The Study Of Ontology Construction And Data Mining Of Professor ZHOU Zhong-ying’s Lung Cancer Treatment Knowledge

This paper mainly includes two parts:the first part:research on ontology construction of Professor ZHOU Zhong-ying’ s lung cancer treatment knowledge;the second part:research on ZHOU Zhong-ying’ s law of differentiating and treating lung cancer based on dedical records and data mining.Part 1:research on ontology construction of Professor ZHOU Zhong-ying’ s lung cancer treatment knowledgeObjectiveIn this study,by accurately and comprehensively constructing Professor ZHOU Zhong-ying’s knowledge ontology of clinical experience in the treatment of lung cancer,the systematic and structured display and retrieval of knowledge was realized,which can help beginners learn and master Professor ZHOU Zhong-ying’s clinical experience in the treatment of lung cancer quickly,accurately and comprehensively,and provide evidence for the results of subsequent data mining.MethodsWith the theme of "ZHOU Zhong-ying" and "lung cancer",the relevant literatures of Professor ZHOU Zhong-ying treating lung cancer were systematically retrieved in CNKI database,Wanfang database and Vip database,and time limit was set from database establishment to August 31,2018.The original information related to etiology,location,nature,pathogenesis,pathological factors,positive and evil relationships,symptoms,treatment principles and methods,syndrome differentiation,medication and metastasis of lung cancer was extracted from eligible literatures.Based on the standard of "Nouns of Traditional Chinese Medicine",the original information is standardized and extracted as concepts,attributes and logical relations.Protege 5.5 was used as ontology modeling tool and OWL as ontology description language,Professor ZHOU Zhong-ying’s ontology relationship model in the field of lung cancer treatment,namely concept tree,was established,and then Professor ZHOU Zhong-ying’s knowledge ontology of clinical experience in the treatment of lung cancer was constructed.ResultsA total of 28 eligible literatures were retrieved and screened from the electronic database.439 specific concepts,31 attributes and 718 logical relationships among concepts were extracted.By combing the relationship between concepts,linking the concepts of traditional Chinese medicine into a semantic network with concepts as nodes and logical relationships as paths,a knowledge ontology of traditional Chinese medicine was constructed,including Professor ZHOU Zhong-ying’s understanding of diseases in the treatment of lung cancer and medication based on differentiations.Through the relevant plug-ins of the software,Professor ZHOU Zhong-ying’s knowledge of lung cancer treatment can be systematically,structurally displayed and intelligently retrieved.The knowledge ontology is open and editable after the completion of the construction,which can realize real-time updating and perfection.ConclusionThe knowledge ontology construction technology was applied to the research of Professor ZHOU Zhong-ying’s experience on lung cancer treatment,① which realized the reconstruction and sharing of clinical experience knowledge,and made it possible to retrieve and display clinical experience knowledge system based on semantics.②This not only can discover,manage and utilize relevant knowledge more scientifically and effectively,improve its transmission efficiency and understanding depth,③ but also is conducive to comprehensive retrieval of Professor ZHOU Zhong-ying’s experience knowledge of lung cancer treatment,database construction and correct interpretation of relevant concepts,and can verify with subsequent data mining results.④ This study has practical reference significance for the display and optimization of the summary of clinical experience and application of clinical learning of famous and veteran TCM masters.It also explored and innovated the inheritance methods of clinical experience and academic ideas of famous and veteran TCM masters.Part 2:research on ZHOU Zhong-ying’s law of differentiating and treating lung cancer based on dedical records and data miningObjectiveTo summarize Professor ZHOU Zhong-ying’ s clinical thoughts and medication rules of syndrome differentiation by data mining technology,and to verify and supplement data mining results with knowledge ontology.MethodsA database of medical records of TCM lung cancer cases treated by Professor ZHOU Zhong-ying was constructd by collecting and sorting out a large number of relevant medical cases from 1994 to 2010.With the help of Excel 2016、IBM SPSS Modeler 18.0 and IBM SPSS Statistics 17.0,using the methods of frequency statistics,association rules,scale-free network analysis,cluster analysis and factor analysis,rules of Professor ZHOU Zhong-ying’ s treatment of lung cancer cases,such as syndrome differentiation,medication,symptoms,surgery,chemotherapy and cancer metastasis were summarized by statistics,analysis and data mining on the collected medical records.The TCM diagnosis and treatment concepts and medication rules with ZHOU Zhong-ying’s own characteristics were excavated,including the basic syndrome types,basic prescriptions,commonly used medication pairs,add or subtract medication according to specific symptoms and syndrome types in TCM treatment of lung cancer.Finally,the results are confirmed by Professor LIN Lin,Professor ZHOU Zhong-ying’s disciple,and complement each other with the constructed knowledge ontology.ResultsA total of 608 cases were included,involving 237 medicines,123 symptoms and 25 syndromes.Professor ZHOU Zhong-ying’s basic syndromes in treating lung cancer cases are deficiency of both Qi and Yin,heat-toxin-phlegm-blood stasis syndrome and phlegm-blood stasis syndrome.Its location is mainly in the lung,involving the stomach,spleen,liver,kidney and brain and other viscera.The main pathological factors were deficiency,phlegm,blood stasis,heat and toxicity.The core prescription contains 12 medicines:Radix Pseudostellariae,Radix Salviae Miltiorrhizae,Radix Salviae Miltiorrhizae,Bombyx Batryticae,Herba Crani,Ophiopogon Japonicus,Turtle Shell,Coix Seed,Hedyotis Diffusa,Ranunculus Cattus,Zecha and Cichorius Volvacea.The main symptoms were cough,phlegm,dry mouth,asthma and chest tightness.Other complications included pleural effusion,chest pain,poor appetite,white phlegm,less phlegm and hemoptysis.The common tongue and pulse symptoms were dark tongue,thin,yellow and delicate tongue moss,fine and smooth pulse.Commonly used medicine pairs and medicine combinations are:Pollen Cateticorum+Faeces Trogopterus,Ligustrum Lucidum+Herba Euphorbiae,Ganoderma lucidum+Caulis Spatholobus,Peach Kernel+Almond,Radix Salviae Miltiorrhizae,Radix Ophiopogonis,Anemarrhena+Cortex Lycii+Morus alba,Clemizate Flower+Madder+Purple Perilla Seed + Luttleberry Seed,Centipede+Woodlouse Insect+Aspongopus+Turmeric,Arisaema+Centipede+Bombyx Mori+ Honeycomb,Baked Glycyrrhiza+ Pyrochore+Codonopsis pilosula+Poria+ Pinellia Ternata+Tangerine Peel+Divine Comedy.Syndrome-medication rules include:Medicated leaven,Deep-fried Atractylodes Macrocephala Koidz,Pericarpium Citri Reticulatae,Poria Cocos,Codonopsis Pilosula,Caulis Spatholobi and Roasted Licorice for stomach-spleen deficiency syndrome;Spiral Flower,Rubia Rubescens,Herba Euphorbiae and Mulberry Bark for lung collateral injury syndrome;Bombyx Batryticatus,Rhizoma Arisaematis,Luhoneycomb,Rhizoma Typhonii,Centipede,Scorpion,Pueraria Root and Chuanxiong for wind-phlegm disturbance syndrome;Astragalus Membranaceus,Rhizoma Anemarrhena for liver-kidney deficiency syndrome;Scutellaria Zi,Morus Alba and Astragalus Membranaceus for chest-hypochondriac effusion syndrome.Symptom-medication rules include:Medicated leaven,Deep-fried Atractylodes Macrocephala Koidz,Tangerine Peel,Poria cocos,Codonopsis Pilosula and Roasted Licorice for poor appetite;Caulis Caulis,Astragalus Membranaceus and Coix Seed for fatigue;Mulberry Bark,Astragalus Membranaceus and Coriander Seed for hydrothorax and abdomen effusion;Glehnia Littoralis,Radix Adenophorae Rotanini and Ophiopogon Japonicus for dry mouth;Bombyx Rigidus,Rhizoma Arisaematis,Honeycomb,White Centipede,Epimedimple Scorpion,Pueraria Lobata and Ligusticum Wallichii for dizziness,headache,numbness,obliquity of mouth and eyes,and astringency of speech.The information related to the metastasis sites are:brain metastasis is mainly identified as wind-phlegm disturbance syndrome and liver-kidney deficiency syndrome;bone metastasis is often identified as liver-kidney deficiency syndrome,and bone metastasis patients are more likely to receive chemotherapy,the associated drugs are centipede,Rhizoma Arisaematis,Dew Beehive and Bombyx Rigidus.The syndromes of patients with chemotherapy are mostly related to Qi-Yin deficiency syndrome,and their medication is mainly composed of tonifying Qi and nourishing Yin medicines,such as Radix Salviae Miltiorrhizae,Radix Ophiopogonis,Radix Pseudostellariae,Herba Craniae,Radix Astragali,Coix Miltiorrhizae,etc.The results of data mining and knowledge ontology were mutually validated,which found that:① knowledge ontology only summarized Professor ZHOU Zhong-ying’ s clinical experience in treating lung cancer,but data mining can acquire new knowledge and rules,which can be used to supplement knowledge ontology.② Knowledge ontology could only list the concepts of the same type indiscriminately.Data mining,through frequency statistics and scale-free network analysis,can distinguish the primary and secondary relationships among concepts and thus allocate weights for the corresponding concepts in knowledge ontology.③ The knowledge ontology did not involve the specific medical record information,however,the empirical rules obtained from data mining all came from the actual clinical cases.Thus,the results can be used to verify whether the relevant content of knowledge ontology is in line with clinical practice.④ Knowledge ontology lacked the construction of the relationship between similar concepts,such as common drug combinations,symptoms combinations and syndromes combinations,which can be supplemented by the results of data mining association analysis and clustering analysis.⑤ The concepts and relationships in the construction of knowledge ontology are more systematic and perfect,which can be used as a reference for the content design of data mining.⑥ The results of data mining have not been confirmed by Professor ZHOU Zhong-ying himself,while the content of knowledge ontology construction had been verified and approved by Professor ZHOU Zhong-ying himself,and combined with the previous published research results,the relevant content could provide evidence for the conclusion of data mining.ConclusionProfessor ZHOU Zhong-ying’s core pathogenesis of lung cancer is deficiency of both Qi and Yin,heat,toxin,phlegm and blood stasis combination.The main medication is tonifying qi and nourishing yin,clearing away heat,detoxifying and resolving phlegm.There is a certain relationship among different types of information,such as syndrome types,symptoms and medication.Professor ZHOU Zhong-ying has a unique law of medication for specific syndrome types,symptoms and metastasis sites.Data mining method can effectively mine Professor ZHOU Zhong-ying’s clinical experience in the treatment of lung cancer,and verify and supplement each other with the established knowledge ontology.