Presented in part at the 101st Annual Meeting of the United States and Canadian Academy of Pathology, Vancouver, March 2012.
Pathological Evaluation of Nipple-Sparing Mastectomies with Emphasis on Occult Nipple Involvement: The Weill-Cornell Experience with 325 Cases
Article first published online: 18 JAN 2014
© 2014 Wiley Periodicals, Inc.
The Breast Journal
Volume 20, Issue 1, pages 15–21, January/February 2014
How to Cite
Eisenberg, R. E.K., Chan, J. S.Y., Swistel, A. J. and Hoda, S. A. (2014), Pathological Evaluation of Nipple-Sparing Mastectomies with Emphasis on Occult Nipple Involvement: The Weill-Cornell Experience with 325 Cases. The Breast Journal, 20: 15–21. doi: 10.1111/tbj.12199
- Issue published online: 18 JAN 2014
- Article first published online: 18 JAN 2014
- breast cancer;
Nipple-sparing mastectomy (NSM) is an increasingly utilized surgical option in managing breast carcinoma; however, data on malignant involvement of a separately submitted nipple margin are scant. Consecutive NSM, including those performed for therapeutic and prophylactic purposes, over a 4-year period (2007–2011), were studied. A separately submitted nipple margin was evaluated by permanent H&E preparations and via frozen section evaluation whenever requested. 325 consecutive NSM specimens, 208 (64%) therapeutic-NSM, and 117 (36%) prophylactic-NSM were studied. All nipples were clinically unremarkable. 86% (179/208) of nipple margins from therapeutic-NSM and 100% (117/117) from prophylactic-NSM showed no histopathologic abnormality. 14% (29/208) of nipple margins from therapeutic-NSM and no nipple margin from prophylactic-NSM showed malignancy. Frozen section evaluation was performed in 188/325 NSM (58%) with a sensitivity of 64% and specificity of 99%. Central tumor location and stage N2/N3 lymph node status were significantly associated with nipple margin positivity (χ2 ≤ 0.05). Subsequent nipple resection was performed in 69% (20/29) of nipple margin-positive cases with residual malignancy found in 40% (8/20, including three cases of invasive carcinoma). In a mean follow-up of 33 months, one invasive carcinoma recurred in the “saved” nipple, 36 months after therapeutic-NSM. 14% (29/208) of nipple margins in therapeutic-NSM and no nipple margin (0/117) in prophylactic-NSM showed malignancy. Central tumor location and N2/N3 stage were significantly associated with nipple margin positivity (χ2 ≤ 0.05).
The surgical management of breast carcinoma has evolved to include operative techniques which offer improved cosmetic outcomes. In 1991, skin-sparing mastectomy, which allowed immediate reconstruction and resulted in comparable local recurrence rates to simple mastectomy, was described [1-3]. In 2003, Gerber et al. described nipple-sparing mastectomy (NSM), a modification of skin-sparing technique, which permitted conservation of nipple areolar-complex (NAC) . In recent years, NSM has become an increasingly utilized procedure, for patients with breast carcinoma and for those seeking prophylactic treatment, such as in BRCA germline mutation carriers .
Despite its popularity, the oncological safety of NSM remains of concern. Such worry stems from the potential presence of terminal-duct lobular units (TDLU) in the conserved portion of the nipple. TDLU is generally regarded as the initial site of mammary malignancies and are reportedly present in 1–17% of nipples, although its true prevalence remains controversial [6, 7]. Therapeutic radiation has been combined with NSM to enhance oncological safety; although it confers an increased risk of nipple necrosis . Overall, local recurrence in NSM is observed in 3–6% of patients at 5 years, a rate similar to conventional mastectomies .
Several studies have evaluated the nipple in conventional mastectomies to predict, via inference, the incidence of its “occult” involvement by malignancy. The results of such studies differ widely, and these differences are attributable to the variable methods of patient selection and techniques employed in nipple evaluation [10-13].
In this retrospective study, we primarily sought to study histopathologic findings in the separately submitted retro-areolar nipple margin in NSM as a proxy for occult involvement of nipple by malignancy.
Materials and Methods
All consecutive NSMs performed in our Medical Center over a 54-month period (January 2007–July 2011) were identified via a retrospective review of our departmental data base. All NSM had been performed either as therapeutic operations for breast carcinoma or as prophylactic procedures in high-risk patients.
A separately submitted retro-areolar cross-section representing the “true” nipple margin was assessed in all cases. This margin was evaluated either by a routine H&E-stained preparation or, whenever requested by intraoperative frozen section examination with subsequent preparation of a permanent control slide. Request for frozen section evaluation was largely based on clinical and radiological findings in a case. All nipple margins were harvested from the nipple base and were submitted en face, but otherwise unoriented, for pathological evaluation. The nipple margins were essentially of similar size (Fig. 1a), approximating the dimension of a “dime,” i.e. 10-cent United States coin (18 × 18 × 1 mm, ±3 mm).
Positivity of nipple margin was defined as involvement by one of the following neoplastic processes: invasive carcinoma of any type, ductal carcinoma in situ (DCIS), and lobular carcinoma in situ (LCIS). The inclusion of LCIS may (or may not) be unanimously regarded as “malignant” involvement of a margin (vide infra).
Archived slides from all nipple margin-positive, and selected nipple margin-negative, cases were reviewed. Pertinent clinicopathological features including type, grade, size, and location of tumor; lymphovascular invasion; and lymph node involvement were tabulated. Results of estrogen receptor, progesterone receptor, and HER2 testing were recorded. Follow-up information on nipple margin-positive cases was obtained.
Statistical analyses were performed using Microsoft Excel (Microsoft, Redwood City, WA) software. Central tendency of each distribution of clinicopathological features were described using mean and SE. Chi-squared tests were then performed and contingency tables were generated for each feature. Associations between clinicopathological data and nipple margin-positive cases were deemed statistically significant at χ2 ≤ 0.05.
The study was approved by Institutional Review Board.
Three hundred and twenty-five (325) NSM were performed at our institution in a 54-month period (January 2007–July 2011). The specimens had been obtained from 215 patients (211 female, four male). Mean age of patients was 50 (range 25–85). NSM were performed for therapeutic purposes in 208 (64%), all of whom had been previously diagnosed with breast carcinoma; or in high-risk individuals for prophylactic purposes in 117 (36%) patients. Bilateral NSM were performed in 110 (51%) cases. A separate retro-areolar cross-section of the nipple margin was submitted for all cases by each of three surgeons.
Preoperatively, nipples in all NSM were clinically unremarkable. No nipple margin in a prophylactic-NSM showed malignancy (0/117). Twenty-nine of 208 (14%) nipple margins in therapeutic-NSM showed occult malignancy: DCIS in 16/208 (8%), LCIS in 5/208 (2%), invasive ductal carcinoma in 4/208 (2%), and invasive lobular carcinoma in 4/208 (2%; Fig. 1b–d).
An intraoperative consultation with frozen section examination of nipple margin was sought in 188 (58%) specimens. The majority (159, or 85%) of frozen sections were requested in the therapeutic-NSM group. One case was false-suspicious—having been diagnosed as “suspicious for carcinoma” on frozen section but which showed tangentially sectioned inactive lactiferous duct on permanent control. Frozen section examination had a sensitivity of 64% (false-negative: 9) and specificity of 99% (false-suspicious: 1). The rate of false-negative and false-suspicious diagnoses were 4.8% and 0.5% respectively. Five of nine false-negative nipple margin cases showed lactiferous ducts involved by LCIS (Fig. 1e and f). The initial breast pathology in the latter set included three invasive lobular carcinoma, one invasive ductal carcinoma, and one DCIS. Three false-negative nipple margin cases were positive for malignancy on permanent control slide only, including two cases of invasive duct carcinoma and one case of DCIS. The remaining false-negative case showed rare invasive lobular cells (confirmed by positive cytokeratin AE1/3 immunohistochemical staining) in a patient status-post neoadjuvant chemotherapy. Sensitivity of frozen section examination was 80% (false-negative: 4) and the false-negative rate was 2.1% when cases of margin involvement by LCIS were excluded. Positive-predictive value (PPV) and negative-predictive value (NPV) of FSE were 94% and 95% respectively.
Key clinical and pathological features of the cases (including location of primary tumor, its type, size, and grade; lymphovascular and lymph node involvement; estrogen receptor; and HER2 status) are detailed in Table 1. We were unable to accurately assess distance from tumor to nipple in a reliable manner due to the retrospective nature of this study; however, its location (central or quadrant-wise) was recorded. The clinical and pathologic parameters that were significantly associated with positive nipple margin in therapeutic-NSM are also shown in Table 1. The location of primary tumor had a nonrandomly distributed relationship with positive nipple margin (χ2 < 0.008). Central location of tumor demonstrated a statistically significant association with positive nipple margin versus location in other quadrants. There was a statistically significant association of positive nipple margin status with nodal involvement stage N2 and N3 (metastasis to four or more lymph nodes; χ2 = 0.05). Age of patient; type, grade, and size of tumor; lymphovascular involvement; estrogen receptor status; and HER2 status were not significantly associated with nipple margin involvement (χ2 > 0.05).
|Clinicopathologic characteristic||n||Total%||Nipple margin benign n (%)||Nipple margin occult malignancy n (%)||χ2|
|<50||122||59||109 (89)||13 (11)||0.131|
|≥50||86||41||70 (81)||16 (19)|
|Location of primary tumor||181|
|UOQ||79||44||72 (91)||7 (9)||0.008b|
|LOQ||28||15||25 (90)||3 (10)|
|UIQ||14||8||11 (79)||3 (21)|
|LIQ||16||9||15 (94)||1 (6)|
|Central||18||10||10 (56)||8 (44)|
|Multicentric||26||14||19 (42)||7 (58)|
|IDC||115||55||100 (87)||15 (13)||0.255|
|ILC||27||13||20 (74)||7 (26)|
|Other types INV CA||8||4||8 (100)||0 (0)|
|DCIS||57||28||50 (88)||7 (12)|
|1||30||20||24 (80)||6 (20)||0.424|
|2||74||50||62 (84)||12 (16)|
|3||45||30||41 (91)||4 (9)|
|NG1||3||5||2 (67)||1 (33)||0.318|
|NG2||31||55||29 (94)||2 (6)|
|NG3||23||40||19 (83)||4 (17)|
|Present||33||16||27 (82)||6 (18)||0.534|
|Absent||168||84||145 (86)||23 (14)|
|Lymph node status||180|
|N0||133||74||115 (86)||18 (14)||0.054b|
|N1||37||20||34 (92)||3 (8)|
|N2/3||10||6||6 (60)||4 (40)|
|Size (INV CA)a||137|
|≤2 cm||107||78||92 (86)||15 (14)||0.622|
|>2 cm||30||22||23 (77)||7 (23)|
|Estrogen receptor (INV CA)a||136|
|Positive||113||83||93 (82)||20 (18)||0.667|
|Negative||23||17||21 (91)||2 (9)|
|HER2/neu (INV CA)a||131|
|Positive||17||14||17 (100)||0 (0)||0.415|
|Negative||114||86||92 (81)||22 (19)|
|N||208||100||179 (86)||29 (14)|
Subsequent nipple resection was performed in 32 (10%) of the NSM cases (Table 2). Of 29 nipple margin-positive therapeutic-NSM, 20 (69%) had nipple resection (13 at the time of surgery and seven in a second procedure). Residual malignancy was found in eight of 20 (40%) of these nipple resections, including two cases of invasive ductal carcinoma, one case of invasive lobular carcinoma, and five cases of DCIS (one with clinically undetected Paget disease).
|Nipple margin||n||Positive n (%)||Negative n (%)|
|Subsequent nipple resection|
|Positive||10||8 (20)||2 (80)|
|INV CA||4||3 (25)||1 (75)|
|DCIS||6||5 (17)||1 (83)|
|Negative||22||12 (45)||10 (55)|
To our knowledge, nine of 29 nipple margin-positive cases did not have a subsequent nipple resection. Four of five patients with classical LCIS in nipple margin did not undergo nipple resection. One patient with LCIS in nipple margin had nipple resection without residual abnormality. Four of remaining five nipple margin-positive patients without subsequent nipple resection had no clinical evidence of recurrence after a mean 10-month follow-up (range 1–24 months). One patient was lost to follow-up.
Subsequent nipple resection was performed in 12 nipple margin-negative cases for various clinical, technical, and cosmetic reasons (nine at the time of surgery and three in a second procedure). Ten of 12 nipple resection specimens showed no diagnostic abnormality. Clinical follow-up was available for only two of these 10 patients; one patient underwent nipple resection for “close” margin and one patient for cosmetic purposes. Two nipple margin-negative cases showed occult malignancy in the nipple resection. One case showed DCIS in NAC, although the separately submitted nipple margin had been “negative” at frozen section and on permanent control. This finding of DCIS may represent a “skip” lesion, and as such is regarded as a false-negative case. The other nipple resection was performed for local recurrence of invasive duct carcinoma in NAC 3 years after therapeutic-NSM.
Nipple-sparing mastectomy is an increasingly exercised surgical option for the therapeutic management of breast carcinoma as well as for the prophylactic management of high-risk patients. NSM offers improved cosmetic outcome through conservation of the breast “envelope,” and ready reconstruction . However, preservation of nipple with this technique is essentially cosmetic, and does not usually allow for maintenance of either erectile function or tactile sensation [15, 16].
Patients in whom NSM is generally not considered an option due to increased chances of nipple necrosis include those who are obese or have breast ptosis, and those with a history of smoking, autoimmune disease, or radiation to the breast . Additional clinical and pathologic parameters which have been reported to influence nipple involvement by malignancy include “short” tumor to nipple distance, “large” size, central location, multifocality or multicentricity, lymphovascular and lymph node involvement, and positivity for HER2 [18-21].
In our study of 325 consecutive NSM, a separately submitted retro-areolar cross-section of the nipple margin was submitted for each case. The plane of the nipple margin surgically taken in this study was located ~3 mm (range 2–5 mm) below the NAC (Fig. 2) . No nipple margin (0/117) from prophylactic-NSM harbored occult malignancy. However, 29 of 208 (14%) nipple margins from therapeutic-NSM showed occult involvement of the nipple by malignancy. The most common malignancy in the nipple margin was DCIS, identified in 55% of nipple margin-positive cases. Central tumor location and metastatic involvement of four or more regional lymph nodes (i.e. N2/N3) were significantly associated with nipple margin involvement (χ2 ≤ 0.05). The statistical significance (or otherwise) of the association of various clinical and pathological factors associated with occult involvement of nipple with malignancy is shown in Table 1.
Frozen section evaluation of nipple margin in therapeutic-NSM group proved to be a specific (99%) but nonsensitive (64%) test in our series. In this series, frozen section slides of nipple margin were largely nonproblematic for interpretation (Fig. 1); however, the rather high false-negative rate of the test (4.8%, which includes LCIS cases) can be attributable to technically suboptimal sectioning of specimen or sampling error. False-negative frozen sections of nipple margin resulted in another operative procedure of nipple resection in three of 208 cases. Subsequent nipple resection is a relatively minor procedure that can be performed under local anesthesia. One frozen section in our series was interpreted as “suspicious”; however, no immediate surgical action was taken based on this misinterpretation of a tangentially sectioned seemingly proliferative appearance of lactiferous ductal epithelium.
The relative accuracy of frozen section evaluation of the nipple margin allows for a single stage NSM procedure in ~85% of cases. In a 5-year retrospective study of frozen sections conducted in this setting, five cases (10%) yielded a true-positive diagnosis. In this study, eight false-negative diagnoses, and no false-positive, report were rendered. Four (50%) of the latter group were due to sampling errors, three were interpretative errors, and one was attributed to “diathermy artifact” . Our study demonstrates that frozen section evaluation may not be indicated in patients undergoing prophylactic-NSM as no nipple in such cases showed any disease, however, it may be indicated if there is clinical or radiological evidence of disease.
Multiple lactiferous ducts, ranging from 15 to 48, each draining a mammary lobe, open into the surface of the nipple . These lactiferous ducts can be involved by DCIS, either via pagetoid extension from underlying diseased ducts (Fig. 2) or by de novo origin. Lobules can be present in ~17% of nipples and can harbor LCIS . Early Paget disease of nipple as well as incipient in situ and minimal invasive carcinoma can remain clinically occult. Lohsiriwat et al. reported seven cases of Paget disease of nipple in the “saved” nipple in 861 NSM in a median follow-up of 50 months . In Lohsiriwat's series, Paget disease was associated with extensive DCIS component, high-grade DCIS, and HER2 overexpression and was heralded by erosion, crusting, or ulceration of the nipple.
The goal of pathologic evaluation of margin in NSM specimens is to detect clinically occult disease, i.e. ensure that the nipple being “saved” is benign. It is obvious that presence of either invasive carcinoma or DCIS in nipple margin is a contraindication to nipple preservation; however, the rare finding of LCIS in the margin of NSM, as occurred in 2% (4/208) of therapeutic-NSM cases in this series, presents an unusual dilemma. If LCIS is considered a “marker” for subsequent development of invasive carcinoma in either breast  then LCIS in the nipple margin presents minimal oncological threat, particularly if bilateral NSM is performed. However, if LCIS is regarded as a precursor of invasive carcinoma , then the affected nipple should be excised. Irrespective of whether LCIS is viewed as a “marker” or as a “precursor” lesion, we recommend nipple resection in such cases since there is inadequate follow-up data of patients in whom a portion of nipple was “saved” after a diagnosis of LCIS in nipple margin. It is for this reason that in this series we have included LCIS in the category of malignant involvement of nipple margin—recognizing that others may hold a contrary view. If LCIS is excluded, then the incidence of malignant involvement is 12% (versus 14% if LCIS is included).
The incidence of NAC recurrence in our study was 0.3% in a mean follow-up of ~33 months (range 11–61 months). There was only one histologically documented recurrent invasive carcinoma in the “saved” nipple, wherein, the primary tumor was a 2 cm poorly differentiated invasive ductal carcinoma located in the lower-outer quadrant without lymphovascular or lymph node involvement. The recurrence presented as nipple retraction 36 months after therapeutic-NSM was performed. Of the two factors that we found to be associated with occult nipple involvement at the time of therapeutic-NSM (namely central location and N2 or N3 lymph node status) neither was operative in this recurrence. The incidence of NAC recurrence in NSM ranges from 0 to 2.4% (Table 3). It is notable that three patients in our series died with widely metastatic disease. All three patients had received neoadjuvant chemotherapy and developed metastases without evidence of local recurrence.
|First author (ref.)||Total NSM (n)||Indication for NSM||NAC recurrence rate (%)||Mean follow-up (months)|
|Therapeutic (n)||Prophylactic (n)||Rate (%)|
|Benediktsson et al. ||216||216||0||0/216 (0)||156|
|Gerber et al. ||61||61||0||1/61 (1.6)||59|
|Voltura et al. ||51||34||17||0/51 (0)||18|
|Crowe et al. ||54||37||17||0/54 (0)||41|
|Sacchini et al. ||192||68||124||0/192 (0)||24.6|
|Paepke et al. ||109||94||15||0/109 (0)||34|
|Babiera et al. ||54||53||1||0/54 (0)||15|
|Harness et al. ||60||40||20||1/40a (2.3a)||18.5|
|Jensen et al. ||149||99||50||3/149 (2)||60|
|Spear et al. ||162||49||113||0/162 (0)||30|
|de Alcantara et al. [34, 35]||353||157||196||0/353 (0)||NA|
|Petit et al. ||934||934||0||10/934 (1.1)||50|
|Kneubil et al. ||948||948||0||2/98b (2.4b)||60|
|Current study||325||208||117||1/325 (0.3)||33|
In summary, NSM is an increasingly performed surgical procedure for treatment of patients with breast carcinoma or those at risk for its development. In our series, 14% of separately submitted nipple margin from therapeutic-NSM showed occult malignancy, the most common form of which was DCIS. Frozen section evaluation of nipple margin was generally accurate (sensitivity of 64% and specificity of 99%). Although recurrences in NSM are rare, close clinical follow-up is warranted in all patients after NSM.
Conflicts of Interest
None of the authors have any commercial interest in the subject of study. No financial or material support was provided to the authors’ from any source.
- 5Prophylactic and therapeutic mastectomy in BRCA mutation carriers: can the nipple be preserved? Ann Surg Oncol 2011;18:3102–9., , , et al.
- 7Lobules in the nipple: frequency and significance for breast cancer treatment. Pathol Annu 1985;20:317–21., .
- 11Cancer of the breast with nipple involvement. Br J Cancer 1989;74:58–61., .
- 16Defining a place for nipple sparing mastectomy in modern breast care: an evidence based review. Breast J 2012; doi:10.1111/j.1524-4741.2011.01220.x. [Epub ahead of print]., .
- 22The accuracy of intraoperative subareolarr frozen section in nipple-sparing mastectomies. Ochsner J 2010;10:188–92., , , et al.