Table 3 Logistic regressions for sentinel node status.

From: Clinical implications of intrinsic molecular subtypes of breast cancer for sentinel node status

 

BasL (n = 170)

Non-BasL (n = 1386)

 

SN−

SN+

OR [95% CI]

Pwithin

SN−

SN+

OR [95% CI]

Pwithin

Pinteration

Menopause

Pre-menopause

47 (69)

21 (31)

Ref

 

199 (61)

128 (39)

Ref

 

0.054

Post-menopause

87 (85)

15 (15)

0.39 [0.18;0.82]

0.01*

687 (65)

372 (35)

0.84 [0.65;1.09]

0.19

 

Tumor size

11–20 mm

78 (83)

16 (17)

Ref

 

621 (72)

245 (28)

Ref

 

0.39

> 20 mm

56 (74)

20 (26)

1.74 [0.83;3.66]

0.14

265 (51)

255 (49)

2.44 [1.94;3.06]

 < 0.0001*

 

Grade

1

0 (0)

0 (0)

-

-

273 (67)

133 (33)

Ref

 

0.85

2

11 (79)

3 (21)

Ref

 

427 (62)

264 (38)

   

3

113 (78)

31 (22)

1.01 [0.26;3.83]

0.99

147 (60)

96 (40)

1.15 [0.87;1.53]

0.33

 

N/A

10 (83)

2 (17)

-

-

39 (85)

7 (15)

-

-

 

BRCA1/2

Negative

68 (78)

19 (22)

Ref

 

N/A

    

Positive

25 (86)

4 (14)

0.57 [0.18;1.85]

0.35

     

N/A

41 (76)

13 (24)

-

-

     

BRCA1

Negative

75 (79)

20 (21)

Ref

 

N/A

    

Positive

18 (86)

3 (14)

0.62 [0.17;2.34]

0.48

     

N/A

41 (76)

13 (24)

-

-

     

BRCA2

Negative

86 (80)

22 (20)

Ref

 

N/A

    

Positive

7 (88)

1 (13)

0.56 [0.07;4.78]

0.59

     

N/A

41 (76)

13 (24)

-

-

     
  1. Multivariate regressions were used to analyze subtype (BasL or Non-BasL) together with each of menopausal state, tumor size, and tumor grade, respectively, and the odds ratio (OR) and 95% confidence interval (CI) within each subtype are reported. Wald tests were used to derive p values within each of the individual subtypes (Pwithin) as well as for the interaction term (Pinteraction). Because only a limited number of non-BasL patients were screened for BRCA1/2 carrier status, univariate logistic regression was performed for these features within the BasL subtype. Row-wise percentages in parentheses.
Back to article page