Table A2.

Late Recurrences Analyses Supportive Materials: Univariable and Multivariable Results for NSABP B-28.

Variable	Univariable Models		Multivariable Model*
	HR (95% CI)†	P†	HR (95% CI)†	P†
Age ≥ 50 years	1.11 (0.76 to 1.62)	0.58	1.01 (0.68 to 1.49)	0.97
≥ 4 Positive nodes	2.15 (1.49 to 3.10)	< .001	2.11 (1.43 to 3.10)	< .001
Tumor size > 2 cm	1.20 (0.83 to 1.73)	0.34	1.12 (0.76 to 1.65)	0.57
Grade		0.04		0.32
Moderate v low	1.81 (0.90 to 3.64)		1.59 (0.79 to 3.22)
Poor v low	2.34 (1.14 to 4.79)		1.70 (0.81 to 3.57)
Treatment AC→P	1.07 (0.74 to 1.55)	0.71	1.03 (0.71 to 1.50)	0.87
Mastectomy	1.10 (0.77 to 1.59)	0.60	0.91 (0.62 to 1.34)	0.64
RS‡	2.45 (1.43 to 4.21)	0.002	2.46 (1.37 to 4.43)	0.005

NOTE: Association of clinical and pathology characteristics and the continuous RS with distant recurrence risk after 5 years in patients with higher ESR1 expression (ESR1 > 9.1) in B-28, in univariable and multivariable Cox proportional hazards models.

Abbreviations: AC→P, doxorubicin plus cyclophosphamide followed by paclitaxel; HR, hazard ratio; NSABP, National Surgical Adjuvant Breast and Bowel Project; RS, recurrence score.

^*Age, tumor size, grade, treatment, and surgery type were not statistically significant in the multivariable models. Similar results were seen for the RS when adjusted for only the significant independent predictors.

^†Based on Cox proportional hazards models of distant recurrence after 5 years in patients with higher ESR1 expression. P values are from likelihood ratio tests; CIs use the Wald method.

^‡RS hazard ratio presented for a 50-point difference.