ORIGINAL ARTICLE

Underestimation of papillary breast lesions by core biopsy:correlation to surgical excision

Stamatia Destounis • Posy Seifert • Patricia Somerville •

Philip Murphy • Renee Morgan • Andrea Arieno •

Wende Logan Young

Received: 11 January 2012 / Accepted: 7 March 2012

� The Japanese Breast Cancer Society 2012

Abstract

Background The purpose of this study was to evaluate

papillary lesions of the breast diagnosed at needle core

biopsy and the outcomes of follow-up imaging and surgical

findings.

Methods Retrospective review of 13,806 needle core

biopsies performed from 2004 to 2010 revealed 352

patients with 368 papillary lesions; 137 of these lesions

underwent surgical excision and 215 had a minimum of

2 year imaging follow-up. Outcomes of imaging follow-up

and underestimation of carcinoma by comparison to sur-

gical excision, as applicable, were determined. Patient

demographics, clinical presentation, BI-RADS� breast

density, palpability, biopsy methods, number of specimens,

and pathology were recorded. A two-tailed Fisher exact test

was used to assess associations between biopsy techniques

and the results of surgical excision as well as the associa-

tion between palpability and the results of surgical exci-

sion. SAS� V 9.1.3 was used to perform the calculations.

Results One hundred and thirty-seven lesions proceeded

to surgical excision. A total of 28 lesions were underesti-

mated; 21 of these were underestimated carcinomas. When

comparing biopsy devices in the underestimated group,

64 % were biopsied with directional vacuum-assisted

devices and 36 % with automated large core devices

(p \ 0.0706). In total 18 % of the palpable lesions were

underestimated, whereas 17 % of the nonpalpable lesions

were underestimated (p \ 0.6560); this was not a signifi-

cant difference.

Conclusions Overall, carcinoma was underestimated at

needle biopsy in 6 % (21/368) of papillary lesions diag-

nosed when compared to surgical excision. In cases that

underwent excision, 34 % (47/137) revealed carcinoma;

45 % (21/47) of these were underestimated carcinomas.

Lesion palpability and biopsy method did not affect

underestimation in this study population.

Keywords Breast imaging � Papillary lesions �Underestimation � Core needle biopsy

Introduction

Single or multiple papillary lesions are uncommon, found

in up to 5 % of biopsies [1], and can be found in any area

of the breast. These lesions at core biopsy result in a wide

range of findings, including benign without atypia, benign

with atypia, papillary ductal carcinoma in situ (DCIS), or

papillary invasive malignancy. Reports describe that pap-

illary malignancies can have areas of benign features,

whereas benign papillomas can have areas of mitoses

resembling atypia and carcinoma [2]. As a consequence,

papillary findings at core needle biopsy (CNB) remain

controversial and excision is not always recommended.

Studies have shown that papillary lesions found at CNB

can be underestimated carcinomas, suggesting that all

papillary lesions diagnosed at CNB should be excised

[3–12]. Conversely, other studies suggest that follow-up of

benign without atypia at CNB is a reasonable management

choice for the patients [13–18]. Our practice protocol is to

follow lesions diagnosed as benign without atypia at CNB

in asymptomatic patients rather than recommend open

surgical biopsy (OSB). OSB is recommended for lesions

diagnosed with atypia, malignancy, or if the patient is

S. Destounis (&) � P. Seifert � P. Somerville � P. Murphy �R. Morgan � A. Arieno � W. L. Young

Elizabeth Wende Breast Care, LLC., 170 Sawgrass Dr.,

Rochester, NY 14620, USA

e-mail: sdestounis@ewbc.com

123

Breast Cancer

DOI 10.1007/s12282-012-0361-2

symptomatic, such as having persistent spontaneous dis-

charge, enlarging palpable lump, or worsening pain. The

purpose of this review was to identify the incidence and

circumstance of pathological underestimation of carcino-

mas at CNB initially diagnosed as papillary lesions.

Materials and methods

This institutional review board approved retrospective

review was conducted in compliance with Health Infor-

mation Portability and Accountability Act. Patient consent

was not required. A total of 13,806 CNBs from January

2004 to December 2010 were identified from an electronic

medical records database at an outpatient facility special-

izing in breast imaging. A comprehensive chart review

identified 352 patients with 368 papillary lesions, which

included papilloma, intraductal papilloma, atypical papil-

loma, papilloma with atypia, papillary neoplasm, papil-

loma with DCIS, papillary DCIS, encapsulated papillary

carcinoma, and solid papillary carcinoma, that fit the study

criteria. Further review demonstrated 137 lesions which

underwent surgical excision and 215 who had imaging

follow-up. If surgical excision was not performed, a min-

imum of 2 years of imaging follow-up was required for

inclusion in the study analysis. In the surgically excised

group, further analysis concentrated on underestimation.

Overestimation was not analyzed as the lesions were

determined to be completely excised at CNB. Study

information collected included patient demographics,

BI-RADS� breast density, lesion characteristics, core biopsy

information, histopathological results, and results of fol-

low-up.

Core needle biopsy procedure

Core needle biopsy procedures were performed by breast

imaging specialists with years of experience ranging from

6 to 35 years (average 12) utilizing stereotactic, ultrasound,

or magnetic resonance imaging (MRI) guidance. Auto-

mated large core (ALC) devices with 14-gauge needles or

directional vacuum-assisted (DVA) devices with 14-, 12-,

or 9-gauge needles were used. Image guidance, biopsy

device, needle size, and number of specimens were selec-

ted on the basis of considerations that included the char-

acteristics of image findings such as type (mass,

calcifications, architectural distortion), size, location, and

palpability, preference of the radiologist, and health con-

cerns of the patient. All CNB procedures were performed

using the facility’s standard operating procedures. Stereo-

tactic core biopsy was performed using a Hologic DSM

prone stereotactic table (LoRad; Hologic, Bedford, MA,

USA). Ultrasound-guided biopsy was performed utilizing

one of the following units: Siemens Antares (Siemens

Medical Solutions USA, Inc., Malvern, PA, USA), Philips

iU22 system (Philips Healthcare, Bothell, WA, USA), or

GE Logiq (GE Healthcare, Milwaukee, WI, USA). The

MRI-guided biopsies were performed on a 1.5-T GE

Echospeed (GE Healthcare) or a 1.5-T Siemens Symphony

(Siemens Medical Solutions USA, Inc.) equipped with a

dedicated surface breast coil. DVA biopsy was performed

using a Suros ATEC or Suros Celero (Hologic, Inc.), as

well as the Mammotome ST system (Ethicon Endo-Sur-

gery, Cincinnati, OH, USA). ALC biopsy utilized either

Manan (Manan Medical Products, Wheeling, IL, USA) or

Bard (CR Bard, Inc., Murray Hill, NJ, USA).

CNB pathology was submitted to one of three local

laboratories for interpretation, depending on insurance

carrier of the patient. The majority, however, were sent to

one lab where six breast pathologists reviewed all cases.

The experience of the pathologists ranged from 4 to

30 years (average 17 years) where a daily conference for

case review was conducted.

All CNB results were reviewed by the attending radi-

ologist for concordance with imaging findings. Recom-

mendation for OSB or follow-up depended upon the core

pathology results and clinical presentation. If the core

pathology results were discordant with imaging and/or

pathology revealed atypia or malignancy, OSB was rec-

ommended. For benign without atypia findings in asymp-

tomatic patients, the patient was instructed to return in

6 months for follow-up examination. For study purposes,

an asymptomatic patient was defined as ‘‘no symptoms;

such as enlarging lump, enhancing mass, or nipple dis-

charge.’’ A subset of patients was recommended for OSB

despite benign findings as a result of clinical presentation

such as continued nipple discharge. Surgical results for

these patients were requested and then also compared with

imaging findings and CNB pathology results by the

attending radiologist.

Statistical analysis

To assess the associations between biopsy techniques and the

results of surgical excision, a two-tailed Fisher exact test was

used. The same test was used for the association between

palpability and the results of surgical excision. SAS� V 9.1.3

was used to perform the calculations. An alpha (a) level of

\0.05 was used to determine significance.

Results

One hundred and thirty-seven lesions diagnosed at CNB

proceeded to surgical excision. Two hundred and fifteen

patients with benign findings at CNB were recommended

Breast Cancer

123

for imaging follow-up, which could include mammo-

graphic, sonographic, or MRI imaging. Follow-up ranged

from 2 to 6 years, with an average of 3.8 years. Three

ipsilateral cancer diagnoses and one contralateral cancer

diagnosis have been made in follow-up, all of which were

in different quadrants from the papillary lesion. One

atypical finding and three benign findings in the ipsilateral

breast have been made. Additionally, one atypical finding

and five benign findings in the contralateral breast have

been made.

Patient demographics and clinical characteristics, as

well as imaging findings of the 137 lesions that proceeded

to surgical excision, are shown in Tables 1 and 2. CNB was

performed on all 137 lesions. DVA was utilized for 78 %

(n = 107) of biopsies with a 9-, 12-, or 14-gauge needle

with an average of 7.8 samples taken. The remainder

(22 %, n = 30) utilized an ALC device with a 14-gauge

needle, with an average of 3.1 samples taken. Ultrasound-

guided biopsy was performed for 53 % (n = 73); 43 %

(n = 59) of the lesions were biopsied stereotactically and

MRI guidance was used for 4 % (n = 5). At CNB, 53 of

the lesions had a benign diagnosis, 52 were atypical, and 32

malignant.

Upon surgical excision, a total of 81/137 (59.1 %)

diagnoses remained the same; 44/81 (54.3 %) of these were

benign, 11/81 (13.6 %) atypical, and 26/81 (32.1 %)

malignant. Comparison between CNB and OSB diagnoses

is shown in Table 3. Of the 53 benign lesions diagnosed at

CNB that were excised, 9 had an upgraded diagnosis; 7 to

atypia and 2 to malignancy (both in situ carcinoma). Fifty-

two atypical lesions were diagnosed at CNB and of those,

19 upgraded to carcinoma; 13 in situ (Figs. 1, 2) and 6

invasive. In total, 28 lesions were upgraded at surgical

excision; 21 (15 %) of these were upgraded to carcinoma.

Underestimated lesions

Of the 20.4 % (n = 28/137) that had an upgraded diag-

nosis at surgery, 25 % (n = 7) went from benign to atyp-

ical, 7 % (n = 2) benign to malignant, and 68 % (n = 19)

atypical to malignant. Eight of the 28 underestimated

lesions were palpable on physical examination and 11

reported discharge upon presentation. Of the 28 underes-

timated lesions, 1 occurred in a fatty breast (3.57 %), 13 in

scattered fibroglandular tissue (46.43 %), 7 in heteroge-

neously dense tissue (25 %), and 7 in extremely dense

tissue (25 %). Eighteen lesions (64 %) were biopsied with

DVA with an average of 9 samples (range 3–12) and 10

(36 %) with ALC, with an average of 3 samples retrieved

(range 1–5).

Specifically, of the 21 underestimated malignancies, 16

lesions presented as a mass at imaging, 10 were visualized

Table 1 Demographic and clinical characteristics of OSB cohort

(n = 130 patients, 137 lesions)

Characteristic Number of subjects Percent

Age, years (n = 130)

18–35 4 3.08

36–45 21 16.15

46–55 24 18.46

56–65 35 26.92

66–75 31 23.85

76? 15 11.54

Reason for presentation (n = 137)

Screening 38 27.74

Diagnostic 99 72.26

Breast density (n = 129)a

Fatty 6 4.62

Scattered 49 37.69

Heterogeneously dense 43 33.08

Extremely dense 31 23.85

Palpability (n = 137)

Nonpalpable 88 64.23

Palpable 45 32.85

Not performed 4 2.92

Discharge (n = 137)

Present 48 35.03

Not present 89 64.96

a Breast density was not available for one patient

Table 2 Imaging findings

Characteristic Number of subjects Percent

Imaging findings

Mass 92 67.15

Calcifications 7 5.12

Mass with calcium 5 3.65

Architectural distortion 1 0.73

Dilated duct/ductal enhancement 2 1.46

MRI enhancement 1 0.73

Filling defects 29 21.17

Total 137 100.00

Table 3 Comparison of histological findings from CNB and OSB

CNB findings OSB findings

Benign

without atypia

Benign

with atypia

Malignant

papillary

Benign without atypia 44 7 2

Benign with atypia 22 11 19

Malignant papillary 4 2 26

Breast Cancer

123

on both mammography and ultrasound; 4 of the 21 malig-

nancies were palpable. Three presented as filling defects

and all were seen on ductography only. Two calcifications

were seen on mammography only. Six of the patients pre-

sented with discharge. When looking at the breast density

composition of the 21 underestimated malignancies, 1 was

found in fatty tissue, 11 in scattered, 4 in heterogeneously

dense, and 5 in extremely dense. Biopsy methods of the

underestimated cancers are shown in Table 4.

Benign papillary lesions without atypia

with and without surgical excision

A total of 215 patients with a benign without atypia CNB

diagnosis were followed up with imaging, and a total of 53

with a benign without atypia diagnosis at CNB were rec-

ommended for surgical excision. In the group of patients

with imaging follow-up, the average patient age was

56 years (range 29–88). Palpability was reported in 209

cases with 19 lesions being palpable and 190 nonpalpable.

Discharge was present in 112 cases. In the group of patients

who had surgery the average age was 54 (range 18–78).

Fifteen lesions were palpable and 36 were nonpalpable; 2

did not have physical exam. Discharge was reported in 34

cases. A breakdown of lesion type for both groups is shown

in Table 5.

Discussion

Reported underestimation rates in the literature range from

6 to 39 % [19]; our findings are in line with such reports as

within the OSB cohort, we found a 20 % (28/137) under-

estimation rate. Our study findings indicated that benign

papillary lesions with atypia diagnosed at CNB had the

lowest accuracy rate (21 %), as a majority (68 %) of the

upgrades in our study came from this group. Many other

studies have had similar findings in regards to papillary

lesions with atypia having a high association with malig-

nancy at OSB [1, 3, 11, 13, 14, 17, 19]. Benign papillary

lesions at CNB had the highest accuracy rate (83 %), fol-

lowed by malignant papillary lesions (81 %).

Fig. 1 A 77-year-old presented

for screening mammogram

(a, b). Mass-like areas were

identified (a–c, arrow).

Ultrasound (d) revealed a

hypoechoic mass. Ultrasound-

guided 14-gauge ALC was

performed, which revealed

papillary neoplasm. OSB

pathology revealed DCIS

Breast Cancer

123

When evaluating several variables to determine a cor-

relation to underestimation of carcinoma, we looked at

biopsy method, lesion palpability, and biopsy device. We

did not find any of these variables to be indicators of

underestimation in this study population. Ciatto et al. [20]

published a study of accuracy and underestimation of

malignancy with over 4,000 biopsies and concluded that

image-guidance method or lesion palpability did not sub-

stantially affect accuracy.

Of the 59 stereotactic guided biopsies that went to sur-

gery, underestimation was noted in 15 (25 %). Of the 73

ultrasound-guided biopsies, underestimation was noted in

12 (16 %). One underestimation was noted of the 5 MRI-

guided biopsies (20 %). Image guidance did not seem to be

a significant factor in underestimation. Arora et al. [19]

also determined that image guidance did not seem to be

associated with underestimation.

Palpability also was not a factor influencing underesti-

mation. Eighteen percent (n = 8/45) of the palpable lesions

were underestimated, whereas 17 % (n = 15/88) of the

nonpalpable lesions were underestimated (p \ 0.6560);

this was not a significant difference.

When comparing biopsy devices in the underestimated

group, we found that 64 % were biopsied with DVA and

36 % with ALC (p \ 0.0706). This does not support that

biopsy device is an influencing factor of underestimation.

When reviewing the underestimated cancers, we found that

52.38 % (n = 11/21) were biopsied with DVA and

Fig. 2 A 76-year-old presented with a palpable mass in the left breast

at 5 o’clock. A mass was identified on mammography (a LCC;

b LMLO). Ultrasound revealed a partially cystic mass (c). Ultra-

sound-guided 12-gauge DVA revealed papillary neoplasm. Surgical

excision yielded multiple foci of DCIS

Table 4 Underestimated cancerous lesions and CNB methods

Characteristic Number of subjects Percent

Stereo 10 47.62

9 DVA 2 20.00

12 DVA 5 50.00

14 ALC 3 30.00

Ultrasound 10 47.62

12 DVA 3 30.00

14 ALC 7 70.00

MRI 1 4.76

9 DVA 1 100.00

Total 21 100.00

Table 5 Lesion type of benign without atypia lesions with and

without surgery

Characteristic Benign without

surgery

Benign with

surgery

Imaging findings

Mass 93 25

Calcifications 37 2

Mass with calcium 7 1

Architectural distortion 5 1

Dilated duct/ductal enhancement 2 1

MRI enhancement 1 0

Filling defects 70 23

Total 215 53

Breast Cancer

123

47.62 % (n = 10/21) with an ALC device. More specifi-

cally, we found a higher percentage of underestimations in

the group with lesions biopsied under ultrasound guidance

with a 14-gauge ALC device, which had an average of 3

samples taken (range 1–5). A previously published work

from our facility reviewed outcomes of 9- and 12-gauge

CNBs and found no difference. The analysis found 8.00 %

underestimation for the 12-gauge group and 8.24 % for the

9-gauge group. These results continue to support our cur-

rent practice of utilizing either needle gauge [21]. Other

studies have also evaluated the difference between devices

and found a range of results. Houssami and colleagues [22]

reported that (11-gauge) vacuum-assisted devices were

significantly less prone to underestimation than (14-gauge)

automated devices and go on to say that using larger core-

needle devices does reduce underestimation to an extent,

but does not eliminate the problem completely. Addition-

ally Philpotts et al. [23] reported no significant difference

between outcomes of (ultrasound-guided) DVA verses

ALC in terms of missed cancers, underestimations, com-

plications, or need for immediate second biopsy. Liberman

et al. [17] found that ALC was accurate with benign lesions

but still stated that OSB was indicated for atypia and

papillary DCIS, which supports our findings.

A large population of patients in this study were rec-

ommended for follow-up rather than OSB when a benign

CNB diagnosis was made. Overall, this proved to be the

correct approach for these patients within the follow-up

cohort; however, we did have three patients with sub-

sequent cancer diagnoses in the ipsilateral breast. These

cancers were located in different quadrants of the breast

and thus did not arise from the papillary lesion. Several

published studies have determined that follow-up of benign

lesions without atypia is adequate and excision is not

necessary. Syndor et al. [13] reported that 25 benign pap-

illary lesions diagnosed at CNB had a minimum of

24 month follow-up with no interval change. Similarly,

Rosen et al. [1] reported that 29 patients with mammo-

graphic follow-up were all found to be stable or had

decreased lesion size during the follow-up period. Mercado

and colleagues [12] found benign without atypia lesions

upgraded to atypia or cancer in 21 % of cases (9/42) and

recommend surgical excision because of the number of

upgrades. Our findings demonstrated that, for the majority

of benign papillary lesions without atypia, imaging follow-

up was appropriate.

Conclusion

Overall, we found a 20 % underestimation rate in the

patient cohort that had surgical excision. As a result of the

high upgrade rate for atypical papillary lesions we feel it is

necessary to recommend surgical excision when such a

diagnosis is made at CNB. Benign papillary lesions in

asymptomatic patients do not need to be surgically excised,

as we found these lesions at CNB had the highest accuracy

rate (83 %); however, close imaging surveillance is

required. Biopsy device, biopsy method, and lesion pal-

pability were not found to be significant indicators of

underestimation.

Acknowledgments The authors wish to thank Karen Ahlstrom,

Anne Horan, Rena Blair, Mary Ventrella, Dawn Riedy, MD and

David Cavanaugh, MA for their contributions to this manuscript.

References

1. Rosen EL, Bentley RC, Baker JA, Soo MS. Imaging-guided core

needle biopsy of papillary lesions of the breast. AJR. 2002;179:

1185–92.

2. Rosen PP, Hoda SD. Breast pathology: diagnosis by needle core

biopsy. 2nd ed. Philadelphia: Lippincott Williams & Wilkins;

2006.

3. Tseng HS, Chen YL, Chen ST, Wu YC, Kuo SJ, Chen LS, et al.

The management of papillary lesion of the breast by core needle

biopsy. EJSO. 2009;35:121–4.

4. Bernik SF, Troob S, Ying BL, Simpson SA, Axelrod DM, Siegel

B, et al. Papillary lesions of the breast diagnosed by core needle

biopsy: 71 cases with surgical follow-up. Am J Surg. 2009;

197:473–8.

5. Jaffer S, Nagi C, Bleiweiss IJ. Excision is indicated for intra-

ductal papilloma of the breast diagnosed on core needle biopsy.

Cancer. 2009;115:2837–43.

6. Skandarajah AR, Field L, Mou AYL, Buchanan M, Evans J, Hart

S, et al. Benign papilloma on core biopsy requires surgical

excision. Ann Surg Oncol. 2008;15(8):2272–7.

7. Shin HJ, Kim HH, Kim SM, Yang HR, Sohn JH, Kwon GY, et al.

Papillary lesions of the breast diagnosed at percutaneous sono-

graphically guided biopsy: comparison of sonographic features

and biopsy methods. AJR. 2008;190:630–6.

8. Rizzo M, Lund MJ, Oprea G, Schniederjan M, Wood WC,

Mosunjac M. Surgical follow-up and clinical presentation of 142

breast papillary lesions diagnosed by ultrasound-guided core-

needle biopsy. Ann Surg Oncol. 2008;15(4):1040–7.

9. Liberman L, Tornos C, Huzjan R, Bartella L, Morris EA, Der-

shaw DD. Is surgical excision warranted after benign, concordant

diagnosis of papilloma at percutaneous breast biopsy? AJR. 2006;

186:1328–34.

10. Ueng SH, Mezzetti T, Tavassoli FA. Papillary neoplasms of the

breast: a review. Arch Pathol Lab Med. 2009;133:893–907.

11. Cheng TY, Chen CM, Lee MY, Lin KJ, Hung CF, Wang PS,

et al. Risk factors associated with conversion from nonmalignant

to malignant diagnosis after surgical excision of breast papillary

lesions. Ann Surg Oncol. 2009;16:3375–9.

12. Mercado CL, Hamele-Bena D, Oken SM, Singer CI, Cangiarella

J. Papillary lesions of the breast at percutaneous core-needle

biopsy. Radiology. 2006;238:801–8.

13. Syndor MK, Wilson JD, Hijaz TA, Massey HD, Shaw de Paredes

ES. Underestimation of the presence of breast carcinoma in

papillary lesions initially diagnosed at core-needle biopsy.

Radiology. 2007;242(1):58–62.

14. Renshaw AA, Derhagopian RP, Tizol-Blanco DM, Gould EW.

Papillomas and atypical papillomas in breast core needle biopsy

specimens. Am J Clin Pathol. 2004;122(2):217–21.

Breast Cancer

123

15. Reynolds HE. Core needle biopsy of challenging benign breast

conditions: a comprehensive literature review. AJR. 2000;174:

1245–50.

16. Ahmadiyeh N, Stoleru MA, Raza S, Lester S, Golshan M.

Management of intraductal papillomas of the breast: an analysis

of 129 cases and their outcome. Ann Surg Oncol. 2009;16:

2264–9.

17. Liberman L, Bracero N, Vuolo M, Dershaw DD, Morris EA,

Abramson AF, et al. Percutaneous large-core biopsy of papillary

breast lesions. AJR. 1999;172:331–7.

18. Kim MJ, Kim EK, Kwak JY, Son EJ, Park BW, Kim SI, et al.

Nonmalignant papillary lesions of the breast at US-guided

directional vacuum-assisted removal: a preliminary report. Eur

Radiol. 2008;18(9):1774–83.

19. Arora N, Hill C, Hoda S, Rosenblatt R, Pigalarga R, Tousimis E.

Clinicopathologic features of papillary lesions on core needle

biopsy of the breast predictive of malignancy. Am J Surg. 2007;

194:444–9.

20. Ciatto S, Houssami N, Ambrogetti D, et al. Accuracy and

underestimation of malignancy of breast core needle biopsy: the

Florence experience of over 4000 consecutive biopsies. Breast

Cancer Res Treat. 2007;101(3):291–7.

21. Destounis S, Skolny M, Morgan R, Arieno A, et al. Rates of

pathological underestimation for 9 and 12 gauge breast needle

core biopsies at surgical excision. Breast Cancer. 2011;18:42–50.

22. Houssami N, Ciatto S, Ellis E, Ambrogetti D. Underestimation of

malignancy of breast core-needle biopsy: concepts and precise

overall and category-specific estimates. Cancer. 2007;109:

487–95.

23. Philpotts L, Hooley R, Lee C. Comparison of automated versus

vacuum-assisted biopsy methods for sonographically guided core

biopsy of the breast. AJR. 2003;180(2):347–51.

Breast Cancer

123