基于SEER数据库构建预测早期老年HER2阳性乳腺癌生存概率的列线图OA北大核心CSTPCD

AIM:To construct column-line plots to predict survival in elderly patients with early-stage HER2-positive breast cancer using the Surveillance,Epidemiology and End Results(SEER)database.METHODS:5 220(based on the era of single-target-ed therapy)and 1 176(based on the era of dual-targeted therapy)patients screened in the SEER da-tabase were randomized into a training group and an internal validation group.COX proportional risk regression was used to screen survival-related pre-dictors and build a column-line graphical model,and the accuracy and utility of the model were test-ed using the consistency index(C-index),calibra-tion curves,and time-dependent ROC curves.Pa-tients receiving chemotherapy and non-chemother-apy were statistically paired using two-group pro-pensity score matching,and subgroup analyses were performed on the screened variables.RE-SULTS:The single-targeted therapy era line graph was constructed from seven variables:age,marital status,T-stage,N-stage,surgery,chemotherapy,and radiotherapy.The dual-targeted therapy era line graph was constructed from five variables:age,AJCC staging,surgery,chemotherapy,and radio-therapy.The results of the subgroup analysis showed that older HER2-positive breast cancer pa-tients who received chemotherapy had better OS.CONCLUSION:Based on the SEER database,an ac-curate column-line graph predicting survival in el-derly patients with early-stage HER2-positive breast cancer was established and validated.This study suggests that chemotherapy increases surviv-al benefit in elderly patients.