摘要
Recent observational studies showed that breast-conserving surgery(BCS) resulted in superior survival compared to mastectomy in breast cancer patients. This study compared the clinical outcomes of BCS and mastectomy using propensity score(PS)matching analysis, which had advantages over conventional methods in reducing bias. Nonmetastatic breast cancer patients who underwent BCS and mastectomy were matched 1:1 based on their PS. We used the Kaplan-Meier method and Cox-regression model to estimate the treatment effects. A total of 2,866 patients with a median follow-up time of 67 months were included in the original study population. Although the mastectomy cohort(N=1,219) had more advanced disease compared to the BCS cohort(N=1,647), LRFS was similar between the two groups(93.8% vs. 92.4%, P>0.05). BCS(vs. mastectomy) was associated with improved DFS(73.8% vs. 58.7%, P<0.01) and CSS(91% vs. 78.2%, P<0.01) in the original population. In the PS-matched population(N=1,668), clinicopathological features were equally distributed between the two cohorts. BCS(vs. mastectomy) was not associated with improved DFS(70.7% vs. 66.9%, P>0.05) or CSS(87.5% vs. 84.9%, P>0.05). We found that PS methods reduce bias when estimating treatment effects using observational data. BCS and mastectomy show equivalent outcomes in nonmetastatic breast cancer patients.
Recent observational studies showed that breast-conserving surgery(BCS) resulted in superior survival compared to mastectomy in breast cancer patients. This study compared the clinical outcomes of BCS and mastectomy using propensity score(PS)matching analysis, which had advantages over conventional methods in reducing bias. Nonmetastatic breast cancer patients who underwent BCS and mastectomy were matched 1:1 based on their PS. We used the Kaplan-Meier method and Cox-regression model to estimate the treatment effects. A total of 2,866 patients with a median follow-up time of 67 months were included in the original study population. Although the mastectomy cohort(N=1,219) had more advanced disease compared to the BCS cohort(N=1,647), LRFS was similar between the two groups(93.8% vs. 92.4%, P>0.05). BCS(vs. mastectomy) was associated with improved DFS(73.8% vs. 58.7%, P<0.01) and CSS(91% vs. 78.2%, P<0.01) in the original population. In the PS-matched population(N=1,668), clinicopathological features were equally distributed between the two cohorts. BCS(vs. mastectomy) was not associated with improved DFS(70.7% vs. 66.9%, P>0.05) or CSS(87.5% vs. 84.9%, P>0.05). We found that PS methods reduce bias when estimating treatment effects using observational data. BCS and mastectomy show equivalent outcomes in nonmetastatic breast cancer patients.
基金
supported by grants from the National Key Research and Development Program of China (2016YFC1302300)
the National Natural Science Foundation of China (81720108029, 81621004, 81490750)
Guangdong Science and Technology Department (2016B030229004)
Guangzhou Science Technology and Innovation Commission (201803040015)
supported by FountainValley Life Sciences Fund of University of Chinese Academy of Sciences Education Foundation
作者简介
Kai Chen,Contributed equally to this work;Zihao Pan,Contributed equally to this work;Liling Zhu,Contributed equally to this work;Corresponding authors:Shunrong Li,email:lishr@mail.sysu.edu.cn;Corresponding authors:Erwei Song,email:songwe@mail.sysu.edu.cn.
