HHS Public Access

Author manuscript
Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
Author Manuscript

Published in final edited form as:

Breast Cancer Res Treat. 2015 September ; 153(2): 397–405. doi:10.1007/s10549-015-3535-8.

Complex Fibroadenoma and Breast Cancer Risk: A Mayo Clinic

Benign Breast Disease Cohort Studya
Dr. Aziza Nassar, MD, Dr. Daniel W. Visscher, MD, Dr. Amy C. Degnim, MD, Messr. Ryan D.
Frank, Messr. Robert A. Vierkant, Dr. Marlene Frost, PhD, Dr. Derek C. Radisky, PhD, Dr.
Celine M. Vachon, PhD, Ms Ruth A. Kraft, Dr. Lynn C. Hartmann, MD, and Dr. Karthik
Ghosh, MD
Author Manuscript

Department of Laboratory Medicine and Pathology (Dr Nassar) and Department of Cancer
Biology (Dr Radisky), Mayo Clinic, Jacksonville, Florida, and Division of Anatomic Pathology (Dr
Visscher), Department of Surgery (Dr Degnim), Division of Biomedical Statistics and Informatics
(Messr Frank and Vierkant), Division of Oncology Research (Dr Frost), Division of Epidemiology
(Dr Vachon), Division of Nephrology and Hypertension (Ms Kraft), Division of Medical Oncology
(Dr Hartmann), and Division of General Internal Medicine (Dr Ghosh), Mayo Clinic, Rochester,
Minnesota

Abstract
Purpose—To examine the breast cancer risk overall among women with simple fibroadenoma or
complex fibroadenoma and to examine the association of complex fibroadenoma with breast
cancer through stratification of other breast cancer risks.
Author Manuscript

Methods—The study included women aged 18 to 85 years from the Mayo Clinic Benign Breast
Disease Cohort who underwent excisional breast biopsy from 1967 through 1991. Within this
cohort, women who had fibroadenoma were compared to women who did not have fibroadenoma.
Breast cancer risk (observed vs expected) across fibroadenoma levels was assessed through
standardized incidence ratios (SIRs) by using age- and calendar-stratified incidence rates from the
Iowa Surveillance, Epidemiology, and End Results registry. Analyses were performed overall,
within subgroups of involution status, with other demographic characteristics (age, year of biopsy,
indication for biopsy, and family history), and with histologic characteristics, including overall
impression (nonproliferative disease, proliferative disease without atypia [PDWA], or atypical
hyperplasia).

Results—Fibroadenoma was identified in 2,136 women (noncomplex, 1,835 [85.9%]; complex,

Author Manuscript

301 [14.1%]). SIR for noncomplex fibroadenoma was 1.49 (95% CI, 1.26–1.74); for complex
fibroadenoma, it was 2.27 (95% CI, 1.63–3.10) (test for heterogeneity in SIR, P=.02). However,

aThe Abstract was previously published in poster form at the CTRC-AACR San Antonio Breast Cancer Symposium, San Antonio,
Texas, December 4–8, 2012.
Reprints: Aziza Nassar, MD, Department of Laboratory Medicine and Pathology, 4500 San Pablo Rd, Jacksonville, FL 32224
([email protected]).
Conflict of Interest
The authors declare that they have no conflicts of interest.
The funding sources had no involvement in the study design; in the collection, analysis, and interpretation of the data; in the writing of
the report; and in the decision to submit the manuscript for publication. The authors are responsible for the content. The content does
not necessarily represent the views of the National Institutes of Health.
 Nassar et al. Page 2

women with complex fibroadenoma were more likely to have other, concomitant high-risk
Author Manuscript

histologic characteristics (eg, incomplete involution and PDWA). In analyses stratified by

involution status and PDWA, complex fibroadenoma was not an independent risk marker for
breast cancer.

Conclusions—Complex fibroadenoma does not confer increased breast cancer risk beyond
other established histologic characteristics.

Keywords
benign breast disease; breast cancer; cancer risk; fibroadenoma; Mayo Clinic Benign Breast
Disease Cohort

Introduction
Author Manuscript

Fibroadenoma (FA) is a benign breast lesion that occurs most commonly in premenopausal
women, but it can occur in women at any age [1]. Several reports have described a higher
risk of subsequent breast carcinoma among patients who have FA, with relative risks (RRs)
ranging from 1.48 to 1.7 [1–5]. At the tissue level, FA is a benign tumor composed of both
epithelial and stromal components. Complex FAs are FAs with associated histologic
characteristics, including cysts (≥3 mm), sclerosing adenosis, epithelial calcifications, or
papillary apocrine metaplasia [2]. A prior report from Dupont et al [2] showed that the RR
of invasive breast cancer (BC) was 2.17 times higher for patients with FA than for matched
controls (95% CI, 1.5–3.2). The RR increased to 3.10 for women with complex FAs (95%
CI, 1.9–5.1) and remained elevated for more than 20 years after diagnosis [2]. Further
increases in risk were reported for patients with FA who had 1) benign proliferative disease
without atypia (PDWA) in the surrounding parenchyma adjacent to the FA (RR, 3.47; 95%
Author Manuscript

CI, 1.8–6.8), 2) atypical hyperplasia in the surrounding parenchyma (RR, 7.29; 95% CI, 2.2–
24), and 3) a family history of BC (RR, 3.72; 95% CI, 1.4–10).

Currently, FA is diagnosed when a patient presents with a palpable lump or has a breast
imaging procedure that shows an abnormality that is investigated with percutaneous biopsy.
Because FA is a benign finding, most patients with FA are subsequently followed by their
primary care providers. Hence, the risk association of FA with BC, and by type of FA
(complex or simple), should be clarified so that women with FA can be appropriately
counseled about future screening and risk reduction strategies.

In the present study, we investigated the BC risk associated with FA, both overall and
stratified into complex and simple FA. In addition, we evaluated the associations stratified
by individual histologic criteria (ie, cysts ≥3 mm, sclerosing adenosis, epithelial
Author Manuscript

calcifications, and papillary apocrine metaplasia) to clarify whether a specific feature of

complex FA was more closely associated with risk. Lastly, we evaluated the effect of
proliferative disease and atypical hyperplasia (within FA or adjacent to FA, or both) to
further stratify risk.

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 3

Methods
Author Manuscript

Study Population
The study was conducted with the Mayo Clinic Benign Breast Disease (BBD) Cohort [6,7].
The characteristics of the patients, the inclusion and exclusion criteria, and the assessment of
the pathologic specimens have been described extensively in earlier reports [6,7]. Briefly,
the study population consisted of 9,076 women aged 18 to 85 years who underwent surgical
excision of a benign breast lesion at Mayo Clinic in Rochester, Minnesota, from January 1,
1967, through December 31, 1991. Details about family history of BC and subsequent BC
events were obtained from medical records and questionnaires sent to patients and next of
kin. Family history of BC was classified as negative (if there was no family history of BC);
strong (if there was ≥1 first-degree relative with BC before age 50 years or ≥2 relatives with
BC, with ≥1 being a first-degree relative); or weak (for any family history not classified as
Author Manuscript

strong).

Histologic Examination
If the woman had more than 1 biopsy during the study period, we used findings from the
earliest biopsy performed. For each participant, stored hematoxylin-eosin–stained sections
were evaluated by an experienced breast pathologist (D.W.V.). The biopsy findings were
classified according to the criteria proposed by Page and Anderson [8] as nonproliferative
disease, PDWA, or atypical hyperplasia (AH) (including atypical ductal hyperplasia or
atypical lobular hyperplasia, or both). PDWA criteria were ductal hyperplasia (more than
mild), papillomas, radial scar, or sclerosing adenosis (≥2 foci of lobular units involved and
≥1 focus with a diameter larger than that of a normal lobule). Other histologic features, such
as cysts, apocrine metaplasia, FA, sclerosing adenosis (<2 lobular units involved), or
Author Manuscript

columnar cell changes, were considered to be nonproliferative unless they also contained 1
of the described proliferative lesions. Sufficient tissue was available to assess the presence
or absence of FA for the 9,076 women. FA was diagnosed when a fibroepithelial lesion with
both ductal epithelial and stromal fibrous proliferations measured at least 4.0 mm in
diameter [9]. Another breast pathologist (A.N.) carried out a second review of all cases with
FA and categorized the parenchyma as either PDWA or AH.

At secondary review, all FAs were classified as complex or noncomplex. According to

previously published criteria [8], FA was classified as complex (Figure) when it contained 1
or more of the following: cyst diameter 3 mm or larger, sclerosing adenosis, epithelial
calcifications, and papillary apocrine metaplasia. In addition, other coexisting histologic
findings were recorded for each biopsy sample: nonproliferative and proliferative lesions (as
described above), and AH. Each sample was evaluated for the presence of surrounding
Author Manuscript

breast parenchyma adjacent to the FA and was defined as sufficient for evaluation if it was
at least 0.5 cm2 in area or if it had 5 lobular units (a lobular unit is a cluster of acini that
drain into a terminal duct lobular unit). Finally, the location of all PDWA and AH findings
was documented (as within the FA, in adjacent parenchyma, or in both locations).

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 4

The adjacent breast parenchymal tissue was also evaluated for extent of lobular involution.
Author Manuscript

Lobular involution was classified as no involution (0% involuted lobules), partial involution
(1%-74% involuted lobules), or complete involution (≥75% involuted lobules) [7].

The Mayo Clinic Institutional Review Board approved the research. Informed consent was
also received from all participants.

Statistical Analysis
Data were summarized descriptively as frequencies and percentages for categorical variables
and as means and ranges for continuous variables. We used χ2 tests of significance to
compare the presence of FA across the levels of categorical variables (including age at
biopsy, era of biopsy, indication for biopsy, histologic impression, extent of lobular
involution, and family history of BC).
Author Manuscript

Duration of follow-up was calculated as the number of days after biopsy of a benign lesion
to the date of BC diagnosis, death, or last contact. In addition, if women had prophylactic
mastectomies or a diagnosis of lobular carcinoma in situ, their data were censored at the date
of such an occurrence. Standardized incidence ratios (SIRs) with 95% CIs were used to
estimate RRs; SIRs were calculated by dividing the actual number of women who had BC
by the number expected to have BC. The expected number was determined by dividing the
patient’s follow-up into 5-year periods according to the patient’s age and according to the
calendar period; this accounted for differences related to the variables. The reference
population was the Iowa Surveillance, Epidemiology, and End Results registry [6]. Like the
Mayo Clinic patients, 80% of the reference population lived in the Upper Midwest, and the 2
groups were similar demographically (eg, >95% of both groups were white) [6]. SIRs were
calculated for the overall group and for subgroups defined by histologic impression and
Author Manuscript

other BBD characteristics. Potential heterogeneity in SIRs across subgroups was assessed
with Poisson regression analysis; the offset term was the log-transformed expected event
rate for each patient. All statistical analyses were 2-sided and used SAS statistical software
(SAS Institute Inc). P values less than .05 were considered statistically significant.

Results
Clinicopathologic Characteristics
A total of 9,076 women comprised the cohort: 6,940 (76.5%) with no FA; 1,835 (20.2%)
with simple FA; and 301 (3.3%) with complex FA. Mean age was 52.2 years for women
without FA, 45.8 years for women with simple FA, and 50.2 years for women with complex
FA. Nearly half the women with simple FA were younger than 45 years, compared with
Author Manuscript

40% for women with complex FA (Table 1).

Simple FA was detected more often as a lump (57.2%) compared to complex FA (49.0%).
Complex FA was more common in the postmammographic era between 1982 and 1991
(65.8%) compared to the premammographic era before 1982 (34.2%). The proportion of
women with FA who also had AH was low overall but approximately 4 times greater in the
complex FA group (6.0%) than in the simple FA group (1.6%).

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 5

Involution differed among women according to FA status, with a general trend toward less
Author Manuscript

complete involution in women with FA compared to those without. Complete involution

was seen less frequently in women with simple FA (n=216; 15.1%) and complex FA (n=27;
11.7%) than in women without FA (n=1,588; 23.3%) (P=.002). Family history of BC was
not common among women with either simple FA or complex FA. Strong family history
was noted for only small percentages of women with simple FA (8.7%) or complex FA
(11.7%).

Association of FA and Other Risk Factors With BC Risk

BC occurred in 151 (8.2%) of the 1,835 women with simple FA and in 40 (13.3%) of the
301 women with complex FA (Table 2). The RR of BC for the entire FA cohort was 1.60
(95% CI, 1.38–1.85), which was similar to the RR for the group without FA (RR, 1.51; 95%
CI, 1.40–1.63) and for the group with simple FA (RR, 1.49; 95% CI, 1.26–1.74). However,
Author Manuscript

the group with complex FA had a slightly increased RR of 2.27 (95% CI, 1.63–3.10)
compared with the general population (test for heterogeneity in RR between simple FA and
complex FA, P=.02).

For women with simple FA, the risk of BC did not differ substantively across subcategories
of age, era of biopsy, involution status, and family history of BC (Table 2). However,
women with simple FA presenting with a breast lump (RR, 1.95; 95% CI, 1.59–2.38) had a
higher risk than those with FA detected on mammography (RR, 1.09; 95% CI, 0.80–1.44;
P<.001).

As expected, there was a significant difference in risk when simple FA was associated with
proliferative disease compared to nonproliferative disease (RR, 1.20; 95% CI, 0.97–1.45),
PDWA (RR, 2.73; 95% CI, 1.99–3.66), or AH (RR, 4.07; 95% CI, 1.49–8.86; P<.001)
Author Manuscript

(Table 2). Similar trends were seen for women with complex FA, but results were not
statistically significant, in part because the sample sizes were smaller (test for heterogeneity
in RR between simple FA and complex FA: P=.35 for NP, P=.32 for PDWA, and P=.59 for
AH). For women with proliferative disease, the location of proliferative histologic findings
in association with FA was also studied (Table 3). The BC risk was not different when the
PDWA was located adjacent to the FA or within the FA (P=.97). Similarly, the risk was not
influenced by the location of AH, whether adjacent to or within the FA (P=.41). Findings
were similar when we looked at simple and complex FA separately, although the numbers in
each category were too small in the subgroup analysis.

Of the 4 proliferative disease components of FA, sclerosing adenosis was predominant

(79.4%; 239 of 301). No individual component conferred an increased risk of BC compared
Author Manuscript

to the others (Table 4), although these analyses were limited by power and should be
interpreted with caution. The presence of any 1 of the histologic criteria was not associated
with increased risk of BC, but when 2 (RR, 2.67; 95% CI, 1.52–4.33) or more (RR, 7.77;
95% CI, 3.13–15.99) histologic features of complex FA were found simultaneously in the
complex FA, there was an increase in BC risk. However, the group of women with 3 lesions
in a complex FA was very small (6.3%; 19 of 301) (Table 4).

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 6

Discussion
Author Manuscript

In this study, the overall risk of BC for women with FA was slightly elevated (RR, 1.60;
95% CI, 1.38–1.85) and similar to the risk of women without FA (RR, 1.51; 95% CI, 1.40–
1.63). Women with complex FA had a slightly higher RR of 2.27 (95% CI, 1.63–3.10) than
those with simple FA (RR, 1.49; 95% CI, 1.26–1.74; test for heterogeneity in RR, 0.022).
However, in analyses stratified by histologic impression, these differences attenuated
greatly, suggesting that the risk of BC with FA was associated with the greatest degree of
epithelial proliferation (PDWA or AH, whether within or adjacent to the FA) rather than
with the mere presence of FA.

Dupont et al [2] described BC risk in women with FAs almost 2 decades ago. In a
retrospective cohort study of 1,835 women with FA biopsied from 1950 through 1968, BC
developed in 87 (4.7%). In a comparison with 2 control groups, 1 of which included the
Author Manuscript

patient’s sister-in-law, Dupont et al concluded that FA confers an overall RR of 2.17 (95%

CI, 1.5–3.2). However, this risk increased to 3.10 (95% CI, 1.9–5.1) for 422 women with
complex FA, including 29 (6.9%) who had BC. Dupont et al [2] further showed that this
increased risk persisted for decades after the initial diagnosis. In the present study, 2,136
women with FAs were studied and BC developed in 191 women (8.9%) during a 20-year
follow-up. The BC risk was slightly elevated for women with FA and more so for women
with complex FA. Similar to the study findings of Dupont et al, although women with
complex FA were at slightly elevated BC risk, the risk was not significantly different
between women with simple FA and women with no FA (Table 2). It is notable that these 2
cohorts are among the largest cohorts of women with FA studied to date.

In the present study, among women with FA, the presence of PDWA or AH was associated
Author Manuscript

with greater BC risk compared to nonproliferative disease (P<.001) (Table 2). PDWA
(including usual ductal hyperplasia other than mild, sclerosing adenosis, papillomas, and
radial scar) was noted in 22.9% (490 of 2,136) of all women with FA and more frequently in
women with complex FA (218 of 301; 72.4%). In comparison, Dupont et al [2] reported an
incidence of 13.7% for PDWA with FA and an increased BC risk of 3.9 for FA with PDWA.
In contrast, AH was seen with 2.2% of all FA but was relatively more common with
complex FA (6.0%) than with simple FA (1.6%). The frequency of atypical ductal
hyperplasia (1.2%) was somewhat lower in the study by Kuijper et al [1] than in the study
by Dupont et al [2] (1.7%). In the present study, the elevated risk for women with FA was
not significantly different from that for women without FA (RR, 1.51; 95% CI, 1.40–1.63).
These findings suggest that the elevated risk for women with FA is similar to the overall
BBD cohort risk rather than specific for the presence of the FA.
Author Manuscript

The study by Dupont et al [2] reported risk association according to the location of the
proliferative disease—within the FA or outside the FA. In our study, PDWA was associated
with a slightly increased risk irrespective of whether it was present within the FA or outside
the FA. The presence of AH within the FA was associated with an elevated risk, but the
number of patients was small (RR, 6.03; 95% CI, 2.21–13.12). Hence, our study suggests
that the location of the PDWA or AH within the FA or outside the FA did not influence BC
risk (Table 3). Once again, these data suggest that the risk of BC in women with FA is

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 7

driven mostly by the histologic features present in association with the FA, such as PDWA
Author Manuscript

or AH.

In studying the association with other known risk factors for BC, we noted that the
association of involution and risk in women with FA (RR, 1.72; 95% CI, 1.27–2.28 for no
involution and RR, 0.80; 95% CI, 0.41–1.40 for complete involution; P=.03) was similar to
the association for women with BBD overall [6]. Although there was no significant
association between family history of BC and BC risk for women with FA (P=.08), women
with a strong family history of BC had higher risk than those with no family history. In
summary, the BC risk for the 3 categories of BBD in this cohort of women with FA was
similar to the risk of women with BBD overall, suggesting that the elevated risk in this
population was likely reflecting the underlying risk associated with BBD rather than with
the FA itself.
Author Manuscript

Dupont et al [2] reported elevated BC risk for women with complex FA. When we assessed
the 4 histologic components of FA for associations with BC risk in the present study, none
of the components of complex FA (cysts >3 mm, sclerosing adenosis, epithelial
calcifications, or papillary apocrine changes) was significantly associated with BC risk
(Table 4). However, the presence of 2 or more of these components increased the risk of BC
compared to only 1 factor (P=.007). BC risk was noted to be slightly higher for women with
complex FA and PDWA as well as for those with complex FA and AH, but the difference
was not significant (P=.17) (Table 2). This is likely because complex FA includes the
histologic findings of PDWA, such as sclerosing adenosis, and the reported increased risk
with complex FA reflects the risk associated with PDWA. As confirmed by a prior study
from our group, sclerosing adenosis was found to be an independent risk marker for BC
(SIR for BC, 2.10) [10].
Author Manuscript

All 2,136 FAs in the present study were excised, and none had concurrent BC. Prior reports
showed the co-occurrence of in situ and invasive cancer ranging from 0.2% to 2% [1,11–
13], and Kuijper et al [1] reported carcinoma in situ (both ductal carcinoma in situ and
lobular carcinoma in situ) in 2.0% of patients with FA. As a result, several authors have
opted for conservative management of FA in younger women [1]. Another study showed
that a large proportion of FAs in women younger than 20 years resolves spontaneously [14].
Studies have also reported that carcinoma in situ arising within FA is found mostly in
women older than 35 years; therefore, removal of FAs in this age group has been suggested
[1]. Although the present report did not specifically compare conservative management with
excision of FA, the finding that malignancy within FA is rare suggests that a conservative
approach with observation of FA is reasonable unless the mass increases in size over time.
Author Manuscript

Ciatto et al [15] noted that the SIR for histologically confirmed FAs was slightly increased
(SIR, 2.0; 95% CI, 1.4–2.7), whereas there was no apparent risk of nonexcised, clinically or
mammographically detected FAs (SIR, 0.97; 95% CI, 0.7–1.4). Those investigators
suggested that BC risk assessment may be biased, specifically when the analysis is limited to
surgically excised FAs. All patients in the present study had surgical excision of their FAs,
and we found a mild increase in BC risk for women with complex FA who presented with a
lump compared with FA detected by mammogram, although the difference was not

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 8

statistically significant (Table 2). The reason for this is unclear. However, surgical excision
Author Manuscript

by itself should not alter risk for FA since the lesion by itself is considered a
nonproliferative histologic entity that is not associated with an increase in BC risk (RR for
NP disease, 1.21; 95% CI, 1.08–1.34) (Table 2).

The strength of the present study is its large BBD cohort exceeding 9,000 women, including
more than 2,100 women with FA and women who were found to have FA in both the
premammographic era and the postmammographic era. Samples were reviewed extensively
by 2 experienced breast pathologists who used established criteria previously reported. The
limitation of the current study is that all women had a surgical biopsy performed, and thus
we cannot comment on the current practice of clinical and radiologic observation alone,
which is advised for mammographically detected small lesions found in some women.

Conclusion
Author Manuscript

The risk of BC for women with FA of the breast is similar to that for women with a
diagnosis of BBD. Complex FA does not confer increased risk of BC beyond that of other
established histologic features such as PDWA and AH. Therefore, the management of
complex FA should be based on the associated histologic findings of BBD, such as PDWA
or AH.

Acknowledgments
The authors thank T. Allers, J. Johnson, and A. Harris and the Mayo Survey Research Center for data collection.

This research was supported by an R01 CA132879 from the National Cancer Institute (L.C.H.), the Mayo Clinic
Breast Cancer Specialized Programs of Research Excellence (SPOREs) grant CA116201 (L.C.H. and D.C.R.), and
the Fred C. and Katherine B. Andersen Foundation (L.C.H.).
Author Manuscript

Abbreviations

AH atypical hyperplasia
BBD benign breast disease
BC breast cancer
FA fibroadenoma
PDWA proliferative disease without atypia
RR relative risk
SIR standardized incidence ratio
Author Manuscript

References
1. Kuijper A, Mommers EC, van der Wall E, van Diest PJ. Histopathology of fibroadenoma of the
breast. Am J Clin Pathol. 2001; 115(5):736–742. [PubMed: 11345838]
2. Dupont WD, Page DL, Parl FF, et al. Long-term risk of breast cancer in women with fibroadenoma.
N Engl J Med. 1994; 331(1):10–15. [PubMed: 8202095]

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 9

3. McDivitt RW, Stevens JA, Lee NC, Wingo PA, Rubin GL, Gersell D. The Cancer and Steroid
Hormone Study Group. Histologic types of benign breast disease and the risk for breast cancer.
Author Manuscript

Cancer. 1992; 69(6):1408–1414. [PubMed: 1540878]

4. Kabat GC, Jones JG, Olson N, et al. A multi-center prospective cohort study of benign breast
disease and risk of subsequent breast cancer. Cancer Causes Control. 2010; 21(6):821–828.
[PubMed: 20084540]
5. El-Wakeel H, Umpleby HC. Systematic review of fibroadenoma as a risk factor for breast cancer.
Breast. 2003; 12(5):302–307. [PubMed: 14659144]
6. Hartmann LC, Sellers TA, Frost MH, et al. Benign breast disease and the risk of breast cancer. N
Engl J Med. 2005; 353(3):229–237. [PubMed: 16034008]
7. Milanese TR, Hartmann LC, Sellers TA, et al. Age-related lobular involution and risk of breast
cancer. J Natl Cancer Inst. 2006; 98(22):1600–1607. [PubMed: 17105983]
8. Page, DL.; Anderson, TJ. Diagnostic Histopathology of the Breast. Edinburgh: Churchill
Livingstone; 1987. p. 362
9. Azzopardi, JG. Fibroadenoma. In: Bennington, JL., editor. Problems in Breast Pathology. Major
problems in pathology. Vol. 11. Philadelphia, PA: Saunders; 1979. p. 39-56.
Author Manuscript

10. Visscher DW, Nassar A, Degnim AC, et al. Sclerosing adenosis and risk of breast cancer. Breast
Cancer Res Treat. 2014; 144(1):205–212. [PubMed: 24510013]
11. Ozzello L, Gump FE. The management of patients with carcinomas in fibroadenomatous tumors of
the breast. Surg Gynecol Obstet. 1985; 160(2):99–104. [PubMed: 2982218]
12. Deschenes L, Jacob S, Fabia J, Christen A. Beware of breast fibroadenomas in middle-aged
women. Can J Surg. 1985; 28(4):372–374. [PubMed: 2990650]
13. Buzanowski-Konakry K, Harrison EG Jr, Payne WS. Lobular carcinoma arising in fibroadenoma
of the breast. Cancer. 1975; 35(2):450–456. [PubMed: 1111920]
14. Cant PJ, Madden MV, Coleman MG, Dent DM. Non-operative management of breast masses
diagnosed as fibroadenoma. Br J Surg. 1995; 82(6):792–794. [PubMed: 7627513]
15. Ciatto S, Bonardi R, Zappa M, Giorgi D. Risk of breast cancer subsequent to histological or
clinical diagnosis of fibroadenoma: retrospective longitudinal study of 3938 cases. Ann Oncol.
1997; 8(3):297–300. [PubMed: 9137802]
Author Manuscript
Author Manuscript

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 10
Author Manuscript
Author Manuscript
Author Manuscript

Figure.
Complex fibroadenoma with sclerosing adenosis (crowded glands in a fibrotic stroma)
(hematoxylin-eosin; original magnification ×10).
Author Manuscript

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Nassar et al. Page 11

Table 1

Clinicopathologic Characteristics of the Entire Cohort by FA Status

Author Manuscript

Status, No. (%)a

Characteristic No FA (n=6,940) Simple FA (n=1,835) Complex FA (n=301) P Valueb

Age at BBD diagnosis, y .007
<45 1,973 (28.4) 888 (48.4) 118 (39.2)
45–55 2,109 (30.4) 390 (21.3) 68 (22.6)
>55 2,858 (41.2) 557 (30.4) 115 (38.2)
Era of biopsy .03
Premammographic 3,490 (50.3) 752 (41.0) 103 (34.2)
Postmammographic 3,450 (49.7) 1,083 (59.0) 198 (65.8)
Indication for biopsy .01
Missing data 388 156 15
Author Manuscript

Lump 4,007 of 6,552 (61.2) 960 of 1,679 (57.2) 140 of 286 (49.0)
Mammogram 2,545 of 6,552 (38.8) 719 of 1,679 (42.8) 146 of 286 (51.0)
Overall histologic impression <.001
NP disease 4,439 (64.0) 1,534 (83.6) 65 (21.6)
PDWA 2,215 (31.9) 272 (14.8) 218 (72.4)
AH 286 (4.1) 29 (1.6) 18 (6.0)
Involution .002

Missing datac 138 401 71

None 1,019 of 6,802 (15.0) 506 of 1,434 (35.3) 60 of 230 (26.1)

Partial 4,195 of 6,802 (61.7) 712 of 1,434 (49.7) 143 of 230 (62.2)
Complete 1,588 of 6,802 (23.3) 216 of 1,434 (15.1) 27 of 230 (11.7)
Family history of breast cancer .20
Author Manuscript

Missing data 30 11 1
None 4,618 of 6,910 (66.8) 1,181 of 1,824 (64.7) 194 of 300 (64.7)
Weak 1,610 of 6,910 (23.3) 484 of 1,824 (26.5) 71 of 300 (23.7)
Strong 682 of 6,910 (9.9) 159 of 1,824 (8.7) 35 of 300 (11.7)

Abbreviations: AH, atypical hyperplasia; BBD, benign breast disease; FA, fibroadenoma; NP, nonproliferative; PDWA, proliferative disease
without atypia.
a
Numbers may not add up to the total number of patients because of missing values for some variables.
b 2
χ test comparing distributions of characteristics across simple FA vs complex FA.
c
No normal terminal duct lobular units were present on the slide, so involution could not be assessed.
Author Manuscript

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 2

Relative Risk of Breast Cancer by FA Status and Other Characteristics

No FA Simple FA Complex FA All FA

Nassar et al.

Obs Exp Obs Exp Obs Exp Obs Exp

Total Cases, Cases, SIRa Total Cases, Cases, SIRa Total Cases, Cases, SIRa Total Cases, Cases, SIRa
Characteristic No. No. No. (95% CI) No. No. No. (95% CI) No. No. No. (95% CI) No. No. No. (95% CI)
Overall 6,940 683 454.76 1.50 (1.39– 1.62) 1,835 151 101.55 1.49 (1.26– 1.74) 301 40 17.58 2.27 (1.63– 3.10) 2,136 191 119.13 1.60 (1.38– 1.85)
Age at BBD diagnosis, y
<45 1,973 150 104.85 1.43 (1.21– 1.68) 888 60 38.28 1.57 (1.20– 2.02) 118 16 5.08 3.15 (1.80– 5.11) 1,006 76 43.36 1.75 (1.38– 2.19)
45–55 2,109 251 163.20 1.54 (1.35– 1.74) 390 48 27.66 1.74 (1.28– 2.30) 68 12 4.52 2.65 (1.37– 4.63) 458 60 32.18 1.86 (1.42– 2.40)
>55 2,858 282 186.71 1.51 (1.34– 1.70) 557 43 35.61 1.21 (0.87– 1.63) 115 12 7.98 1.50 (0.78– 2.63) 672 55 43.59 1.26 (0.95– 1.64)

P valueb .78 .20 .12 .07

Era of biopsy
Premammographic 3,490 412 242.99 1.70 (1.54– 1.87) 752 75 44.32 1.69 (1.33– 2.12) 103 17 6.57 2.59 (1.51– 4.14) 855 92 50.89 1.81 (1.46– 2.22)
Postmammographic 3,450 271 211.77 1.28 (1.13– 1.44) 1,083 76 57.23 1.33 (1.05– 1.66) 198 23 11.01 2.09 (1.32– 3.14) 1,281 99 68.24 1.45 (1.18– 1.77)

P valueb .003 .14 .51 .13

Indication for biopsy

Lump 4,007 422 267.07 1.58 (1.43– 1.74) 960 100 51.18 1.95 (1.59– 2.38) 140 23 7.66 3.00 (1.90– 4.51) 1,100 123 58.84 2.09 (1.74– 2.49)
Mammogram 2,545 242 163.02 1.48 (1.30– 1.68) 719 47 43.26 1.09 (0.80– 1.44) 146 17 9.3 1.83 (1.06– 2.93) 865 64 52.56 1.22 (0.94– 1.55)

P valueb .44 <.001 .12 <.001

Overall histologic impression

NP disease 4,439 347 289.92 1.20 (1.07– 1.33) 1,534 100 83.62 1.20 (0.97– 1.45) 65 6 3.32 1.81 (0.66– 3.93) 1,599 106 86.93 1.22 (1.00– 1.47)
PDWA 2,215 262 148.59 1.76 (1.56– 1.99) 272 45 16.46 2.73 (1.99– 3.66) 218 29 13.38 2.17 (1.45– 3.11) 490 74 29.84 2.48 (1.95– 3.11)
AH 286 74 16.25 4.55 (3.58– 5.72) 29 6 1.47 4.07 (1.49– 8.86) 18 5 0.89 5.65 (1.84– 13.16) 47 11 2.36 4.67 (2.33– 8.35)
<.001 <.001 .17 <.001

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
P valueb
Involution
None 1,019 126 65.44 1.93 (1.60– 2.29) 506 38 25.01 1.52 (1.08– 2.09) 60 10 2.9 3.45 (1.65– 6.35) 566 48 27.9 1.72 (1.27– 2.28)
Partial 4,195 425 279.56 1.52 (1.38– 1.67) 712 66 43.14 1.53 (1.18– 1.95) 143 18 9.33 1.93 (1.14– 3.05) 855 84 52.47 1.60 (1.28– 1.98)
Complete 1,588 110 101.86 1.08 (0.89– 1.30) 216 10 13.27 0.75 (0.36– 1.39) 27 2 1.68 1.19 (0.14– 4.28) 243 12 14.95 0.80 (0.41– 1.40)

P valueb <.001 .06 .22 .03

Family history of breast cancer

Page 12
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

No FA Simple FA Complex FA All FA

Obs Exp Obs Exp Obs Exp Obs Exp

Total Cases, Cases, SIRa Total Cases, Cases, SIRa Total Cases, Cases, SIRa Total Cases, Cases, SIRa
Characteristic No. No. No. (95% CI) No. No. No. (95% CI) No. No. No. (95% CI) No. No. No. (95% CI)
None 4,618 375 292.36 1.28 (1.16– 1.42) 1,181 84 61.54 1.36 (1.09– 1.69) 194 20 10.93 1.83 (1.12– 2.83) 1,375 104 72.47 1.44 (1.17– 1.74)
Nassar et al.

Weak 1,610 205 110.57 1.85 (1.61– 2.13) 484 46 29.02 1.59 (1.16– 2.11) 71 12 4.2 2.86 (1.48– 4.99) 555 58 33.22 1.75 (1.33– 2.26)
Strong 682 101 50.61 2.00 (1.63– 2.42) 159 21 10.4 2.02 (1.25– 3.09) 35 8 2.29 3.49 (1.51– 6.87) 194 29 12.7 2.28 (1.53– 3.28)

P valueb <.001 .27 .24 .08

Abbreviations: AH, atypical hyperplasia; BBD, benign breast disease; Exp, expected; FA, fibroadenoma; NP, nonproliferative; Obs, observed; PDWA, proliferative disease without atypia; SIR, standardized incidence ratio.
a
The standardized incidence ratio compares the observed number of breast cancer events with the number expected on the basis of Iowa Surveillance, Epidemiology, and End Results data. All analyses account for the effects of age and calendar period.
b
Test for heterogeneity in risk of breast cancer across levels of attribute.

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
Page 13
 Nassar et al. Page 14

Table 3

Risk of Breast Cancer According to Location of Proliferative Histologic Findings Within the FA or Outside
Author Manuscript

the FA

All FA

Characteristic Total No. Obs Cases, No. Exp Cases, No. SIRa (95% CI)

PDWA in relation to FA
Outside the FA 295 45 18.08 2.49 (1.81–3.33)
Within the FA 195 29 11.75 2.47 (1.65–3.54)

P valueb .97

AH in relation to FA
Outside the FA 28 5 1.36 3.67 (1.19–8.54)
Within the FA 19 6 0.99 6.03 (2.21–13.12)

P valueb .41
Author Manuscript

Abbreviations: AH, atypical hyperplasia; Exp, expected; FA, fibroadenoma; Obs, observed; PDWA, proliferative disease without atypia; SIR,
standardized incidence ratio.
a
The standardized incidence ratio compares the observed number of breast cancer events with the number expected on the basis of Iowa
Surveillance, Epidemiology, and End Results data. All analyses account for the effects of age and calendar period.
b
Test for heterogeneity in risk of breast cancer was only between outside the FA and within the FA.
Author Manuscript
Author Manuscript

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
 Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Table 4

Association of Individual Complex FA Criteria With Risk of Breast Cancer Among Women With a Complex FA

Characteristic Women, No. Person-years, No. Observed Events, No. Expected Events, No. SIRa (95% CI) P Valueb
Nassar et al.

Cysts >3 mm .10

No 211 3,976 24 12.53 1.92 (1.23–2.85)
Yes 90 1,573 16 4.85 3.30 (1.89–5.36)
Sclerosing adenosis .09
No 62 1,138 4 3.42 1.17 (0.32–2.99)
Yes 239 4,412 36 13.95 2.58 (1.81–3.57)
Epithelial calcifications .35
No 270 4,995 37 15.29 2.42 (1.70–3.34)
Yes 31 554 3 2.08 1.44 (0.30–4.20)
Papillary apocrine metaplasia .07
No 221 4,060 25 13.09 1.91 (1.24–2.82)
Yes 80 1,490 15 4.29 3.50 (1.96–5.77)
Complex FA lesions, No. .007
1 181 3,385 17 10.48 1.62 (0.94–2.60)
2 101 1,850 16 6.00 2.67 (1.52–4.33)
3 19 315 7 0.90 7.77 (3.13–15.99)

Abbreviations: FA, fibroadenoma; SIR, standardized incidence ratio.

a
The standardized incidence ratio compares the observed number of breast cancer events with the number expected on the basis of Iowa Surveillance, Epidemiology, and End Results data. All analyses
account for the effects of age and calendar period.
b
Test for heterogeneity in risk of breast cancer across levels of attribute.

Breast Cancer Res Treat. Author manuscript; available in PMC 2016 September 01.
Page 15