Breast cancer risk prediction in women aged 35–50 years: impact of including sex hormone concentrations in the Gail model

Tess V. Clendenen; Wenzhen Ge; Karen L. Koenig; Yelena Afanasyeva; Claudia Agnoli; Louise A. Brinton; Farbod Darvishian; Joanne F. Dorgan; A. Heather Eliassen; Roni T. Falk; Göran Hallmans; Susan E. Hankinson; Judith Hoffman-Bolton; Timothy J. Key; Vittorio Krogh; Hazel B. Nichols; Dale P. Sandler; Minouk J. Schoemaker; Patrick M. Sluss; Malin Sund; Anthony J. Swerdlow; Kala Visvanathan; Anne Zeleniuch-Jacquotte; Mengling Liu

doi:10.1186/s13058-019-1126-z

Breast Cancer Research

Table 1 Descriptive characteristics of invasive breast cancer cases and matched controls

From: Breast cancer risk prediction in women aged 35–50 years: impact of including sex hormone concentrations in the Gail model

	Cases (n = 1762)	Controls (n = 1890)
Cohort, n
BGS	230	230
CLUE II	87	87
CSB	69	69
Guernsey	124	124
NHS	93	93
NHS II	248	250
NSMSC	31	31
NYUWHS	493	496
ORDET	214	224
Sister	173	286
Age at blood donation, years, n (%)
35–40	472 (26.8)	487 (25.8)
41–45	708 (40.2)	752 (39.8)
46–50^a	582 (33.0)	651 (34.5)
Race/ethnicity, n (%)
White	1587 (90.1)	1696 (89.7)
Black/African American	76 (4.3)	73 (3.9)
Other or missing	99 (5.6)	121 (6.4)
Age at diagnosis, years, n (%)
35–45	287 (16.3)
46–50	579 (32.9)
51–55	436 (24.7)
56–60	235 (13.3)
61–65	141 (8.0)
> 65	84 (4.8)
Lag time between blood donation and diagnosis, years, n (%)
0–2	274 (15.6)
3–5	420 (23.8)
6–10	443 (25.1)
11–15	286 (16.2)
16–20	201 (11.4)
> 20	138 (7.8)
Age at menarche, years, n (%)
< 12	376 (21.3)	411 (21.7)
12–13	976 (55.4)	1012 (53.5)
≥14 or missing^b	410 (23.3)	467 (24.7)
Age at first live birth, years, n (%)
< 20 or missing^b	114 (6.5)	143 (7.6)
20–24	457 (25.9)	521 (27.6)
25–29^c	473 (26.8)	511 (27.1)
≥ 30	304 (17.3)	307 (16.2)
Nulliparous	414 (23.5)	408 (21.5)
Number of benign breast biopsies, n (%)
0 or missing^b	1339 (76.0)	1559 (82.5)
≥ 1	423 (24.0)	331 (17.5)
0	1311 (74.4)	1415 (74.9)
1^d	382 (21.7)	412 (21.8)
> 1^d	69 (3.9)	63 (3.3)
BMI, kg/m², n (%)
< 25	1097 (59.9)	1124 (62.6)
25–29	420 (24.8)	465 (24.0)
≥ 30	234 (15.4)	289 (13.4)
Missing	11	12
AMH cohort-specific quartiles, n(%)
Q1	365 (20.7)	480 (25.4)
Q2	444 (25.1)	468 (24.8)
Q3	453 (25.7)	468 (24.8)
Q4	500 (28.4)	474 (25.1)
Testosterone cohort-specific quartiles, n (%)
Q1	423 (24.0)	511 (27.0)
Q2	414 (23.5)	464 (24.6)
Q3	452 (25.7)	460 (24.3)
Q4	473 (26.8)	455 (24.1)
BCRAT 5-year risk score (%), n (%)^e
< 0.6%	296 (16.8)	332 (17.6)
0.6–0.99%	679 (38.5)	765 (40.5)
1–1.66%	525 (29.8)	517 (27.3)
1.67–1.99%	110 (6.2)	130 (6.9)
2–2.99%	115 (6.5)	115 (6.1)
≥ 3%	37 (2.1)	31 (1.6)
ER status, n (%)
ER-positive	1139 (79.8)
ER-negative	289 (20.2)
Unknown	334

Note: Cases and controls were matched 1:1 for all cohorts except for Sister Study which matched 1:2
^aAll cases had age at blood donation ≤ 50, though for 24 sets, matched controls ages were ≤ 51.2 years at blood donation
^bTo be consistent with BCRAT, which imputes missing data to the lowest risk category, we imputed missing data as follows: age at menarche: ≥ 14 for 35 cases (1.5%) and 49 controls (1.9%); age at first live birth: < 20 for 5 cases (0.2%) and 7 (0.3%) controls; number of breast biopsies: 0 for 42 cases (1.8%) and 40 controls (1.6%)
^cAs done in BCRAT, nulliparous and women who were 25–29 at first birth were combined in all models
^dThe number of first-degree family members with breast cancer was coded as 0, 1, or > 1 affected relatives. For cohorts that collected family history as a no/yes variable, “yes” answers were assigned to the intermediate category (1 affected relative)
^eCalculated using the following variables: race, age at menarche, age at first live birth, number of breast biopsies, and number of first-degree family members with breast cancer, history of atypical hyperplasia was missing for all cohorts and set to “no.” Gail model 2 rates and parameters were used as described in [14]

Back to article page