Skip to main content
. 2020 Nov 9;10:592998. doi: 10.3389/fonc.2020.592998

Table 3.

Overall results and subgroup/sensitivity analyses.

Variables Heterogeneity Meta-analyses
I-square p value Model OR (95% CI) p value
Overall 88.1% <0.001 Random 1.340 (0.6772.653) 0.400
Stage I - III 93.6% <0.001 Random 1.535 (0.4335.441) 0.507
II - III 0.0% 0.557 Fixed 1.068 (0.7541.512) 0.711
Subtype HER2- 0.0% 0.698 Fixed 0.935 (0.7161.221) 0.622
TNBC 0.0% 0.800 Fixed 1.028 (0.7791.356) 0.846
Study type* PCS 95.7% <0.001 Random 2.241 (0.41012.259) 0.352
RCT 0.0% 0.487 Fixed 0.930 (0.6741.283) 0.659
Quality Score* Low 95.7% <0.001 Random 2.241 (0.41012.259) 0.352
High 0.0% 0.487 Fixed 0.930 (0.6741.283) 0.659
Study population& Subgroup 0.0% 0.800 Fixed 1.028 (0.7791.356) 0.846
Whole cohort 96.1% <0.001 Random 1.950 (0.21217.913) 0.555
BRCA status* BRCA1 only 95.7% <0.001 Random 2.241 (0.41012.259) 0.352
BRCA1 & 2 0.0% 0.487 Fixed 0.930 (0.6741.283) 0.659
Experiment regimen& Combination 0.0% 0.800 Fixed 1.028 (0.7791.356) 0.846
Single 96.1% <0.001 Random 1.950 (0.21217.913) 0.555
Platinum& Carboplatin 0.0% 0.800 Fixed 1.028 (0.7791.356) 0.846
Cisplatin 96.1% <0.001 Random 1.950 (0.21217.913) 0.555

CI, Confidence interval; OR, Odds ratio; PCS, Prospective cohort study; RCT, Randomized control trial; TNBC, Triple negative breast cancer.

*Collinearity existed among these three variables, so the subgroup analyses based on these three variables drew the same conclusion.

&Collinearity existed among these three variables, so the subgroup analyses based on these three variables drew the same conclusion.