- Split View
-
Views
-
Cite
Cite
Tia R. Milanese, Lynn C. Hartmann, Thomas A. Sellers, Marlene H. Frost, Robert A. Vierkant, Shaun D. Maloney, V. Shane Pankratz, Amy C. Degnim, Celine M. Vachon, Carol A. Reynolds, Romayne A. Thompson, L. Joseph Melton, Ellen L. Goode, Daniel W. Visscher, Age-Related Lobular Involution and Risk of Breast Cancer, JNCI: Journal of the National Cancer Institute, Volume 98, Issue 22, 15 November 2006, Pages 1600–1607, https://doi.org/10.1093/jnci/djj439
- Share Icon Share
Abstract
Background : As women age, the lobules in their breasts undergo involution or regression. We investigated whether lobular involution in women with benign breast disease was associated with subsequent breast cancer risk. Methods : We examined biopsy specimens of 8736 women in the Mayo Benign Breast Disease Cohort from whom biopsy samples were taken between January 1, 1967, and December 31, 1991. Median follow-up for breast cancer outcomes was 17 years. We classified lobular involution in the background breast tissue as none (0% involuted lobules), partial (1%–74%), or complete (≥75%). Subsequent breast cancer events and data on other risk factors were obtained from medical records and follow-up questionnaires. To estimate relative risks (RRs), standardized incidence ratios were calculated by use of incidence rates from the Iowa Surveillance, Epidemiology, and End Results (SEER) Registry. All statistical tests were two-sided. Results : Distribution of extent of involution was none among 1627 (18.6%) women, partial among 5197 (59.5%), and complete among 1912 (21.9%). Increased involution was positively associated with increased age and inversely associated with parity (both P <.001). The relative risk for the entire cohort of 8736 women, compared with the Iowa SEER population, was 1.40 (95% CI = 1.30 to 1.51). Risk of breast cancer was associated with the extent of involution (for no involution, RR [i.e., observed versus expected] = 1.88, 95% confidence interval [CI] = 1.59 to 2.21; for partial involution, RR = 1.47, 95% CI = 1.33 to 1.61; and for complete involution, RR = 0.91, 95% CI = 0.75 to 1.10; test for heterogeneity P <.001). Lobular involution modified risk in all subsets (e.g., among women with atypia, for no involution, RR = 7.79, 95% CI = 3.56 to 14.81; for partial involution, RR = 4.06, 95% CI = 3.03 to 5.33; and for complete involution, RR = 1.49, 95% CI = 0.41 to 3.82; P = .003). Conclusions : In this large cohort of women with benign breast disease, lobular involution was associated with reduced risk of breast cancer. Aberrant involution may be a biologically important phenomenon in breast cancer biology.
The mammary gland undergoes profound physiologic changes throughout the phases of a woman's life, including puberty, pregnancy, lactation, postlactational involution, and aging ( 1 , 2 ) . The epithelial tissue of the human breast is organized into 15–20 major lobes, each made up of lobules that contain the milk-forming acini. As a woman ages, breast lobules regress, or involute, with a reduction in the number and size of acini per lobule and replacement of the delicate intralobular stroma with the more dense collagen of breast connective tissue ( Fig. 1 ) ( 1 , 3 – 5 ) . Over time, there is progressive fatty replacement of glandular elements and collagen ( 1 , 5 ) . This process differs greatly from postlactational involution. After lactation, there is regression of all breast tissues as secretory activity is curtailed, but there is no substantive loss of glandular tissue ( 1 , 2 ) .
Although involution of the breast involves a consistent sequence of histologic changes, the rate and extent of involution vary considerably among individual women ( 3 , 4 ) . This age-related lobular involution has been documented in women younger than age 40 years and thus involves factors not limited to the onset of menopause ( 1 , 3 – 5 ) . Cowan et al. ( 4 ) studied age-related involution in breast tissue obtained at autopsy and speculated that obstetrical and lactational history may be more important than age at influencing the onset of involution, but they did not provide specific obstetrical data to support their claim. Geschickter ( 5 ) studied more than 100 breast specimens obtained at autopsy or surgery and noted the degree of involution by age and obstetrical history. He observed early changes consistent with lobular involution in 33% of women younger than age 40 years. He also found that repeated pregnancies were associated with the persistence of lobules, whereas lobule size and number declined in the absence of childbearing.
It has been hypothesized that age-related lobular involution, with its loss of glandular elements, may be associated with a decreased risk of breast cancer ( 6 , 7 ) . To test this hypothesis, we characterized the extent of lobular involution by age group in a large cohort of women with benign breast disease and examined the association between involution and breast cancer risk.
P ARTICIPANTS AND M ETHODS
Study Population
Inclusion and exclusion criteria for the study cohort have been previously described ( 8 ) . Briefly, the study population consisted of women aged 18–85 years who had benign breast disease (i.e., a breast biopsy examination with benign findings) diagnosed via surgical excision at Mayo Clinic between January 1, 1967, and December 31, 1991. The Mayo Benign Breast Disease Cohort included 9087 women with 15 years of follow-up at initial report ( 8 ) . Since that report, we obtained cancer follow-up data for an additional 289 women, for whom that information was lacking previously, bringing the cohort to a total of 9376 women with a median of 17 years of follow-up. For 640 breast biopsy samples, the biopsy specimen consisted entirely of the index lesion; there was no background breast tissue in which to determine the degree of lobular involution. Thus, the final cohort for this analysis included 8736 women.
All protocol procedures and patient contact materials were reviewed and approved by the Institutional Review Board of the Mayo Clinic. Return of the patient contact materials was considered implied consent.
Histology
All slides were reviewed by a breast pathologist (DWV) without knowledge of patient age, cancer outcome, or original histologic diagnosis. Biopsy findings were classified by the most extreme degree of hyperplasia as nonproliferative, proliferative disease without atypia, or atypical hyperplasia, as previously reported ( 8 ).
Each biopsy specimen was also categorized according to the extent of lobular involution in the background breast tissue. Involuted terminal duct lobular units (TDLUs) contain only a few to several small acini that may be distended by cystic change ( Fig. 1 ). Involuted lobules also have flattened inconspicuous acinar epithelium with fibrosis of specialized intralobular stroma. The degree of involution for each specimen was categorized as none (0% TDLUs involuted), partial (1%–74% TDLUs involuted), or complete ( ≥ 75% TDLUs involuted). These cut points were set by the pathologists at the initiation of the study to best distinguish the extremes of no involution from near-complete involution.
In general, viewing five to six lobules was sufficient to assess the extent of involution. One slide from a breast specimen typically contained a dozen or more lobules. There are two exceptions to this statement: 1) when involution was extensive and there are only a few lobular remnants on the slide (which is sufficient to state that complete lobular involution has occurred) and 2) when the entire sample consists of an epithelial hyperplastic lesion, as was the case for 640 (6.8%) of the 9376 women in our original cohort.
Risk Factor Information and Follow-up
To obtain information about family history, reproductive history, and use of hormone replacement therapy, a study-specific questionnaire was sent to patients; 5352 (61%) of the 8736 women or their next of kin returned the questionnaire. Follow-up for breast cancer events was obtained through comprehensive (inpatient and outpatient) Mayo medical records and the questionnaire.
Family history of breast cancer was categorized as strong, weak, or negative. A strong family history was defined as the patient having 1) at least one first-degree relative with breast cancer diagnosed before age 50 years or 2) two relatives with breast cancer at any age, with at least one being a first-degree relative. Patients with family history of breast cancer who did not meet the above criteria were categorized as having a weak family history ( 8 ).
Statistical Analysis
Data were summarized descriptively by use of frequencies and percentages. We initially compared the unadjusted distribution of breast cancer risk factors across levels of involution with chi-square tests of statistical significance. Subsequent comparisons were made after accounting for the effects of age by use of multicategorical, nominal logistic regression analysis ( 9 ) . We summarized results from these analyses by use of adjusted percents, carried out by calculating log odds estimates for each 10-year age category (<40, 40–49, 50–59, 60–69, 70–79, or ≥80 years), back-transforming to percent estimates, and then averaging the corresponding percents across all sets of age. This approach was similar to a least-squares means estimate in an analysis of variance setting. Among 245 women with synchronous bilateral biopsy examinations, we assessed the level of agreement across the two readings by use of weighted kappa statistics and their corresponding 95% confidence intervals (CIs).
The length of follow-up for each woman in the study was calculated as the number of days from her biopsy examination to the date of her breast cancer diagnosis, death, or last contact. We estimated relative risks (RRs) on the basis of standardized incidence ratios by dividing the observed numbers of incident breast cancers by expected numbers of population-based incident breast cancers. Expected values were calculated by apportioning each woman's person-years of follow-up into 5-year age and calendar-period categories and multiplying these by the corresponding breast cancer incidence rates from the Iowa Surveillance, Epidemiology, and End Results (SEER) Registry. This reference population was chosen because of its demographic similarities to the Mayo Clinic population (80% of cohort members reside in the upper Midwest). Potential heterogeneity in relative risks across levels of involution was assessed by use of Poisson regression analysis, with the log-transformed expected event rate for each individual modeled as the offset term.
In addition to assessing overall breast cancer risk, we also compared rates of ipsilateral to contralateral breast cancer in relation to the side of the benign lesion, both overall and by levels of involution. When calculating incidence for ipsilateral cancer, individuals with contralateral cancer were censored at their date of diagnosis, and vice versa. Women with missing laterality information, bilateral biopsy examination results, or bilateral breast cancer were censored for both events in these analyses. This approach yielded identical numbers of person-years for each type of event. As a result, the length of follow-up was not a factor in the analysis, and the rate comparisons reduced to simple comparisons of the number of events. Thus, we were able to assess whether the relative rate of ipsilateral cancer (compared with contralateral cancer) differed across levels of involution using simple chi-square tests of statistical significance. All statistical tests were two-sided, and all analyses were carried out with the SAS software system (SAS Institute, Inc, Cary, NC).
R ESULTS
Extent of Lobular Involution
We characterized the extent of lobular involution in the benign breast biopsies of a cohort of 8736 women with tissue sampled between January 1, 1967, and December 31, 1991, at the Mayo Clinic. The distribution of the patients by level of lobular involution was as follows: no involution among 1627 (18.6%) women, partial involution among 5197 (59.5%) women, and complete involution among 1912 (21.9%) women.
Factors Associated With Involution
As shown in Table 1 , the degree of lobular involution increased progressively with age at diagnosis of benign breast disease ( P <.001). Complete involution of lobular units was observed in only 19 (3.4%) of the 566 women who were younger than 30 years at their benign biopsy; in 53 (5.1%) of the 1037 women aged 30–39 years; in 142 (5.8%) of the 2446 women aged 40–49 years; in 455 (21.6%) of the 2109 women aged 50–59 years; in 724 (45.3%) of the 1600 women aged 60–69 years; and in 519 (53.1%) of the 978 women aged 70 years or older. The gradual nature of the involution process is apparent in that it is already present at least to a partial degree in more than half of the women younger than 40 years and is still ongoing in women older than 70 years.
. | Extent of lobular involution, No. (%) . | . | . | . | ||
---|---|---|---|---|---|---|
Characteristic . | None . | Partial . | Complete . | P value † . | ||
Overall | 1627 (18.6) | 5197 (59.5) | 1912 (21.9) | |||
Age at BBD, y | <.001 | |||||
18–29 | 308 (54.4) | 239 (42.2) | 19 (3.4) | |||
30–39 | 417 (40.2) | 567 (54.7) | 53 (5.1) | |||
40–49 | 643 (26.3) | 1661 (67.9) | 142 (5.8) | |||
50–59 | 218 (10.3) | 1436 (68.1) | 455 (21.6) | |||
60–69 | 29 (1.8) | 847 (52.9) | 724 (45.3) | |||
≥70 | 12 (1.2) | 447 (45.7) | 519 (53.1) | |||
Family history of breast cancer ‡ | <.001 | |||||
None or weak | 796 (21.1) | 2717 (56.4) | 911 (22.4) | |||
Strong | 223 (24.5) | 566 (55.9) | 139 (19.5) | |||
Parity § | <.001 | |||||
Nulliparous | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
Parous | 893 (22.6) | 2780 (56.2) | 836 (21.2) | |||
No. of children § | <.001 | |||||
0 | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
1 | 75 (17.2) | 269 (54.7) | 131 (28.0) | |||
2 | 324 (22.2) | 891 (56.2) | 258 (21.5) | |||
3 | 257 (23.9) | 742 (55.8) | 205 (20.4) | |||
≥4 | 235 (26.6) | 871 (55.3) | 242 (18.0) | |||
Children breastfed ‖ | .428 | |||||
No | 464 (21.6) | 1555 (57.4) | 436 (21.1) | |||
Yes | 431 (23.3) | 1202 (55.9) | 364 (20.8) | |||
Hormone replacement therapy ¶ | .016 | |||||
Ever | 481 (22.5) | 1330 (57.1) | 494 (20.3) | |||
Never | 516 (21.8) | 1837 (55.4) | 458 (22.9) |
. | Extent of lobular involution, No. (%) . | . | . | . | ||
---|---|---|---|---|---|---|
Characteristic . | None . | Partial . | Complete . | P value † . | ||
Overall | 1627 (18.6) | 5197 (59.5) | 1912 (21.9) | |||
Age at BBD, y | <.001 | |||||
18–29 | 308 (54.4) | 239 (42.2) | 19 (3.4) | |||
30–39 | 417 (40.2) | 567 (54.7) | 53 (5.1) | |||
40–49 | 643 (26.3) | 1661 (67.9) | 142 (5.8) | |||
50–59 | 218 (10.3) | 1436 (68.1) | 455 (21.6) | |||
60–69 | 29 (1.8) | 847 (52.9) | 724 (45.3) | |||
≥70 | 12 (1.2) | 447 (45.7) | 519 (53.1) | |||
Family history of breast cancer ‡ | <.001 | |||||
None or weak | 796 (21.1) | 2717 (56.4) | 911 (22.4) | |||
Strong | 223 (24.5) | 566 (55.9) | 139 (19.5) | |||
Parity § | <.001 | |||||
Nulliparous | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
Parous | 893 (22.6) | 2780 (56.2) | 836 (21.2) | |||
No. of children § | <.001 | |||||
0 | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
1 | 75 (17.2) | 269 (54.7) | 131 (28.0) | |||
2 | 324 (22.2) | 891 (56.2) | 258 (21.5) | |||
3 | 257 (23.9) | 742 (55.8) | 205 (20.4) | |||
≥4 | 235 (26.6) | 871 (55.3) | 242 (18.0) | |||
Children breastfed ‖ | .428 | |||||
No | 464 (21.6) | 1555 (57.4) | 436 (21.1) | |||
Yes | 431 (23.3) | 1202 (55.9) | 364 (20.8) | |||
Hormone replacement therapy ¶ | .016 | |||||
Ever | 481 (22.5) | 1330 (57.1) | 494 (20.3) | |||
Never | 516 (21.8) | 1837 (55.4) | 458 (22.9) |
For age at diagnosis of benign breast disease (BBD), percentage values were unadjusted. For all other variables, percentage values were adjusted for age.
For age at BBD, P values were calculated using chi-square tests of statistical significance. For all other variables, P values were calculated by use of multicategorical logistic regression analyses, accounting for the effects of age. All statistical tests were two-sided.
Information on family history was available for 5352 of 8376 women.
Information on parity was available for 5220 of 8736 women. Specific number of children was available for 4500 of the 4509 parous women.
Information on breastfeeding was available for 4452 of 8736 women.
Information on hormone replacement therapy was available for 5116 of 8736 women.
. | Extent of lobular involution, No. (%) . | . | . | . | ||
---|---|---|---|---|---|---|
Characteristic . | None . | Partial . | Complete . | P value † . | ||
Overall | 1627 (18.6) | 5197 (59.5) | 1912 (21.9) | |||
Age at BBD, y | <.001 | |||||
18–29 | 308 (54.4) | 239 (42.2) | 19 (3.4) | |||
30–39 | 417 (40.2) | 567 (54.7) | 53 (5.1) | |||
40–49 | 643 (26.3) | 1661 (67.9) | 142 (5.8) | |||
50–59 | 218 (10.3) | 1436 (68.1) | 455 (21.6) | |||
60–69 | 29 (1.8) | 847 (52.9) | 724 (45.3) | |||
≥70 | 12 (1.2) | 447 (45.7) | 519 (53.1) | |||
Family history of breast cancer ‡ | <.001 | |||||
None or weak | 796 (21.1) | 2717 (56.4) | 911 (22.4) | |||
Strong | 223 (24.5) | 566 (55.9) | 139 (19.5) | |||
Parity § | <.001 | |||||
Nulliparous | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
Parous | 893 (22.6) | 2780 (56.2) | 836 (21.2) | |||
No. of children § | <.001 | |||||
0 | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
1 | 75 (17.2) | 269 (54.7) | 131 (28.0) | |||
2 | 324 (22.2) | 891 (56.2) | 258 (21.5) | |||
3 | 257 (23.9) | 742 (55.8) | 205 (20.4) | |||
≥4 | 235 (26.6) | 871 (55.3) | 242 (18.0) | |||
Children breastfed ‖ | .428 | |||||
No | 464 (21.6) | 1555 (57.4) | 436 (21.1) | |||
Yes | 431 (23.3) | 1202 (55.9) | 364 (20.8) | |||
Hormone replacement therapy ¶ | .016 | |||||
Ever | 481 (22.5) | 1330 (57.1) | 494 (20.3) | |||
Never | 516 (21.8) | 1837 (55.4) | 458 (22.9) |
. | Extent of lobular involution, No. (%) . | . | . | . | ||
---|---|---|---|---|---|---|
Characteristic . | None . | Partial . | Complete . | P value † . | ||
Overall | 1627 (18.6) | 5197 (59.5) | 1912 (21.9) | |||
Age at BBD, y | <.001 | |||||
18–29 | 308 (54.4) | 239 (42.2) | 19 (3.4) | |||
30–39 | 417 (40.2) | 567 (54.7) | 53 (5.1) | |||
40–49 | 643 (26.3) | 1661 (67.9) | 142 (5.8) | |||
50–59 | 218 (10.3) | 1436 (68.1) | 455 (21.6) | |||
60–69 | 29 (1.8) | 847 (52.9) | 724 (45.3) | |||
≥70 | 12 (1.2) | 447 (45.7) | 519 (53.1) | |||
Family history of breast cancer ‡ | <.001 | |||||
None or weak | 796 (21.1) | 2717 (56.4) | 911 (22.4) | |||
Strong | 223 (24.5) | 566 (55.9) | 139 (19.5) | |||
Parity § | <.001 | |||||
Nulliparous | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
Parous | 893 (22.6) | 2780 (56.2) | 836 (21.2) | |||
No. of children § | <.001 | |||||
0 | 113 (17.6) | 421 (55.3) | 177 (27.1) | |||
1 | 75 (17.2) | 269 (54.7) | 131 (28.0) | |||
2 | 324 (22.2) | 891 (56.2) | 258 (21.5) | |||
3 | 257 (23.9) | 742 (55.8) | 205 (20.4) | |||
≥4 | 235 (26.6) | 871 (55.3) | 242 (18.0) | |||
Children breastfed ‖ | .428 | |||||
No | 464 (21.6) | 1555 (57.4) | 436 (21.1) | |||
Yes | 431 (23.3) | 1202 (55.9) | 364 (20.8) | |||
Hormone replacement therapy ¶ | .016 | |||||
Ever | 481 (22.5) | 1330 (57.1) | 494 (20.3) | |||
Never | 516 (21.8) | 1837 (55.4) | 458 (22.9) |
For age at diagnosis of benign breast disease (BBD), percentage values were unadjusted. For all other variables, percentage values were adjusted for age.
For age at BBD, P values were calculated using chi-square tests of statistical significance. For all other variables, P values were calculated by use of multicategorical logistic regression analyses, accounting for the effects of age. All statistical tests were two-sided.
Information on family history was available for 5352 of 8376 women.
Information on parity was available for 5220 of 8736 women. Specific number of children was available for 4500 of the 4509 parous women.
Information on breastfeeding was available for 4452 of 8736 women.
Information on hormone replacement therapy was available for 5116 of 8736 women.
We also found a strong, inverse association ( P <.001) between lobular involution and parity ( Table 1 ). Specifically, the likelihood of complete involution was 27.1% (95% CI = 24.1% to 30.1%) in nulliparous women, 28.0% (95% CI = 24.7% to 31.4%) in women who had one child, 21.5% (95% CI = 19.3% to 23.8%) in women who had two children, 20.4% (95% CI = 17.8% to 23.0%) in women who had three children, and 18.0% (95% CI = 16.1% to 20.0%) in women who had four children or more.
Separating women into categories of ever versus never breastfeeding did not reveal any relationship with extent of lobular involution ( P = .428). Women who reported having used hormone replacement therapy were slightly less likely to have complete involution (20.3%) than those with no history of hormone replacement therapy use (22.9%) ( P = .016). Breast tissue from women with a strong family history of breast cancer was less likely than that from women with no or a weak family history of breast cancer to demonstrate lobular involution; i.e., after adjustment for age, more women with a strong family history had no involution (24.5%) than those with no or a weak family history (21.1%), and fewer women with a strong family history had complete involution (19.5%) than those with no or a weak family history (22.4%) (logistic regression analysis comparing distribution of involution across levels of family history, P <.001).
We also examined the extent of lobular involution by category of benign breast disease. Among women with nonproliferative disease, 27.2% had complete involution. However, among women with proliferative disease without or with atypia, only 11.5% had complete lobular involution ( P <.001).
Lobular Involution and Breast Cancer Risk
This cohort of women with benign breast disease was, overall, at increased risk of breast cancer when compared with age-matched women in the general population. Specifically, the relative risk for the entire cohort of 8736 women, compared with the Iowa SEER population, was 1.40 (95% CI = 1.30 to 1.51). In our cohort, degree of involution was associated inversely with breast cancer risk ( Table 2 ; e.g., for no involution, RR = 1.88, 95% CI = 1.59 to 2.21; for partial involution, RR = 1.47, 95% CI = 1.33 to 1.61; and for complete involution, RR = 0.91, 95% CI = 0.75 to 1.10; test for heterogeneity P <.001).
Characteristic . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
Degree of involution | |||||
None (0%) | 1627 | 32 271 | 150 | 79.6 | 1.88 (1.59 to 2.21) |
Partial (1%–74%) | 5197 | 90 409 | 440 | 300.1 | 1.47 (1.33 to 1.61) |
Complete (≥75%) | 1912 | 28 376 | 106 | 116.5 | 0.91 (0.75 to 1.10) |
Histologic type | |||||
Nonproliferative | 5736 | 101 201 | 355 | 321.5 | 1.10 (0.99 to 1.23) |
Proliferative without atypia | 2677 | 45 418 | 276 | 158.1 | 1.75 (1.55 to 1.96) |
Proliferative with atypia | 323 | 4436 | 65 | 16.6 | 3.91 (3.02 to 4.98) |
Age at biopsy, y | |||||
<45 | 2682 | 52 055 | 158 | 108.4 | 1.46 (1.24 to 1.70) |
45–55 | 2559 | 49 246 | 254 | 169.0 | 1.50 (1.32 to 1.70) |
≥55 | 3495 | 49 754 | 284 | 218.8 | 1.30 (1.15 to 1.46) |
Family history of breast cancer ‡ | |||||
None or weak | 4424 | 81 514 | 329 | 269.4 | 1.22 (1.09 to 1.36) |
Strong | 928 | 18 385 | 115 | 59.5 | 1.93 (1.59 to 2.32) |
Age at birth of first live child, y § | |||||
Nulliparous | 711 | 13 021 | 71 | 41.1 | 1.73 (1.35 to 2.18) |
<30 | 4121 | 77 710 | 327 | 257.9 | 1.27 (1.13 to 1.41) |
≥30 | 388 | 7091 | 30 | 22.4 | 1.34 (0.90 to 1.92) |
Characteristic . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
Degree of involution | |||||
None (0%) | 1627 | 32 271 | 150 | 79.6 | 1.88 (1.59 to 2.21) |
Partial (1%–74%) | 5197 | 90 409 | 440 | 300.1 | 1.47 (1.33 to 1.61) |
Complete (≥75%) | 1912 | 28 376 | 106 | 116.5 | 0.91 (0.75 to 1.10) |
Histologic type | |||||
Nonproliferative | 5736 | 101 201 | 355 | 321.5 | 1.10 (0.99 to 1.23) |
Proliferative without atypia | 2677 | 45 418 | 276 | 158.1 | 1.75 (1.55 to 1.96) |
Proliferative with atypia | 323 | 4436 | 65 | 16.6 | 3.91 (3.02 to 4.98) |
Age at biopsy, y | |||||
<45 | 2682 | 52 055 | 158 | 108.4 | 1.46 (1.24 to 1.70) |
45–55 | 2559 | 49 246 | 254 | 169.0 | 1.50 (1.32 to 1.70) |
≥55 | 3495 | 49 754 | 284 | 218.8 | 1.30 (1.15 to 1.46) |
Family history of breast cancer ‡ | |||||
None or weak | 4424 | 81 514 | 329 | 269.4 | 1.22 (1.09 to 1.36) |
Strong | 928 | 18 385 | 115 | 59.5 | 1.93 (1.59 to 2.32) |
Age at birth of first live child, y § | |||||
Nulliparous | 711 | 13 021 | 71 | 41.1 | 1.73 (1.35 to 2.18) |
<30 | 4121 | 77 710 | 327 | 257.9 | 1.27 (1.13 to 1.41) |
≥30 | 388 | 7091 | 30 | 22.4 | 1.34 (0.90 to 1.92) |
Number of events expected on the basis of Iowa Surveillance, Epidemiology, and End Results breast cancer incidence data.
All analyses account for the effects of age and calendar period. CI = confidence interval.
Information on family history of breast cancer was available for 5352 of 8736 women.
Information on parity was available for 5220 of 8736 women.
Characteristic . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
Degree of involution | |||||
None (0%) | 1627 | 32 271 | 150 | 79.6 | 1.88 (1.59 to 2.21) |
Partial (1%–74%) | 5197 | 90 409 | 440 | 300.1 | 1.47 (1.33 to 1.61) |
Complete (≥75%) | 1912 | 28 376 | 106 | 116.5 | 0.91 (0.75 to 1.10) |
Histologic type | |||||
Nonproliferative | 5736 | 101 201 | 355 | 321.5 | 1.10 (0.99 to 1.23) |
Proliferative without atypia | 2677 | 45 418 | 276 | 158.1 | 1.75 (1.55 to 1.96) |
Proliferative with atypia | 323 | 4436 | 65 | 16.6 | 3.91 (3.02 to 4.98) |
Age at biopsy, y | |||||
<45 | 2682 | 52 055 | 158 | 108.4 | 1.46 (1.24 to 1.70) |
45–55 | 2559 | 49 246 | 254 | 169.0 | 1.50 (1.32 to 1.70) |
≥55 | 3495 | 49 754 | 284 | 218.8 | 1.30 (1.15 to 1.46) |
Family history of breast cancer ‡ | |||||
None or weak | 4424 | 81 514 | 329 | 269.4 | 1.22 (1.09 to 1.36) |
Strong | 928 | 18 385 | 115 | 59.5 | 1.93 (1.59 to 2.32) |
Age at birth of first live child, y § | |||||
Nulliparous | 711 | 13 021 | 71 | 41.1 | 1.73 (1.35 to 2.18) |
<30 | 4121 | 77 710 | 327 | 257.9 | 1.27 (1.13 to 1.41) |
≥30 | 388 | 7091 | 30 | 22.4 | 1.34 (0.90 to 1.92) |
Characteristic . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
Degree of involution | |||||
None (0%) | 1627 | 32 271 | 150 | 79.6 | 1.88 (1.59 to 2.21) |
Partial (1%–74%) | 5197 | 90 409 | 440 | 300.1 | 1.47 (1.33 to 1.61) |
Complete (≥75%) | 1912 | 28 376 | 106 | 116.5 | 0.91 (0.75 to 1.10) |
Histologic type | |||||
Nonproliferative | 5736 | 101 201 | 355 | 321.5 | 1.10 (0.99 to 1.23) |
Proliferative without atypia | 2677 | 45 418 | 276 | 158.1 | 1.75 (1.55 to 1.96) |
Proliferative with atypia | 323 | 4436 | 65 | 16.6 | 3.91 (3.02 to 4.98) |
Age at biopsy, y | |||||
<45 | 2682 | 52 055 | 158 | 108.4 | 1.46 (1.24 to 1.70) |
45–55 | 2559 | 49 246 | 254 | 169.0 | 1.50 (1.32 to 1.70) |
≥55 | 3495 | 49 754 | 284 | 218.8 | 1.30 (1.15 to 1.46) |
Family history of breast cancer ‡ | |||||
None or weak | 4424 | 81 514 | 329 | 269.4 | 1.22 (1.09 to 1.36) |
Strong | 928 | 18 385 | 115 | 59.5 | 1.93 (1.59 to 2.32) |
Age at birth of first live child, y § | |||||
Nulliparous | 711 | 13 021 | 71 | 41.1 | 1.73 (1.35 to 2.18) |
<30 | 4121 | 77 710 | 327 | 257.9 | 1.27 (1.13 to 1.41) |
≥30 | 388 | 7091 | 30 | 22.4 | 1.34 (0.90 to 1.92) |
Number of events expected on the basis of Iowa Surveillance, Epidemiology, and End Results breast cancer incidence data.
All analyses account for the effects of age and calendar period. CI = confidence interval.
Information on family history of breast cancer was available for 5352 of 8736 women.
Information on parity was available for 5220 of 8736 women.
Figure 2 illustrates the observed associations between the extent of involution and breast cancer risk among strata of age, histology, family history, and parity. Extent of lobular involution modified age-related breast cancer risk (e.g., for a woman older than 55 years with no involution, RR = 3.21, 95% CI = 1.90 to 5.08, and for a similar woman with complete lobular involution, RR = 0.92, 95% CI = 0.74 to 1.14). The same pattern was observed in all age groups.
Similarly, progressive increases in lobular involution in background breast tissue was associated with reduced risk of breast cancer among women with benign proliferative disease, even those with atypia ( Fig. 2 ). Among women with atypia, no involution was more strongly associated with a higher risk of breast cancer (RR = 7.79, 95% CI = 3.56 to 14.81) than complete involution (RR = 1.49, 95% CI = 0.41 to 3.82) or partial involution (RR = 4.06, 95% CI = 3.03 to 5.33) (test for heterogeneity, P = .003). Among women with proliferative disease without atypia, no involution was also associated with a higher risk of breast cancer (RR = 2.94, 95% CI = 2.26 to 3.75) than complete involution (RR = 1.11, 95% CI = 0.68 to 1.71). The same pattern held true for those with nonproliferative breast disease; i.e., those with no involution had a higher risk than those with complete or partial involution.
Lobular involution modified the influence of family history on risk of breast cancer ( Fig. 2 ). Among women with a strong family history of breast cancer, no involution was associated with the highest risk of breast cancer (RR = 2.77, 95% CI = 1.94 to 3.84), followed by partial involution (RR = 1.72, 95% CI = 1.32 to 2.20) and then by complete involution (RR = 1.61, 95% CI = 0.92 to 2.61). Among women with no or a weak family history of breast cancer and complete involution, the risk of breast cancer (RR = 0.59, 95% CI = 0.41 to 0.81) was approximately half of that for the general population, which was based on Iowa SEER data, and approximately fivefold less than the risk of those with strong family history and no involution (i.e., RR = 2.77, 95% CI = 1.94 to 3.84).
Lobular involution also modified the risk associated with parity or age at birth of the first live child. Among nulliparous women and women whose first live child was born when she was at least 30 years old, no lobular involution was associated with increased risks of breast cancer (RR = 2.41, 95% CI = 1.25 to 4.21, and RR = 2.74, 95% CI = 1.31 to 5.03, respectively). However, among these same two groups of women, when complete lobular involution had occurred, there was no increase in risk (RR = 1.02, 95% CI = 0.53 to 1.78, and RR = 0.48, 95% CI = 0.10 to 1.40, respectively). Among women whose first live child was born when she was younger than 30 years old, complete lobular involution was associated with a decreased risk of breast cancer (RR = 0.65, 95% CI = 0.44 to 0.91).
We also asked whether the era in which the biopsy examination was performed—namely, before or after widespread adoption of mammography—affected the results. In the first 15 years of the cohort (from 1967 through 1981), 78% of biopsy examinations were done because of a palpable concern (i.e., a palpable lump detected during a clinical breast examination or by the patient), and 22% were done because of an abnormal mammogram. From 1982 through 1991, 40% of the biopsy examinations were done because of a palpable concern, and 60% were done because of an abnormal mammogram. The relative risks of breast cancer by involution status and by dates ( Table 3 ) indicated that associations between extent of involution and risk were similar in the pre- and postmammography time periods.
Extent of involution and year of biopsy . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
No involution | |||||
1967–1981 | 957 | 21 886 | 105 | 56.9 | 1.85 (1.51 to 2.23) |
1982–1991 | 670 | 10 384 | 45 | 22.7 | 1.98 (1.44 to 2.65) |
Partial involution | |||||
1967–1981 | 2381 | 49 080 | 204 | 157.3 | 1.53 (1.34 to 1.73) |
1982–1991 | 2816 | 41 329 | 200 | 142.8 | 1.40 (1.21 to 1.61) |
Complete involution | |||||
1967–1981 | 740 | 12 524 | 55 | 46.8 | 1.18 (0.89 to 1.53) |
1982–1991 | 1172 | 15 851 | 51 | 69.7 | 0.73 (0.54 to 0.96) |
Extent of involution and year of biopsy . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
No involution | |||||
1967–1981 | 957 | 21 886 | 105 | 56.9 | 1.85 (1.51 to 2.23) |
1982–1991 | 670 | 10 384 | 45 | 22.7 | 1.98 (1.44 to 2.65) |
Partial involution | |||||
1967–1981 | 2381 | 49 080 | 204 | 157.3 | 1.53 (1.34 to 1.73) |
1982–1991 | 2816 | 41 329 | 200 | 142.8 | 1.40 (1.21 to 1.61) |
Complete involution | |||||
1967–1981 | 740 | 12 524 | 55 | 46.8 | 1.18 (0.89 to 1.53) |
1982–1991 | 1172 | 15 851 | 51 | 69.7 | 0.73 (0.54 to 0.96) |
Number of events expected on the basis of Iowa Surveillance, Epidemiology, and End Results breast cancer incidence data.
All analyses account for the effects of age and calendar period. CI = confidence interval.
Extent of involution and year of biopsy . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
No involution | |||||
1967–1981 | 957 | 21 886 | 105 | 56.9 | 1.85 (1.51 to 2.23) |
1982–1991 | 670 | 10 384 | 45 | 22.7 | 1.98 (1.44 to 2.65) |
Partial involution | |||||
1967–1981 | 2381 | 49 080 | 204 | 157.3 | 1.53 (1.34 to 1.73) |
1982–1991 | 2816 | 41 329 | 200 | 142.8 | 1.40 (1.21 to 1.61) |
Complete involution | |||||
1967–1981 | 740 | 12 524 | 55 | 46.8 | 1.18 (0.89 to 1.53) |
1982–1991 | 1172 | 15 851 | 51 | 69.7 | 0.73 (0.54 to 0.96) |
Extent of involution and year of biopsy . | No. of women . | No. of person-years . | No. of observed events . | No. of expected events * . | Relative risk (95% CI) † . |
---|---|---|---|---|---|
No involution | |||||
1967–1981 | 957 | 21 886 | 105 | 56.9 | 1.85 (1.51 to 2.23) |
1982–1991 | 670 | 10 384 | 45 | 22.7 | 1.98 (1.44 to 2.65) |
Partial involution | |||||
1967–1981 | 2381 | 49 080 | 204 | 157.3 | 1.53 (1.34 to 1.73) |
1982–1991 | 2816 | 41 329 | 200 | 142.8 | 1.40 (1.21 to 1.61) |
Complete involution | |||||
1967–1981 | 740 | 12 524 | 55 | 46.8 | 1.18 (0.89 to 1.53) |
1982–1991 | 1172 | 15 851 | 51 | 69.7 | 0.73 (0.54 to 0.96) |
Number of events expected on the basis of Iowa Surveillance, Epidemiology, and End Results breast cancer incidence data.
All analyses account for the effects of age and calendar period. CI = confidence interval.
Lobular Involution: Localized or Field Effect
To address whether or not the degree of involution was relevant only to the area of the biopsy or was representative of the field of breast tissue, we examined 1) whether, for women with bilateral benign biopsy examination results, involution results were concordant and 2) whether the degree of involution at the benign biopsy site was associated with the risk of ipsilateral breast cancer or with both ipsi- and contralateral breast cancers. A subset of 245 women had bilateral biopsy examinations performed at the same time. In 203 (83%) of these women, the same category of involution (no, partial, or complete) was found in the biopsy tissue from both breasts. In 41 (17%), there was a difference of one category between the two breasts. Only one individual had complete involution in the biopsy tissue of one breast and no involution in the contralateral sample. These results indicate a high level of agreement in involution measured across multiple biopsy specimens within a woman (kappa coefficient = 0.72, 95% CI = 0.64 to 0.80; test for agreement beyond that expected by chance P <.001).
We next investigated the extent of involution and the laterality of subsequent breast cancers. In our cohort overall, there is a slight predominance of ipsilateral breast cancers (55.5%) over contralateral breast cancers (44.5%), as reported previously ( 8 ) ; this result is thought to reflect the presence of some direct precursors among these lesions. To determine whether involution at the site of the benign breast disease was relevant to the contralateral breast, we examined the ratio of ipsilateral to contralateral events by degree of involution. With no involution, the ratio was 53.6% ipsilateral to 46.4% contralateral; for partial involution, the ratio was 55.9% to 44.1%; and for complete involution, the ratio was 53.5% to 46.5% (chi-square test for difference in percent ipsilateral across involution status, P = .85). Thus, the relationship between involution extent and breast cancer risk was observed in both the ipsi- and contralateral breast.
D ISCUSSION
We characterized the degree of lobular involution in the background breast tissue in a large cohort of women with benign breast disease. Our data demonstrate a strong, inverse relationship between degree of involution and breast cancer risk. To our knowledge, this is the first study to systematically examine age-related involution in the context of breast cancer risk. Furthermore, greater degrees of involution reduced breast cancer risk even in high-risk subsets defined by age, atypia, reproductive history, or family history. There was a strong direct association between involution and increasing age. There was an inverse association between involution and parity.
As in this study, others have found that older women tend to have fewer lobules or only lobule remnants ( 4 , 5 ) . Cowan and Herbert ( 4 ) performed a detailed autopsy study of the breast tissue from 102 women, aged 50–104 years, who died without known breast disease. Although considerable individual variability was present, they described a progressive loss of lobules with increasing age. Earlier reports state that age-related involution has already begun in women under the age of 40 years ( 1 , 5 ) . Our data confirm that this process is present, at least to a partial degree, in many younger women.
We hypothesize that the degree of involution detected at the benign biopsy site reflects that of the overall field of a woman's breast tissue. We believe that this hypothesis is reasonable because of our results showing a similar likelihood of contralateral and ipsilateral breast cancers by involution status at the site of the benign breast disease and because of the high concordance in involution status in women who had bilateral biopsy examinations. However, our study design cannot answer this question definitively. To do so would require examination of the extent of involution throughout all of a woman's breast tissue.
It is widely appreciated that, as women age, their risk of breast cancer increases. But the rate of increase of breast cancer slows appreciably at approximately age 50 years ( 10 , 11 ) , which has been attributed to a reduction in ovarian hormonal production ( 12 ) . We observed a definite increase in the process of involution at approximately age 50 years (with complete involution present in 5.8% of women aged 40–49 years and in 21.6% of women aged 50–59 years). These data raise the possibility that involution may be contributing to the slowing in the rate of increase of breast cancer among women older than 50 years, as speculated by Henson and Tarone ( 7 ) .
We examined various factors besides age for their association with degree of involution. We found an inverse association between lobular involution and parity. Others have also reported that the more children a woman has, the more likely she is to have persistence of lobular structures ( 1 , 5 ) , which we found was associated with increased risk of breast cancer. Yet, multiparity is generally considered to reduce the risk of breast cancer ( 13 , 14 ) . Several factors may explain this apparent contradiction. First, we do not have data on the age at each child's birth for the women in our cohort. Some epidemiologic work has suggested that full-term pregnancies after 35 years of age are associated with an increased risk of breast cancer ( 7 , 15 ) . Thus, data on a woman's age at each pregnancy and on her age at breast biopsy examination would help to evaluate more definitively the relationships of parity, involution, and breast cancer risk. In addition, our study was limited by the relatively large size of the group of women categorized as having partial involution. More specific, quantitative measures of degree of involution should be explored to determine whether the association between parity status and degree of involution can be defined more precisely. Given the inverse association between complete involution and multiparity and given that both are associated with reduced risk of breast cancer, we hypothesize that the breast cancer risk modification associated with parity is independent of involution status.
There are several biologic mechanisms by which involution or lack thereof could alter a patient's breast cancer risk. The decrement in epithelial cell number that accompanies involution may decrease breast cancer risk simply because there are fewer epithelial cells to undergo malignant transformation. Another possibility is that aberrant involution may be a marker or phenotype reflecting underlying constitutional susceptibility for breast cancer that is present in the epithelial or stromal compartment or in their relationship with each other. Yet another possibility is that failure to undergo timely or appropriate involution allows prolonged exposure of epithelial cells to intrinsic and/or extrinsic mutagenic stresses ( 16 – 20 ) . In this model, the prime targets of such mutagenic processes, such as stem cells or early progenitors, may become quiescent during the process of involution. Experiments to characterize the epithelial and stromal mediators present in tissue with and without involution, in women with and without subsequent cancer, should help to clarify the mechanism of risk reduction.
For our work to date, we divided extent of involution into three categories. We recognized that, although the morphologic patterns of age-related lobular involution have been defined ( 1 , 4 , 5 ) , no histologic standard exists for evaluating the extent of breast involution. In particular, there is no well-characterized method for grading partial degrees of involution. For this reason, we attempted to classify degree of involution with the least amount of subjectivity. Thus, by deciding only whether breast tissue had no lobular involution versus almost complete involution and then by combining the remainder into one category of partial involution, we minimized the subjectivity inherent in judging percent involution.
Our study has several limitations. First, these findings do not necessarily pertain to all women because the cohort studied included women who had a breast biopsy because of some concern. Nevertheless, women with benign breast disease make up a large population who are understandably concerned about their breast cancer risk (estimated at 1 million US women each year) ( 21 – 23 ) . Another limitation lies in our current very broad category of partial involution. This category encompasses a wide range of involution extent (1%–74% of lobules involuted). We expect that more specific gradations would support more refined association studies. Finally, we did not have complete risk factor data for all the women in the cohort, largely because the women with biopsy examinations in the earlier years of the cohort are now elderly or deceased. Fortunately, for purposes of this report, we did not have to depend on the questionnaire for involution status or for cancer outcomes (which were available from our comprehensive Mayo medical record). We had completed questionnaires for 63.8% of the patients with breast cancer and 61.6% of the patients without breast cancer in the cohort.
There are other approaches to the study of involution and breast cancer risk. Henson and Tarone ( 7 ) suggested an autopsy case–control series to look at involution as a possible risk factor for breast cancer. Although this approach would provide access to extensive amounts of breast tissue, the availability of clinical risk factor information and of a sufficient number of subjects could be limiting. Other women who have breast tissue removed in the course of clinical care are those who have reduction mammoplasty or prophylactic mastectomy. These women are, respectively, those who have breast hypertrophy or a hereditary predisposition to breast cancer. Although involution (or lack thereof) in these women is of considerable interest, their tissue is not necessarily representative of the general population.
The mechanisms controlling age-related involution are of considerable interest. Molecular programs that control postlactational involution in rodents have been studied extensively ( 24 ) . With postlactational involution, there is dramatic reversal of the developmental changes wrought by pregnancy. Specifically, there is widespread apoptosis of alveolar epithelial cells followed by removal of apoptotic debris and remodeling of the stroma and extracellular matrix ( 24 ) . These events occur within a matter of days of abrupt weaning and restore the gland to its prepregnancy state. In contrast, the molecular orchestration of age-related involution, to our knowledge, has not yet been characterized.
In the past, for women with benign breast biopsy results, the type and extent of epithelial proliferation present in their biopsy has been the principal way to stratify their risk. Results of our study indicate that assessing the status of lobular involution in the biopsied tissue may ultimately add to risk prediction capabilities. It is notable, as shown in Fig. 2, C , that some of the most extreme risk estimates are observed in women whose involution status is unusual for their age—namely, young women with complete involution (RR = 0.43, 95% CI = 0.13 to 1.55) and women older than 55 years with no involution (RR = 3.2, 95% CI = 1.90 to 5.08). It is tempting to speculate that the process of complete involution may be protective and, conversely, that lack of involution identifies higher risk groups. However, confidence intervals were wide around the estimates for these less common categories.
In summary, we have evaluated the extent and effect of age-related lobular involution in a cohort of approximately 9000 women who had a benign breast biopsy examination. We observed a statistically significant reduction in risk of breast cancer among those women whose breast tissue had undergone extensive lobular involution, which was apparently independent of other markers of risk. Among women with benign breast disease, assessment of extent of involution may help to fine-tune current risk prediction approaches. Elucidation of the mechanism of lobular involution may reveal ways to promote the process as a means of risk reduction.
Supported by a Department of Defense Center of Excellence Grant (FEDDAMD17-02-1-0473-1), a grant (R01 CA46332) from the National Institutes of Health, and grants from the Susan G. Komen Breast Cancer Foundation (BCTR 99-3152), Andersen Foundation, Breast Cancer Research Foundation, and Regis Foundation for Breast Cancer Research. The authors take full responsibility for the study design, collection of the data, analysis and interpretation of the data, the decision to submit the manuscript, and the writing of the manuscript.
We are indebted to Joel Worra and Dr Piet de Groen for database development; to Wilma Lingle, PhD, and the Biospecimens Core for tissue processing; to Teresa Allers, Mary Amundsen, Mary Campion, Joanne Johnson, Melanie Kasner, Margie Loprinzi, and Lois Penheiter for data collection; to Ann Harris and the Survey Research Center for patient follow-up; and to Vicki Shea for help in preparing the manuscript.
Funding to pay the Open Access publication charges for this article was provided by a grant from the Department of Defense.
References
Vorrherr H. The breast: morphology, physiology, and lactation. New York (NY): Academic Press;
Hutson SW, Cowen PN, Bird CC. Morphometric studies of age related changes in normal human breast and their significance for evolution of mammary cancer.
Cowan DF, Herbert TA. Involution of the breast in women aged 50 to 104 years: a histopathological study of 102 cases.
Henson DE, Tarone RE. On the possible role of involution in the natural history of breast cancer.
Henson DE, Tarone RE. Involution and the etiology of breast cancer.
Hartmann LC, Sellers TA, Frost MH, Lingle WL, Degnim AM, Ghosh K, et al. Benign breast disease and the risk of breast cancer.
Hosmer D, Lemeshow S. Applied logistic regression. New York (NY): John Wiley & Sons;
Clemmesen J. The Danish Cancer Registry. Problems and results.
Cutler SY, Young JL. Third National Cancer Survey: incidence data.
Pike MC, Pearce CL, Wu AH. Prevention of cancers of the breast, endometrium, and ovary.
Kelsey JL, Gammon MD, John EM. Reproductive factors and breast cancer.
Ursin G, Bernstein L, Lord SJ, Karim R, Deapen D, Press MF, et al. Reproductive factors and subtypes of breast cancer defined by hormone receptor and histology.
Trichopoulos D, Hsieh CC, MacMahon B, Lin TM, Lowe CR, Mirra AP, et al. Age at any birth and breast cancer risk.
Russo J, Tay LK, Russo IH. Differentiation of the mammary gland and susceptibility to carcinogenesis.
Liu S, Dontu GA, Wicha MS. Mammary stem cells, self-renewal pathways, and carcinogenesis.
Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. Prospective identification of tumorigenic breast cancer cells.
Tlsty TD, Crawford YG, Holst CR, Fordyce CA, Zhang J, McDermott K, et al. Genetic and epigenetic changes in mammary epithelial cells may mimic early events in carcinogenesis.
Russo J, Balogh GA, Chen J, Fernandez SV, Fernbaugh R, Heulings R, et al. The concept of stem cell in the mammary gland and its implication in morphogenesis, cancer, and prevention.
Kerlikowske K, Smith-Bindman R, Ljung B-M, Grady D. Evaluation of abnormal mammography results and palpable breast abnormalities.
Maxwell AJ, Pearson JM, Bishop HM. Crude open biopsy rates for benign screen detected lesions no longer reflect breast screening quality—time to change the standard.
American Cancer Society. “Mammograms and Other Breast Imaging Procedures.” Available at: http://www.cancer.org/docroot/CRI/content/CRI_2_6X_Mammography_and_other_Breast_Imaging_Procedures_5.asp . [Last accessed: October 5,