Accepted Manuscript

Criteria for Excision of Suspected Fibroadenomas of the Breast

Jennifer L. Hubbard, MD Kathleen Cagle, RN, MPH James W. Davis, MD Krista L.Kaups, MD Miya Kodama

PII: S0002-9610(14)00153-6

DOI: 10.1016/j.amjsurg.2013.12.037

Reference: AJS 11129

To appear in: The American Journal of Surgery

Received Date: 30 July 2013

Revised Date: 25 October 2013

Accepted Date: 22 December 2013

Please cite this article as: Hubbard JL, Cagle K, Davis JW, Kaups KL, Kodama M, Criteria for Excisionof Suspected Fibroadenomas of the Breast, The American Journal of Surgery (2014), doi: 10.1016/j.amjsurg.2013.12.037.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service toour customers we are providing this early version of the manuscript. The manuscript will undergocopyediting, typesetting, and review of the resulting proof before it is published in its final form. Pleasenote that during the production process errors may be discovered which could affect the content, and alllegal disclaimers that apply to the journal pertain.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 1 of 16

Criteria for Excision of Suspected Fibroadenomas of the Breast

Jennifer L. Hubbard, MD, Kathleen Cagle, RN, MPH, James W. Davis, MD, Krista L. Kaups,

MD, Miya Kodama

UCSF Fresno Department of Surgery, Fresno, CA

Please send correspondence to

Jennifer Hubbard, MD Assistant Professor of Surgery University of California San Francisco -- Fresno 2823 Fresno Street 1st Floor Department of Surgery Fresno, CA 93721 Email: jhubbard@fresno.ucsf.edu. Phone: (559) 459-3722 Fax: (559) 459-3719

Disclosures: none (for all authors)

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 2 of 16

Keywords: fibroadenoma, breast, excision, criteria

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 3 of 16

Brief title: Criteria for observation of fibroadenomas

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 4 of 16

Summary: A retrospective review of 723 patients who underwent surgical excision for suspected

fibroadenomas was used to identify risk factors for non-fibroadenoma pathology. Patients

should undergo excision when age > 35, immobile or poorly circumscribed mass, size > 2.5 cm,

or pre-operative biopsy not definitive for fibroadenoma.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 5 of 16

Abstract

Background: Fibroadenomas are benign breast tumors; however, more aggressive lesions may

mimic or arise within fibroadenomas. We sought to define criteria identifying patients who

should undergo surgical excision. Methods: Patients with a pre-operative diagnosis of

fibroadenoma, who underwent surgical excision between 2002 through 2011, were

retrospectively reviewed. Patients with final pathologic diagnosis of fibroadenoma were

compared with those with non-fibroadenoma pathology. Results: Of 723 patients, 681 (94%)

had fibroadenomas on final pathology. The incidence of non-fibroadenoma pathology was 6%

(42 patients), and included benign phylloides (23), malignant phylloides (2), atypical ductal

hyperplasia (1), intraductal papilloma (5), and other benign pathology (11). No cases of

adenocarcinoma were identified. Non-fibroadenoma pathology was associated with age >35,

immobile or poorly-circumscribed mass, size > 2.5 cm, and biopsy not definitive for

fibroadenoma. Conclusions: Patients with age >35 years, immobile or poorly-circumscribed

mass, size > 2.5 cm, or biopsy not definitive for fibroadenoma should undergo surgical excision.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 6 of 16

Background

Fibroadenomas are benign, solid neoplasms of the breast consisting of fibroepithelial

elements. Their size is hormonally influenced, as evidenced by fluctuation in size with the

menstrual cycle and regression in postmenopausal women. Fibroadenomas are often solitary

masses but 25% of patients present with multiple lesions. They have a characteristic clinical

presentation: rubbery, mobile, and firm. Despite this, previous reports have indicated that

diagnosis by clinical exam is accurate in only 50-75% of patients1.

One of the clinical dilemmas facing both surgeons and patients is the concern that the

mass is something more ominous than a fibroadenoma. Both benign and malignant phylloides

tumors may mimic fibroadenomas. Additionally, published reports have described

adenocarcinoma and DCIS arising within fibroadenomas or misdiagnosed as fibroadenomas.

Because of the potential for more aggressive pathology masquerading as fibroadenomas,

management has been debated and recommendations changed several times in recent decades.

Until the mid-1980s, standard practice was excision of all fibroadenomas2. Subsequent studies

in the 1980s and 1990s demonstrated the safety of observing the presumed fibroadenomas in

women under age 35 who had a fine needle aspirate biopsy that did not contain malignant or

suspicious cells1, 3. More recently, the question has been asked whether biopsy is even

necessary. Smith and Burrows concluded that patients under the age of 25 with benign

ultrasound findings could be safely observed without a biopsy4.

We hypothesized that risk factors for more aggressive pathology could be identified in

women with presumed fibroadenomas. Identification of these risk factors would assist

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 7 of 16

physicians in counseling and treating women with presumed fibroadenomas and allow the

development of guidelines for biopsy or excision.

Materials and Methods

The study was approved by the institutional review board of the University of California

San Francisco, Fresno. The patient population was drawn from a large community hospital

system in Fresno County that includes several outpatient clinics and two major hospitals. We

reviewed the electronic and handwritten medical records of all patients with a preoperative

diagnosis of fibroadenoma who underwent surgical excision between 2002 and 2011. Men were

excluded from the study. Patient demographics and risk factors for breast cancer were collected.

Past or current use of hormonal therapy, including contraceptives, was recorded. Additionally,

the patient’s ultrasound results, estimated preoperative size of the mass, preoperative biopsy

results, and final size and pathology were recorded. In patients who underwent an ultrasound,

mass characteristics including increasing size on serial studies, height versus width, BI-RADS

category, and echogenicity, were also collected. Because of the variability of percutaneous

biopsy pathologic interpretation, we categorized results as either fibroadenoma or non-

fibroadenoma. Non-fibroadenoma pathology included fibroepithelial neoplasm, spindle-cell

neoplasm, possible phylloides tumor, and other non-specific diagnoses.

Most of the data was collected from pre-operative history and physical examinations that

were handwritten, prior to the implementation of electronic medical records. In order to make

the greatest use of our data, we made some a priori assumptions that history was negative if not

otherwise documented and physical examination was normal if not documented. Indications for

surgical excision were based on size (or increase in size), characteristics of the mass (ill-defined,

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 8 of 16

poorly circumscribed, hard), patient discomfort, and/or ultrasound characteristics, with the final

determination made by the attending surgeon.

Statistical analyses were performed using paired student’s t test, z-test, Fisher’s exact

test, Mann-Whitney U test, Chi-square analysis, and Mantel-Haenszel estimates (odds ratio

analysis). A p value of 0.05 was considered significant. Statistical analysis was performed using

IBM SPSS software.

Following analysis and identification of significant risk factors, criteria for observation

vs. excision were created and applied retrospectively to the patient database to calculate how

many patients could have safely avoided an operation, and how many with more aggressive

pathology would have been missed in our dataset.

Results

Characteristics of the study cohort, including demographics, are listed in Table I. Of the

723 patients meeting inclusion criteria, 94% (681patients) had a fibroadenoma on final

pathology. The other 6% (42 patients) had pathology demonstrating benign phylloides (23),

malignant phylloides (2), atypical ductal hyperplasia (1), intraductal papilloma (5), and other

benign pathology (tubular adenoma, mastitis, adenomyoepithelioma, benign neurofibroma,

fibrocystic changes, stromal sclerosis, and non-specific fibroadipose tissue). We did not identify

any cases of adenocarcinoma or carcinoma in situ in this cohort of patients.

Patient ethnicity was largely Hispanic and Caucasian and mirrors the ethnicity of Fresno

County, which is 50.9% Hispanic/Latino and 32.4% Caucasian (US Census Bureau, 2011 Fresno

County Data). Most of the known risk factors for breast cancer were not associated with non-

fibroadenoma pathology (Table II). Family history of cancer (other than breast cancer),

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 9 of 16

nulliparity, smoking history, breastfeeding history, early menarche, personal history of breast

cancer, history of prior breast biopsy, and first pregnancy with advanced maternal age, was not

significantly different between the fibroadenoma and non-fibroadenoma group. However,

patients aged > 35 years were significantly more likely to have non-fibroadenoma pathology (OR

2.8, p=0.002, see Table II).

Physical exam

The classical description of a fibroadenoma as a well-circumscribed and mobile mass was

found to hold true among the patients in our study (Table 2). The positive predictive value of a

mobile mass being a fibroadenoma was 93%. A well-circumscribed mass correlated with a 95%

positive predictive value and a sensitivity of 99%. Conversely, patients with an immobile mass

were 9 times more likely to have non-fibroadenoma pathology, and patients with poorly defined

masses were 15 times more likely to have non-fibroadenoma pathology.

Imaging

None of the criteria recorded on ultrasound (height versus width, BI-RADS categories,

and echogenicity) were found to be statistically significant. There was great heterogeneity in

both the reporting of ultrasound diagnoses and descriptions of the mass. Table III shows the

distribution of BI-RADS assessments given for fibroadenoma and non-fibroadenoma pathology.

There was no statistical significance between all BI-RADS scores. Patients with masses that

required needle-localization had no difference in likelihood of non-fibroadenoma pathology

versus those with palpable masses (p = 0.69).

Size

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 10 of 16

Pre-operative size of the mass was recorded based on ultrasound findings or physical

exam, if ultrasound was not available. Size was significantly smaller in the fibroadenoma group

(mean 2.3 + 1.4 cm) vs. the non-fibroadenoma group (3.1 + 3.4, p=0.003). Using Fisher’s exact

test, a cutoff size of greater than 2.5 cm was significantly more likely to have non-fibroadenoma

pathology (OR 2.3, p < 0.03).

Preoperative Biopsy

Of the 723 patients in the study, 182 had a pre-operative biopsy, with the majority

consisting of core-needle biopsies. Only patients with a pre-operative diagnosis of fibroadenoma

were included in this study and no pre-operative biopsies were definitive for non-fibroadenoma

pathology. In our cohort, 159 biopsies (87%) were definitive for fibroadenoma and 23 biopsies

(13%) were not definitive for fibroadenoma or were suspicious for phylloides, but clinically were

felt to be fibroadenomas. Of the patients with definitive pre-operative biopsies, the final

pathology was fibroadenoma in 92%. In patients with a pre-operative biopsy that was not

definitive for fibroadenoma, only 48% had final pathology demonstrating fibroadenoma, despite

the clinical diagnosis (Table II, p < 0.001).

Criteria for Observation

Using the risk factors identified, we developed a set of criteria which allow safe

observation of the suspected fibroadenoma: age less than 35 years, physical examination

demonstrating a mobile and well-circumscribed mass, size < 2.5 cm, and, if present, a biopsy

definitive for fibroadenoma. If these criteria had been applied to the patients in this study, and

all criteria were met, observation instead of excision would have occurred in 391 of the 723

patients (54%). If these criteria had been utilized, the misdiagnosis rate would have been 4%,

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 11 of 16

compared to the actual misdiagnosis rate of 6% (those that had non-fibroadenoma on final

pathology). The misdiagnosed masses would have been benign phylloides tumor (10), tubular

adenoma, mastitis, atypical ductal hyperplasia, intraductal papilloma (2), and fibrocystic

changes.

Discussion

Fibroadenomas of the breast are common, benign lesions, yet indications for excision are

not well-defined3, 5- 8. While some fibroadenomas are excised because they cause pain or

distortion of the breast, excision often occurs because of concern by either the clinician or the

patient that it may be cancer or another aggressive lesion. Invasive carcinoma and carcinoma in

situ within a fibroadenoma are rare, with about 100 cases published in case reports and small

case series9. In addition, most of the published studies were performed prior to the development

of the American College of Radiology BI-RADS system in 1993, which was intended to aid in

standardization of mammographic reporting.

Another reason for lack of good guidelines for excision is that previously reported series

did not use surgical pathology as the gold standard. Wilkinson and colleagues described 110

patients, all under 35 years of age, of whom only 77 underwent surgical excision1. Another

series followed a cohort of 70 women with 87 fibroadenomas, of which 53 were eventually

excised3. In a more recent series by Park and colleagues, although they included a large cohort

of almost 1400 patients, only 125 underwent surgical biopsy or excision10. Thus our study is the

largest to date that includes definitive pathology for all patients.

In the current literature, the factor most predictive of carcinoma, or other aggressive

lesion within a fibroadenoma, is patient age, but even this is inconsistent across studies6, 11. In

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 12 of 16

the present study, older age was again identified as a risk factor for non-fibroadenoma pathology.

We used a cutoff age of 35 years and found it to be statistically significant. Ethnicity has not

been shown to be a significant risk factor except in a single recently published paper12. In that

series, Hispanic ethnicity was found to be a risk factor for non-fibroadenoma pathology,

specifically phylloides tumors. In contrast, our data showed a trend toward fibroadenoma

pathology with Hispanic ethnicity. Given this conflicting data and lack of data in any other

paper, we do not recommend using ethnicity as criteria for observation or excision.

Ultrasound is a commonly employed diagnostic tool for breast masses, and prior

publications have reported a high degree of accuracy in diagnosing fibroadenomas4.

Characteristics of benign lesions include oval shape, circumscribed margin, parallel orientation

to the skin surface, hypoechoic or isoechoic lesion, posterior enhancement, absence of

microcalcifications, and no surrounding tissue changes10. In reviewing available ultrasound

reports in our study, these characteristics were poorly and inconsistently documented. The most

commonly documented findings were echogenicity and size. Rarely was the orientation of the

lesion documented. Use of the BI-RADS classification was also highly variable, even in more

recent years. If the patient had a biopsy prior to the most recent ultrasound, the BI-RADS was

usually listed as 2 or 3, whereas lesions without pathologic diagnosis were variably diagnosed as

3, 4a or 4. Because of this variability in reporting, we were unable to find any statistical

significance based on ultrasound findings. Vagueness in diagnosis and variability of BI-RADS

reporting may have led to decisions to obtain tissue diagnosis, either via biopsy or excision, but

this was not documented in the medical record. We propose that more defined reporting should

be instituted. Until then, caution and clinical judgment should be used when basing diagnosis on

ultrasound findings.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 13 of 16

Many physicians use biopsy as a definitive diagnostic study for suspected breast masses.

In our series, pre-operative biopsy had a significant incidence of being non-diagnostic. In

patients with a non-diagnostic biopsy, the chance of the final pathology showing something other

than a fibroadenoma was 52%. Knowing this, the clinician should have a heightened suspicion

when a pathology report shows fibroepithelial neoplasm, spindle-cell neoplasm, or other vague

and indeterminate diagnoses, as approximately half of these patients will have some other

etiology than fibroadenoma. As with all breast biopsies, pathologic discordance with clinical or

radiographic findings should also lead to further tissue sampling, either via repeat percutaneous

biopsy, surgical biopsy, or excision.

There are limitations to our study. As mentioned previously, we made some a priori

assumptions that history was negative if not otherwise documented and physical examination

was normal if not documented. It is certainly possible that some positive findings were not

documented, and thus were not included in the analysis. As with all retrospective studies, our

study is subject to selection bias. Applying these criteria to our patient cohort would have

dramatically decreased the number of excisions. However, since none of the patients in this

cohort were observed, if our criteria are applied to all patients presenting in clinical practice with

suspected fibroadenomas, the number of surgical procedures may actually increase. We also

recognize that fibroadenomas are often excised because of pain and patient preference.

Unfortunately, this indication was rarely documented in the chart, so it could not be analyzed in

our study. We hope that providing additional reassurance of the rarity of malignancy and other

aggressive pathology will help to reassure patients who wish to have a suspected fibroadenoma

excised solely because of anxiety. Other decisions may also play into the decision to proceed

with surgical excision, such as pain associated with the fibroadenoma or poor cosmesis.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 14 of 16

The present study is the largest to date reviewing patients with suspected fibroadenomas

and associated risk factors for the occurrence of more aggressive lesions on final pathology. Our

study affirms that the widely utilized methods of diagnosing fibroadenomas, namely history,

physical examination, ultrasound, and in some cases, biopsy, have a high degree of accuracy.

Furthermore, using the risk factors identified, guidelines are provided to aid the clinician in

counseling patients when it is safe to observe suspected fibroadenomas and when surgical

excision or biopsy should be recommended.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 15 of 16

References

1. Wilkinson S, Anderson TJ, Rifkind E, et al. Fibroadenoma of the breast: a follow-up of

conservative management. Br J Surg. 1989 Apr;76(4):390-1.

2. Wilkinson S, Forrest AP. Fibro-adenoma of the breast. Br J Surg. 1985 Oct;72(10):838-

40.

3. Carty NJ, Carter C, Rubin C, et al. Management of fibroadenoma of the breast. Ann R

Coll Surg Engl. 1995 Mar;77(2):127-30.

4. Smith GE, Burrows P. Ultrasound diagnosis of fibroadenoma - is biopsy always

necessary? Clin Radiol. 2008 May;63(5):511-5; discussion 516-7.

5. Maxwell AJ, Pearson JM. Criteria for the safe avoidance of needle sampling in young

women with solid breast masses. Clin Radiol. 2010 Mar;65(3):218-22.

6. Dent DM, Cant PJ. Fibroadenoma. World J Surg. 1989 Nov-Dec;13(6):706-10.

7. Dixon JM, Dobie V, Lamb J, et al. Assessment of the acceptability of conservative

management of fibroadenoma of the breast. Br J Surg. 1996 Feb;83(2):264-5.

8. Resetkova E, Khazai L, Albarracin CT, Arribas E. Clinical and radiologic data and core

needle biopsy findings should dictate management of cellular fibroepithelial tumors of

the breast. Breast J. 2010 Nov-Dec;16(6):573-80.

9. Ooe A, Takahara S, Sumiyoshi K, et al. Preoperative diagnosis of ductal carcinoma in

situ arising within a mammary fibroadenoma: a case report. Jpn J Clin Oncol. 2011

Jul;41(7):918-23.

10. Park YM, Kim EK, Lee JH, et al. Palpable breast masses with probably benign

morphology at sonography: can biopsy be deferred? Acta Radiol. 2008 Dec;49(10):1104-

11.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Page 16 of 16

11. Diaz NM, Palmer JO, McDivitt RW. Carcinoma arising within fibroadenomas of the

breast. A clinicopathologic study of 105 patients. Am J Clin Pathol. 1991 May;95(5):614-

22.

12. Gould DJ, Salmans JA, Lassinger BK, et al. Factors associated with phyllodes tumor of

the breast after core needle biopsy identifies fibroepithelial neoplasm. J Surg Res. 2012

Nov;178(1):299-303.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Table I. Demographics and diagnostic data

Total

N=723

Non-FA

N=42

FA

N=681

p value O.R.*

Age, mean ± SD 32 38 ± 15 32 ± 13 0.008 N/A

Ethnicity

Hispanic 370 (51%) 15 (36%) 355 (52%) 0.06 N/A

Caucasian 249 (34%) 19 (45%) 230 (34%) NS N/A

Black 43 (6%) 3 (7%) 40 (6%) NS N/A

Asian 31 (4%) 3 (7%) 28 (4%) NS N/A

Eastern Indian 10 (1%) 2 (5%) 8 (1%) NS N/A

Native American 1 (<1%) 0 1 (<1%) NS N/A

Other/Unknown 19 (3%) 0 19 (3%) N/A N/A

# with pre-op biopsy

performed

190 (27%) 23 (55%) 167 (25%) <.001 N/A

# with pre-op biopsy results

recorded

181 (25%) 22 (52%) 159 (23%) <.001 N/A

# with ultrasound performed 567 (78%) 35 (83%) 532 (78%) NS N/A

# with US results recorded 481 (67%) 30 (71%) 451 (66%) NS N/A

Pre-op size, mean ± SD 2.3 3.1 ± 3.4 2.3 ± 1.4 0.003 1.2 (1.04-1.4)

Final size, mean ± SD 2.5 3.3 ± 2.6 2.4 ± 1.5 0.001 1.3 (1.1-1.5)

*calculated using regression analysis. FA = fibroadenoma.

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Table II. Clinical risk factors for non-fibroadenoma pathology

Non-FA

FA

p

value*

OR (95% C.I.) # Missing

documentation

Age > 35 26 (62%) 249 (37%) 0.002 2.8 (1.5-5.4) 0

Age < 35 16 (38%) 432 (63%)

Family hx breast cancer 0 8 (1%) 1.0 N/A 109

No family hx 42 (100%) 673 (99%)

Hormone Therapy 10 (37%) 140 (28%) 0.96 N/A 198

No hormone therapy 17 (63%) 358 (72%)

Fixed mass 2 (10%) 3 (1%) 0.046 9.4 (1.5-59.7) 431

Mobile mass 19 (90%) 268 (99%)

Ill-defined mass 3 (43%) 4 (5%) 0.008 15.4 (2.5-93.2) 630

Well-circumscribed

mass

4 (57%) 82 (95%)

U/s diagnosis non-FA 18 (62%) 200 (49%) 0.25 N/A 288

U/s diagnosis FA 11(38%) 206 (51%)

Pre-op size > 2.5 cm 17 (50%) 154 (30%) 0.03 2.3 (1.1-4.6) 181

Pre-op size < 2.5 cm 17 (50%) 354 (70%)

Pre-op biopsy non-

diagnostic

12 (52%) 12 (8%) <.001 11.2 (4.1-30.8) 8

Pre-op biopsy FA 11 (48%) 147 (92%)

Non-Palpable 9 (21%) 125 (18%) 0.69 N/A 0

Palpable 33 (79%) 556 (82%)

Final size > 2.5cm 21 (57%) 237 (35%) 0.01 2.4 (1.2-4.7) 10

Final size < 2.5cm 16(43%) 439 (65%)

*Chi Square analysis or Fisher’s Exact test. FA = fibroadenoma, Hx = history, u/s = ultrasound

MANUSCRIP

T

ACCEPTED

ACCEPTED MANUSCRIPT

Table III. Ultrasound characteristics

Total

N=549

Non-FA

N=37

FA

N=512

p value

BI-RADS

0 1 (<1%) 0 1 (<1%) 0.09

1 0 0 0 NA

2 45 (8%) 3 (8%) 42 (8%) 0.75

3 125 (23%) 6 (16%) 119 (23%) 0.43

4 194 (35%) 17 (46%) 177 (35%) 0.24

5 3 (<1%) 0 3 (<1%) 0.50

No score or missing 181 (33%) 11 (30%) 170 (33%) 0.85

Ultrasound diagnosis

Non-diagnostic for FA 220 (40%) 19 (51%) 201 (39%) 0.21

Fibroadenoma 216 (39%) 12 (32%) 204 (40%) 0.43

No diagnosis or missing 113 (21%) 6 (16%) 107 (21%) 0.61

FA = fibroadenoma.