Prognostic Models for Nonmetastatic Triple-Negative Breast Cancer Based on the Pretreatment Serum Tumor Markers with Machine Learning
Table 4
Univariable and multivariable analyses for DFS in stage I-III TNBC patients.
Characteristics
Univariable analysis
p value
Multivariable analysis
p value
HR
95%CI
HR
95%CI
Age at diagnosis
<40
Ref
0.242
-
40∼50
0.449
0.184–1.096
-
50∼60
0.437
0.178–1.075
-
≥60
0.630
0.257–1.548
-
Grade
I
Ref
0.169
-
II
1.743
0.234–12.953
-
III
0.917
0.121–6.983
-
Ki-67
<30%
Ref
0.646
-
≥30%
0.857
0.444–1.654
-
T-stage
1
Ref
0.093
Ref
0.382
2
1.876
1.037–3.393
1.587
0.842–2.990
3
2.249
0.764–6.622
2.090
0.686–6.364
4
1.896
0.254–14.175
2.044
0.268–15.580
N-stage
0
Ref
<0.001∗∗∗
Ref
<0.001∗∗∗
1
0.903
0.396–2.057
0.658
0.263–1.650
2
3.732
1.705–8.171
2.767
1.218–6.288
3
7.775
3.547–17.040
4.980
2.081–11.917
Stage
I
Ref
<0.001∗∗∗
-
II
1.651
0.816–3.340
-
III
5.380
2.564–11.288
-
TMRS groups
Low TMRS
Ref
<0.001∗∗∗
Ref
0.002∗∗
High TMRS
3.173
1.718–5.862
2.847
1.473–5.506
Cox’s proportional hazard analysis was carried out for univariable and multivariable analyses to identify independent prognostic factors for DFS in stage I-III TNBC patients. Multivariable analysis was performed further for the factor whose in univariable analysis. , indicate a significant difference. DFS: disease-free survival; TNBC: triple-negative breast cancer; HR: hazard ratio; CI: confidence interval; TMRS: tumor marker risk score.