. 2006 Dec 29;8(6):R73. doi: 10.1186/bcr1631

Table 1.

Clinicopathological characteristics and univariate analysis of ability to predict failure of tamoxifen therapy

	Population-based cohort				Case-control cohort
			RFS				RFS

Variable	n^a	Percentage^b	8 years	P^c	n^a	Percentage^b	8 years	P^c

All patients	9/60	15	79		13/31	42	67
Nodal status	9/60	15		0.04	13/28	46		0.30
Negative	4/45	9	85		8/15	53	57
Positive	5/15	33	64		5/13	38	71
AJCC stage	9/60	15		0.16	ND^d	ND^d
1	2/28	7	92
2a	0/3	0	100^e
2b	7/27	26	58
3a	0/1	0	100^e
Tumor stage	9/60	15		0.01	13/31	42		0.17
1	2/32	6	93		3/12	25	74
2	4/21	21	55		10/19	53	62
3	3/4	75	50^e
4	0/1	0	100
Tumor grade	9/60	15		0.06	13/31	42		0.56
1	0/5	0	100		0/1	0	100
2	2/22	9	82		8/19	45	63
3	7/26	27	69		5/11	40	70
PR status	9/60	15		0.06	13/31	47		0.96
Negative	0/16	0	100		3/6	50	67
Positive	6/36	17	71		10/25	40	67
Age	9/60	15		0.15	13/31	42		0.56
<50 years	2/6	33	83^e		2/2	100	100
≥50 years	7/54	13	84		11/29	38	64

^a9/60 indicates that there were 9 treatment failures out of 60 patients with known values for the indicated clinical characteristic. ^bThe percentage of treatment failures for the indicated clinical characteristic. ^cP is given for log-rank (or log-rank trend) univariate analysis. Sensitivity tests conducted with missing values provided no evidence for selection bias. ^dClinical characteristic was not determined. ^eFor patients with AJCC stage 2a and 3a disease and patients less than 50 years old, the RFS is reported at 7 years. For patients with tumor stage 3 disease, the RFS is reported at 6 years. AJCC, American Joint Committee on Cancer; PR, progesterone receptor; RFS, relapse-free survival.