Table 2.

Impact of discordant estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 status between cores on agreement between tissue microarrays and the clinical record in phase III of the Carolina Breast Cancer Study

Central TMA				Agreement (%)	Kappa (95 % CI)
Clinical ER status	ER-negative, n (%)		ER-positive, n (%)
All cases, n = 1085
ER-negative, n (%)	259 (90)		42 (5)	94	0.84 (0.80–0.87)
ER-positive, n (%)	28 (10)		756 (95)
Concordant, n = 985
ER-negative, n (%)	217 (96)		32 (4)	96	0.89 (0.86–0.92)
ER-positive, n (%)	8 (4)		728 (96)
Discordant, n = 100
ER-negative, n (%)	42 (68)		10 (26)	70	0.39 (0.22–0.57)
ER-positive, n (%)	20 (32)		28 (74)
Clinical PR status	PR-negative, n (%)		PR-positive, n (%)
All cases, n = 1085
PR-negative, n (%)	305 (80)		41 (6)	89	0.76 (0.72–0.80)
PR-positive, n (%)	75 (20)		664 (94)
Concordant, n = 916
PR-negative, n (%)	246 (91)		27 (4)	95	0.87 (0.83–0.90)
PR-positive, n (%)	23 (9)		620 (96)
Discordant, n = 169
PR-negative, n (%)	59 (53)		14 (24)	61	0.25 (0.12–0.38)
PR-positive, n (%)	52 (47)		44 (76)
Clinical HER2 status	Negative	Equivocal	Positive
All cases, n = 1085
Negative, n (%)	864 (93)	16 (80)	28 (21)	88	0.57 (0.51–0.63)
Equivocal, n (%)	49 (5)	2 (10)	13 (10)
Positive, n (%)	17 (2)	2 (10)	94 (70)
Concordant, n = 889
Negative, n (%)	768 (99)	7 (88)	22 (21)	96	0.78 (0.72–0.84)
Equivocal, n (%)	4 (0)	0 (0)	1 (1)
Positive, n (%)	5 (1)	1 (12)	81 (78)
Discordant, n = 196
Negative, n (%)	96 (63)	9 (75)	6 (19)	57	0.16 (0.07–0.28)
Equivocal, n (%)	45 (29)	2 (17)	12 (39)
Positive, n (%)	12 (8)	1 (8)	13 (42)

Abbreviations: ER estrogen receptor, HER2 human epidermal growth factor receptor 2, PR progesterone receptor, TMA tissue microarray

Concordant cases are those with the same biomarker status across all cores for a given case, while discordant cases are those with discordant biomarker status between any two cores for a given case