贝伐珠单抗治疗非小细胞肺癌疗效预测模型的构建及验证OA

Objective To develop and validate a predictive model for bevacizumab efficacy in non-small cell lung cancer(NSCLC)treatment,incorporating clinical biomarkers including vascular endothelial growth factor A(VEGFA),cell adhesion molecules(CAMS),cytokeratin 19 fragment antigen(Cyfra21-1),and glutathione S-transferase P1(GSTP1)gene.Methods This study enrolled 103 NSCLC patients treated with BEV combined with paclitaxel and cisplatin chemotherapy.After one-year follow-up,patients were categorized into responders and non-responders based on treatment efficacy.Clinical indicators were collected,and VEGFA,CAMS,Cyfra21-1,and GSTP1 gene polymorphisms were analyzed to identify predictive factors for BEV treatment in NSCLC.A nomogram model integrating clinical data and genetic polymorphisms was developed to validate its clinical value through ROC analysis.Results Patients in the effective and ineffective groups showed significant differences in age,TNM stage,number of treatment lines,and genotype composition(P＜0.05).Binary logistic regression analysis revealed that TNM stage,VEGFA rs699947 gene,CAMs rs1799969 gene,Cyfra21-1 rs17561 gene and GSTP1 rs1 695 gene were independent factors influencing the efficacy of BEV therapy for NSCLC(P＜0.05).The nomogram model developed based on these five independent predictors demonstrated excellent calibration(x2=1.593,P=0.114),with high accuracy in predicting treatment outcomes,a-chieving an AUC of 0.930.Conclusion The short-term efficacy of BEV in treating NSCLC patients is strongly associated with TNM staging and multiple genetic factors.Therefore,the nomogram predictive model developed can accurately assess short-term therapeutic outcomes.