium of Good Tumor Formulas, Chinese Pharmacopoeia, Ministry of Health Drug Standards for Chinese Medicine Formulas and National Compilation of Standards for Proprietary Chinese Medicines were selected and organized to construct a database for tumor prescriptions, and the data mining techniques were applied to investigate the prescription regularity of colorectal cancer prescriptions. The formula data were extracted after screening in strict accordance with the inclusion and exclusion criteria, and were then analyzed with Microsoft Excel 2010 for frequency statistics, Apriori block provided by SPSS Clementine 12.0 software for correlation rule analysis, and arules and arulesViz packages in R 4.0.2 software for correlation rule visualization. In addition, SPSS 18.0 software was used for cluster analysis and factor analysis, in which cluster analysis was performed by Ochiai algorithm with bicategorical variables in systematic clustering method and factor analysis was performed mainly with principal component analysis. A total of 285 prescriptions were included in the statistical analysis, and the frequency statistics showed that 43 herbs had been used more than 16 times. The association rules analysis showed that 26 high-frequency me-dicine pair rules were obtained, and the association rules for those dispelling evil spirits, strengthening the body, resolving stasis, dispelling dampness, etc. were visualized. In the cluster analysis, we generated a dendrogram from which 7 groups of traditional Chinese medicines with homogeneity were extracted. 10 common factors were obtained in the factor analysis. The types of herbal medicines involved in the colorectal cancer prescription included anti-cancer antidotes, strengthening and tonifying medicines, blood-regulating medicines, and expectorant medicines, corresponding to the treatment for eliminating evil spirits, strengthening, resolving stasis, and expectorating dampness. The prescriptions for anti-cancer detoxification were normally based on the pairs composed of Scutellaria barbata-Hedyotis diffusa and Sophora flavescens, Sargentodoxa cuneata, S. barbata, often combined with stasis relieving drug and dampness eliminating drug, reflecting the characteristics of treatment for both toxicity and stasis, dampness and toxicity simultaneously. The prescriptions for strengthening the righteousness and tonifying the deficiency were composed of Astragalus membranaceus and Atractylodes macrocephala mainly, exerting the effect of benefiting Qi, strengthening the spleen and drying dampness, tonifying kidney and essence, tonifying blood and invigorating blood. Meanwhile, anti-cancer detoxification medicines shall be reduced as much as possible. The compatibility of the medicines for the intestinal tract reflected the principle of using the right medicine for the right condition and eliminating evil spirits or strengthening the body, as appropriate.
选取《肿瘤方剂大辞典》《肿瘤良方大全》《中国药典》《卫生部中药成方制剂标准》及《国家中成药标准汇编》中收录的肿瘤方剂,构建肿瘤方剂数据库,并运用数据挖掘技术探讨结直肠癌方剂的用药规律。方剂数据严格按照纳入与排除标准筛选提取后,采用Microsoft Excel 2010进行频数统计,运用SPSS Clementine 12.0软件的Apriori模块进行关联规则分析,利用R 4.0.2软件的arules和arulesViz包进行关联规则可视化。此外,使用SPSS 18.0软件进行聚类分析和因子分析,其中聚类分析采用系统聚类法中的Ochiai算法对二分类变量进行分析,因子分析主要采用主成分分析法。共有285首方剂纳入统计分析,频数统计显示有43味药使用频次超过16次。关联规则分析得到26条高频药对规则,并对祛邪、扶正、化瘀、祛湿等关联规则进行了可视化展示。聚类分析生成了聚类图,从中提取出7组具有同质性的中药。因子分析得到10个公因子。结直肠癌方剂涉及的中药类型有抗癌解毒药、补益药、活血药、化痰药,分别对应祛邪、扶正、化瘀、化痰的治法。抗癌解毒方剂多以半枝莲 - 白花蛇舌草、苦参、败酱草、半枝莲药对为基础,常配伍化瘀药和祛湿药,体现了攻毒兼化瘀、祛湿的用药特点;扶正补虚方剂以黄芪、白术为主,发挥益气健脾燥湿、补肾填精、养血活血的功效,同时尽量减少抗癌解毒药的使用。肠道用药的配伍体现了因病制宜、祛邪或扶正的原则。