Value of Constructing Prediction Model Based on Serological Markers in Diagnosis of Colorectal Cancer
Objective To explore the application value of prediction model based on serological markers in the diagnosis of colorectal cancer.Methods The clinical data of total 132 patients with suspected colo-rectal cancer who underwent colonoscopic biopsy in our hospital from January 2022 to May 2023 were ret-rospectively analyzed,and were divided into 78 patients in the colorectal cancer group and 54 patients in the non-colorectal cancer group according to the pathological examination results as the reference standard.The clinical data,blood routine indexes,serum biochemical indexes and serological markers of the two groups were recorded and compared.The indexes with differences between the two groups were included in the bi-nary logistic regression analysis and the prediction model was established.The receiver operating character-istic curve(ROC)was used to compare the diagnostic efficiency of the single index and the combined pre-diction model.Results There was no significant difference in clinical data such as age,gender,blood rou-tine index and serum biochemical index between the two groups(P＞0.05).Serum levels of carbohydrate antigen 125(CA125),carbohydrate antigen 199(C A199),carcinoembryonic antigen(CEA)and alpha-fe-toprotein(AFP)in the colorectal cancer group were higher than those in the non-colorectal cancer group(P＜0.05).Multivariate Logistic regression analysis showed that AFP[OR(95％CI):2.140(1.422～3.332),P＜0.001],CEA[OR(95％CI):1.926(1.289～2.637),P＜0.001],CA125[OR(95％CI):1.688(1.147～2.023),P＜0.001]and CA199[OR(95％CI):1.854(1.232～2.408),P＜0.001]were independent factors for the occurrence of colorectal cancer.ROC curve results showed that the AUC of the prediction model constructed by AFP,CEA,CA125 and CA199 for the diagnosis of colorectal cancer was 0.910,the specificity was 96.3％,and the accuracy was 84.7％.Conclusion The prediction model based on serological markers of AFP,CEA,CA125 and CA199 has high diagnostic value for colorectal cancer and is recommended for clinical application.
NETL
NSTL
万方数据
