Application Of Combined Tumor Marker Model In Early Diagnosis And Staging Of Lung Cancer

ObjectiveTo explore the value of serum tumor markers in the diagnosis and staging of lung cancer,to establish a logistic regression model of multiple serum tumor markers and to explore its application in the early diagnosis of lung cancer.Methods1.Select the subjectsFrom October 2019 to December 2022,250 patients with histopathologically confirmed primary lung cancer(PLC)in Yunnan Cancer Hospital were selected as the first diagnosis,and another 101 patients with benign lung diseases(BPD)in Yunnan Cancer Hospital during the same period were selected as the control group.2.Information collection and tumor marker detectionTNM stage,clinical stage,age,sex,smoking history and drinking history of lung cancer patients were recorded.Enzyme-linked immunosorbent assay(ELISA)or electrochemiluminescence was used to determine the values of various tumor markers(CEA,CA125,CA153,CA199,CA242,CA724,NSE,Pro GRP,CYFRA 211,Fer,SCC)in 250 patients with primary lung cancer(PLC)and 101 patients with benign lung disease.3.Result analysis and model establishmentThe differences and correlations of the expression levels of 11 tumor markers between the lung cancer group and the benign lung disease group were analyzed;the differences in the expression levels of 11 tumor markers between different stages of lung cancer were further compared;the diagnostic value of CEA,CA125,CA153,CA99,CA242,CA724,NSE,Pro GRP,CYFRA211,Fer,and SCC for lung cancer was compared with receiver operating characteristic curve(ROC);the combined screening of tumor markers and the establishment,validation,and evaluation of logistic regression models were performed.Results1.Spearman rank correlation was used to analyze the correlation between age,smoking history,drinking history,CEA,CA125,CA153,CAAl99,CA242,CA724,NSE,Pro GRP,CYFRA211,Fer,SCC and lung cancer.The results showed that age,smoking history,drinking history,CEA,CA125,CA153,NSE,Pro GRP,CYFRA211,Fer,SCC were positively correlated with lung cancer,and the correlation coefficients r were 0.224,0.187,0.187,0.444,0.323,0.234,0.419,0.278,0.469,0.211,and 0.257,respectively,and the differences were statistically significant(P<0.05).Age,CA199,CA242,and CA724 were not associated with lung cancer,and the differences were not statistically significant(P>0.05).2.The levels of age,sex,smoking,drinking,CEA,CA125,CA153,NSE,Pro GRP,CYFRA211,Fer and SCC in lung cancer group and benign lung disease group were significantly higher than those in benign lung disease group(P<0.05).Common chi-square tests for gender(male),smoking,and alcohol consumption showedχ~2=11.53,15.65,and 10.897,P<0.05.The results of Mann Whitney U test for CEA,CA125,CA153,NSE,Pro GRP,CYFRA211,Fer,and SCC showed that the Z values were:-7.836,-5.699,-4.095,-7.146,-2.678,-8.136,-3.519,and-4.517,P<0.05.CAL99,CA242,and CA724 levels were not significantly different between the two groups,and the difference was not statistically significant(P>0.05).3.Compared between different TNM stages in the lung cancer group,CEA,CA125,CA153,CA199,NSE,CYFRA211,Fer,and SCC levels showed a trend to increase from T1 to T4(P<0.05),CEA,CA125,CA153,CA199,NSE,CYFRA211,and Fer levels showed a trend to increase from N0 to N3(P<0.05),and CEA,CA125,CA153,NSE,CYFRA211,and Fer levels were significantly higher in M1 than in M0(P<0.05),and NSE levels were significantly increased in stage I patients in the lung cancer group compared with the benign disease group after pairwise comparison analysis by Kruskal Wallis test(P<0.05),and CEA,CA125,CA153,NSE,CYFRA211,and Fer levels were significantly increased in stage III and IV patients compared with stage I patients(P<0.05).4.Among the 11 serum tumor markers,CEA,CA125,CA153,NSE,Pro GRP,CYFRA211,Fer,and SCC had high diagnostic efficacy(AUC≥0.6),with CYFRA211having the highest diagnostic efficacy(AUC=0.811),61.3%sensitivity and 94.8%specificity at a cut-off value of 3.25 ng/m L,and the highest positive likelihood ratio of 11.788,with the strongest judgment of the disease.The P values of CA199,CA242,and CA724 were 0.111,0.452,and 0.724,respectively,and the difference was not statistically significant(P>0.05).5.Using the three models with the highest scores established by binary logistic regression analysis,the AUC was 0.871 for Group1(CEA+NSE),0.890 for Group2(CEA+NSE+CYFRA211),and 0.900 for Group3(CEA+NSE+CYFRA211+CA125).The diagnostic efficacy of combined detection of each marker was higher than that of single marker.Among them,Group3(CEA+NSE+CYFRA211+CA125)had the highest area under the curve of 0.900,sensitivity 71%,specificity 97.4%,and the highest Youden index of 0.689,which was the marker combination with the strongest diagnostic ability and was the optimal combined diagnostic model,and its obtained predictive probability P value is:P=0.427*CEA+0.249*NSE+0.602*CYFRA211+0.060*CA125-6.544Conclusion1.Serum tumor marker levels are closely related to TNM stage and clinical stage of patients and are important for the early diagnosis of lung cancer.2.There are differences in the combined diagnostic efficacy of serum tumor markers established by logistic regression,of which CEA+NSE+CYFRA211+CA125 four indicators have the best combined diagnostic efficacy and have certain value in the early diagnosis of lung cancer,which can provide a reference for clinical decision-making optimization.