Objective To investigate the tongue image characteristics of patients in postoperative colorectal cancer and significant tongue image features using a multinomial Logistic regression model.Methods Totally 213 postoperative patients with colorectal cancer were included.TFDA-1 digital tongue diagnostic instrument was used to collect the tongue images of patients,and the statistical differences of tongue image characteristics of patients with qi deficiency syndrome,yin deficiency syndrome,phlegm dampness syndrome and blood stasis syndrome were analyzed.Multiple Logistic regression was used to analyze the different tongue image indexes of different syndrome types.Results In the tongue coating index,perAll had the following distribution order:yin deficiency syndromeqi deficiency syndrome>blood stasis syndrome.In the texture indexes,CON,ENT and MEAN had the following distribution order:phlegm-dampness syndrome>yin deficiency syndrome>qi deficiency syndrome>blood stasis syndrome.ASM had the following distribution order:phlegm-dampness syndromeqi deficiency syndrome>blood stasis syndrome.TB-a had the following distribution order:phlegm-dampness syndromeqi deficiency syndrome>phlegm-dampness syndrome>blood stasis syndrome.TC-L had the following distribution order:blood stasis syndromeqi deficiency syndrome>blood stasis syndrome>phlegm-dampness syndrome.TC-b had the following distribution order:yin deficiency syndrome>qi deficiency syndrome>blood stasis syndrome.In the multinomial Logistic regression model,compared with the qi deficiency syndrome,the probability of being diagnosed with the yin deficiency syndrome increased when TB-L and TC-a decreased.Additionally,the probability of being diagnosed with the phlegm-dampness syndrome increased when TB-b and TC-a decreased and TC-b increased;TB-b and TC-L decreased,principal component 1 of texture index increased,and the probability of syndrome differentiation as blood stasis syndrome was higher.Conclusion The tongue image characteristics of different TCM syndromes in postoperative colorectal cancer patients exhibit certain patterns,which can serve as an objective reference for TCM differentiation and diagnosis.

Objective To investigate the tongue image characteristics of patients in postoperative colorectal cancer and significant tongue image features using a multinomial Logistic regression model.Methods Totally 213 postoperative patients with colorectal cancer were included.TFDA-1 digital tongue diagnostic instrument was used to collect the tongue images of patients,and the statistical differences of tongue image characteristics of patients with qi deficiency syndrome,yin deficiency syndrome,phlegm dampness syndrome and blood stasis syndrome were analyzed.Multiple Logistic regression was used to analyze the different tongue image indexes of different syndrome types.Results In the tongue coating index,perAll had the following distribution order:yin deficiency syndromeqi deficiency syndrome>blood stasis syndrome.In the texture indexes,CON,ENT and MEAN had the following distribution order:phlegm-dampness syndrome>yin deficiency syndrome>qi deficiency syndrome>blood stasis syndrome.ASM had the following distribution order:phlegm-dampness syndromeqi deficiency syndrome>blood stasis syndrome.TB-a had the following distribution order:phlegm-dampness syndromeqi deficiency syndrome>phlegm-dampness syndrome>blood stasis syndrome.TC-L had the following distribution order:blood stasis syndromeqi deficiency syndrome>blood stasis syndrome>phlegm-dampness syndrome.TC-b had the following distribution order:yin deficiency syndrome>qi deficiency syndrome>blood stasis syndrome.In the multinomial Logistic regression model,compared with the qi deficiency syndrome,the probability of being diagnosed with the yin deficiency syndrome increased when TB-L and TC-a decreased.Additionally,the probability of being diagnosed with the phlegm-dampness syndrome increased when TB-b and TC-a decreased and TC-b increased;TB-b and TC-L decreased,principal component 1 of texture index increased,and the probability of syndrome differentiation as blood stasis syndrome was higher.Conclusion The tongue image characteristics of different TCM syndromes in postoperative colorectal cancer patients exhibit certain patterns,which can serve as an objective reference for TCM differentiation and diagnosis.

Objective To design and construct a syndrome differentiation knowledge graph that integrates TCM facial color diagnosis information;To explore the hidden relationships between the knowledge.Methods The literature data came from ancient classics,textbooks,as well as standard literature related to inspection included in the CNKI,VIP and Wanfang Data from the establishment of the databases to December 31,2022.The clinical data was sourced from 30 elderly individuals who underwent TCM health examinations at the Community Health Service Center in Jiading Industrial Zone,Shanghai in September 2022.Facial image acquisition was performed using TFDA-1 digital tongue and facial diagnostic instrument.By following the steps of knowledge extraction,knowledge fusion and quality assessment to construct a graph,and with the assistance of TCM experts for interpretation,using Access 2019 to integrate qualitative textual data and quantitative objective image digital information,a syndrome differentiation knowledge graph integrating TCM facial diagnosis was designed and completed in the Neo4j graph database.In addition,a method was designed to shift facial diagnosis knowledge from qualitative to quantitative.Results There were a total of 194 nodes under 8 entity term types and 12 entity term labels,as well as 361 relationships under 13 semantic relationships in knowledge graph.The Neo4j graph database provided a visualized TCM facial color diagnosis and differentiation,which could be queried and fed back using Cypher language.Conclusion The knowledge graph constructed based on the theory of TCM facial color diagnosis visually shows the complex correlation between facial color diagnosis and syndrome differentiation diagnosis,with a knowledge representation model that forms qualitative data of image features → semantic relationships → syndrome differentiation diagnosis forms.