Study On Diagnostic Features And Risk Prediction Model Of Tongue Image In Patients With Non-Small Cell Lung Cancer

Goal:The characteristics of tongue image diagnosis in patients with Non-small Cell Lung Cancer(NSCLC)were the main research contents.T he changes of tongue image features in patients with NSCLC during the occurrence and development of lung cancer and during treatment were analyzed.This paper analyzes the different pathological types,pathological stages,intervention schemes,tongue image features of TCM syndrome types,and the dynamic relationship between tongue image features and blood indexes,and establishes a risk prediction model of non-small cell lung cancer based on tongue image features,which provides technical support for the establishment of integrated traditional Chinese and western medicine lung cancer risk early warning,disease syndrome diagnosis and curative effect evaluation system.Methods:(1)The tongue image features of 309 patients(1235 cases)with NSCLC and 206 control groups were collected and analyzed.T test or rank sum test and Logistic regression analysis were used to analyze the characteristics of different pathological classification,pathological stage,intervention scheme and tongue image index of TCM syndrome type;(2)Taking the progress of the disease as the observation point,38 patients(338times)with NSCLC were monitored dynamically.The "tongue image-blood" index was analyzed by linear stepwise regression analysis,and the dynamic changes of tongue image characteristics and blood indexes during the progress of NSCLC were observed.(3)Logistic regression analysis was used to screen the characteristic tongue image index,and the NSCLC related risk prediction model based on tongue image feature was established to determine whether NSCLC,had distant transfer or not.Risk Assessment Plot visualization method was used to evaluate and analyze the classification ability of risk model,and Naive Bayes,Random Forest,SVM data mining method was used to verify the classification and prediction ability of the model.Results:(1)Compared with the control group,the features of tongue image in NSCLC group were different.PerAll and TC-L,TC-a in NSCLC group were higher than those in healthy control group,suggesting that the tongue quality,the brightness of tongue coating,the decrease of red composition,the thickening of tongue coating and the increase of yellow component in NSCLC group were higher than those in control group.The results of NSCLC tongue image in different pathological stages showed that TB-b and TB-a in stage ? were lower than those in stage ?,and TB-b in stage ? was lower than that in stage ? ?,suggesting that the blue and red components of tongue in ? stage were increased,and the bluish and purple tongue might appear.The results showed that compared with the healthy control group,the TB-H,TC-a and other indexes decreased,the tongue quality was lighter and the tongue coating was white in the qi deficiency syndrome group,the TB-Cr,perAll index in the yin deficiency syndrome group was increased,the tongue color was reddish and the tongue coating thickened,the perAll in the phlegm dampness syndrome group was increased,and the tongue coating was the thickest in the phlegm dampness syndrome group.In the blood stasis syndrome group,TB-Cr,TC-b increased,tongue color red component increased,tongue coating yellowed,suggesting red tongue,yellow coating.The results of tongue image analysis showed that TB-a was negatively correlated with operation,suggesting that the tongue color of operation group was lighter than that of non-operation group.In IV patients,TB-b decreased and tongue color tended to be bluish in patients taking targeted drugs,which was related to the age and course of disease.(2)38 cases(338 cases)were observed dynamically,and the changes of blood indexes and tongue image indexes before and after tumor progress were analyzed.The results showed that there were statistical differences in blood index and tongue image index before and after progress,CEA,per All,TB-a and other indexes increased.Linear stepwise regression analysis showed that there was a linear relationship between CEA and perAll,NSE and TC-L.The "tongue image-blood" index map showed that there was a common change trend between tongue image index and blood index in the process of disease progress.(3)The risk prediction model of non-small cell lung cancer(NSCLC)was constructed by using tongue image and clinical indexes.Logistic regression showed that the specific index for NSCLC classification was perAll and TC-b,after adding tongue image index on the basis of clinical index,the classification ability of NSCLC risk prediction model was significantly improved.Logistic regression showed that the specific index for NSCLC distant transfer classification was TB-a,TB-b,TC-b,and the classification and prediction ability of NSCLC distant transfer risk prediction model was significantly improved after adding tongue image index.Conclusions:1.There are differences in tongue image features in different pathological states,intervention schemes and TCM syndromes in patients with non-small cell lung cancer,which can provide quantitative basis for clinical diagnosis and treatment of non-small cell lung cancer.2.Tongue image features have certain regularity in the progress of non-small cell lung cancer(NSCLC),and have certain correlation with the changes of blood biochemistry and tumor markers.In the process of integrated traditional Chinese and western medicine in the treatment of non-small cell lung cancer(NSCLC),continuous monitoring of tongue image features and blood indexes is of auxiliary significance for prognosis judgment.3.Based on the features of tongue image,the diagnosis and distant metastasis risk prediction model of non-small cell lung cancer can be established.The model can be used to assist diagnosis,disease monitoring,prognosis judgment and so on.