Table 2.

Logistic regression analyses for the prediction of the presence of lymph node metastasis by variables of tumor chronology and biology

Fitted model (N = 3002)	Predictor	Effect	Odds ratio (95% CI)	P-value	AUC (95% CI)
cT	cT	1 vs 4	0.123 (0.058; 0.2610)	<.0001	0.6231 (0.6048; 0.6414)
		2 vs 4	0.270 (0.128; 0.572)
		3 vs 4	0.519 (0.234; 1.152)
pT	pT	1 vs 4	0.129 (0.050; 0.332)	<.0001	0.6700 (0.6522; 0.6877)
		2 vs 4	40.352 (0.137; 0.906)
		3 vs 4	1.338 (0.501; 3.575)
LVI	LVI	Yes vs No	6.782 (5.587; 8.234)	<.0001	0.6620 (0.6466; 0.6775)
Grade	Grade	1 vs 3	0.499 (0.396; 0.630)	<.0001	0.5531 (0.5341; 0.5721)
		2 vs 3	0.830 (0.710; 0.971)
ER	ER	Neg vs Pos	0.975 (0.797; 1.194)	0.8076	0.5016 (0.4885; 0.5148)
PR	PR	Neg vs Pos	1.008 (0.854; 1.189)	0.9248	0.5008 (0.4848; 0.5168)
HER–2	HER-2	Neg vs Pos	0.703 (0.559; 0.884)	0.0026	0.5176 (0.5059; 0.5292)
grade, ER, PR, HER–2					0.5649 (0.5448; 0.5850)
cT, grade, ER, PR, HER–2					0.6406 (0.6206; 0.6606)
LVI, grade, ER, PR, HER–2					0.6852 (0.6657; 0.7047)
pT, LVI, grade, ER, PR, HER–2					0.7462 (0.7279; 0.7645)