Table 3. Pairwise comparisons between different combinations of size and hormone receptor statuses relative to breast cancer-specific mortality^a.

Variable	Hormone receptor status
	ER+PR+		ER+PR−		ER−PR+		ER−PR−
	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P	HR (95% CI)	P
Tumor size (mm)
0–10	0.176 (0.156–0.198)	< 0.001	0.144 (0.118–0.175)	< 0.001	0.191 (0.127–0.290)	< 0.001	0.205 (0.183–0.229)	< 0.001
11–20	0.308 (0.276–0.343)	< 0.001	0.285 (0.239–0.340)	< 0.001	0.339 (0.235–0.490)	< 0.001	0.349 (0.317–0.384)	< 0.001
21–30	0.525 (0.471–0.585)	< 0.001	0.461 (0.388–0.548)	< 0.001	0.575 (0.399–0.828)	0.003	0.514 (0.468–0.565)	< 0.001
31–40	0.739 (0.660–0.827)	< 0.001	0.622 (0.520–0.745)	< 0.001	0.690 (0.473–1.008)	0.055	0.637 (0.577–0.702)	< 0.001
41–50	0.831 (0.737–0.936)	0.002	0.713 (0.588–0.863)	0.001	0.888 (0.593–1.330)	0.564	0.777 (0.701–0.862)	< 0.001
51–60	0.904(0.794–1.029)	0.127	0.769 (0.625–0.946)	0.013	0.931 (0.610–1.421)	0.740	0.856 (0.767–0.956)	0.006
61–70	Reference	–	Reference	–	Reference	–	Reference	–
71–80	1.139 (0.976–1.329)	0.099	0.828 (0.639–1.072)	0.152	0.898 (0.541–1.491)	0.678	1.060 (0.927–1.213)	0.392
>80	1.070 (0.923–1.239)	0.370	1.253 (1.009–1.555)	0.041	1.590 (1.012–2.498)	0.044	1.356(1.207–1.523)	< 0.001

Abbreviations: CI, confidence interval; ER, estrogen receptor; HR, hazard ratio; PR, progesterone receptor

^aThe results of the different combinations of size (rows) and hormone receptor status (columns) are presented at the intersections of the rows and columns. All the results were adjusted using Cox proportional hazards models for year of diagnosis, race, marital status, age, laterality, grade, node, and radiation history.