Generally accepted absolute contraindications to NSM include extensive tumor involvement of the skin, inflammatory breast cancer, Paget’s disease of the nipple, and clinical [2–4] and/or imaging evidence of direct NAC involvement [5]. Some authors advocate for routine intraoperative frozen section of subareolar tissue or a nipple core specimen [6, 7]. Detection of invasive or noninvasive cancer in frozen or permanent section analysis of subareolar or nipple core tissue is generally considered an indication for NAC excision, although alternative approaches such as margin re-excision or radiation of the NAC have been reported [8]. Some institutional protocols do allow for NSM in patients with initial skin involvement that shows excellent response to neoadjuvant chemotherapy [9, 10]; however, applicability to patients with initial direct NAC involvement or inflammatory breast cancer has not been reported.

There is conflicting evidence about the relationship between tumor size and prevalence of occult nipple involvement. Mallon and colleagues have published a comprehensive review of clinicopathologic factors and their association with nipple involvement [11]. Their review is somewhat limited by inclusion of several studies from an era when more women presented with locally advanced disease, as well as studies that did not exclude women for clinical and/or radiographic evidence of direct nipple involvement. Nonetheless, their review is useful as perhaps the most complete analysis to date of risk factors for NAC involvement.

Mallon’s meta-analysis of 16 studies examining tumor size found that the incidence of nipple involvement was 9.8 % for tumors smaller than 2 cm, 13.3 % for tumors 2–5 cm, and 31.8 % for tumors greater than 5 cm (p < 0.05) [11]. However, some studies have suggested that this relationship is mitigated by adjustment for potential confounding factors such as tumor-to-nipple distance or subareolar location [5, 12].

At least six multivariate analyses in large groups of patients have identified central or retroareolar tumor location as a risk factor for occult nipple involvement [13–18]. Mallon’s meta-analysis of data from five of these studies found that 35.2 % of central/retroareolar tumors had nipple involvement, compared to 9.7 % of peripheral tumors (p < 0.05) [11]. Retroareolar tumors can be close to or far from the NAC, which may influence the risk of NAC involvement associated with this factor.

In the relevant literature, tumor-to-nipple distance is probably the most uniformly reported clinicopathologic risk factor for nipple involvement [15, 16, 18–32]. Studies have variously reported tumor-to-nipple distances of 2–4 cm as a threshold value below which the prevalence of nipple involvement is significantly increased [20, 21, 23, 25, 27, 33]. However, a recent report of a large, prospectively collected database (n = 392) found that, although tumor-to-nipple distance (assessed by mammogram, ultrasound, or MRI) was associated with nipple involvement on univariate analysis, on multivariate analysis only imaging or clinical evidence of direct NAC involvement remained independently associated with positive nipple pathology [5].

Mallon’s meta-analysis of six studies found a higher incidence of occult nipple involvement in multicentric versus unicentric tumors (29.6 % vs. 12.4 %, p < 0.05) [11]. However, these data should be applied with caution in modern-era patients in whom mammographically occult multicentricity is detected on MRI. In pooled data from two studies, there was no association between multifocality and occult nipple involvement [11].

Mallon’s analysis of 14 studies found that in most there was no correlation between histologic type and occult nipple involvement [11]. Two reports showed a higher incidence of occult nipple involvement in infiltrating ductal carcinomas with extensive intraductal component [20, 21], and a third report showed a higher prevalence of occult nipple involvement in micropapillary tumors [15].

The superficial lymphatic plexus of the breast is found in the dermis and subcutaneous region and drains primarily to the ipsilateral axillary lymph nodes [34, 35]. Thus, there is a theoretical possibility of a correlation between occult nipple involvement and axillary nodal metastasis. Mallon’s meta-analysis of 16 studies found a higher incidence of occult nipple involvement in cases of lymph node-positive disease (24.4 %), compared to lymph node-negative disease (10 %, p < 0.05) [11]. Furthermore, four studies found increasing likelihood of nipple involvement with heavier nodal disease burden [13, 14, 21, 27]. A more recent, large study (n = 466) not included in the meta-analysis found an association between axillary nodal metastasis and nipple involvement, but this relationship was attenuated by multivariate analysis accounting for other clinicopathologic factors, including tumor size and lymphovascular invasion, which are themselves associated with nodal positivity [18].

Pooled analysis of four studies showed a higher prevalence of NAC involvement in patients with HER2 positive disease (19.7 %), compared to HER2 negative disease (10.1 %, p = 0.0001) [11]. Six studies correlating estrogen receptor (ER) /progesterone receptor (PR) status to NAC involvement showed disparate results. One of these six studies had a sample size (n = 2028) exceeding that of the other five studies combined [15]. Meta-analysis of these six studies reflects the findings of this largest study, showing a weak but significant association between both ER negativity and PR negativity and occult nipple involvement [11]. While no rationale was proposed for these associations, it is consistent with the more aggressive tumor biology of HER2 positive and hormone receptor negative disease.

Mallon’s meta-analysis of seven studies found that the incidence of nipple involvement was higher for tumors with lymphovascular invasion, compared to those without lymphovascular invasion (35.6 % vs. 12.4 %, p < 0.05). His analysis of 11 studies found that increasing grade was associated with increasing likelihood of NAC involvement: the incidence of NAC involvement for grades 1, 2, and 3 tumors were 8.7, 12.6, and 17.2 %, respectively (p < 0.05) [11].

Terminal duct lobular units, from which both ductal and lobular breast cancers are thought to arise, are found in 9–26 % of NACs from total or skin-sparing mastectomy specimens [36–38]. This has led to concern that NSM may be associated with increased risk of recurrent or de novo cancer, particularly in BRCA mutation carriers.

In the largest reported series of NSM in BRCA mutation carriers, Yao and colleagues retrospectively reviewed results of 397 NSM in 201 BRCA mutation carriers (62.2 % BRCA1, 37.8 % BRCA2). Of 150 patients who underwent NSM for risk reduction, four (2.7 %) had an incidentally discovered cancer, and in none was the NAC margin involved. At a mean follow-up of 32.6 months, three of 51 cancer patients experienced a local recurrence (5.8 %), and 1 risk-reduction patient experienced a new breast cancer diagnosis (0.7 %); none of these involved the NAC [39]. Several other studies have reported on therapeutic or risk-reducing NSM in smaller numbers of BRCA-positive patients (range, n = 6–89) over relatively short follow-up times (mean/median, range 10.4–51 months). No new breast cancers occurred after risk-reducing NSM [40–44]. The only study reporting a local recurrence (a case of Paget’s disease in the retained NAC) included women without BRCA mutations, and it was unclear if this recurrence occurred in a BRCA mutation carrier [40].

Longer follow-up is available in two earlier studies of BRCA mutation carriers who underwent risk-reducing subcutaneous mastectomy [45, 46]. Subcutaneous mastectomy preserves a margin of breast tissue below the NAC, whereas modern NSM removes most of the breast tissue under the NAC and within the nipple. Nevertheless, in these studies rates of de novo cancer in the NAC were also very low [45, 46]. Rebbeck and colleagues evaluated l05 BRCA 1 or 2 mutation carriers who underwent bilateral risk-reducing mastectomy, at least 30 % of which were subcutaneous. At a mean 6.4 years of follow-up, two women (1.9 %), both of whom underwent subcutaneous mastectomy, had a new breast cancer diagnosis [45]. Hartmann and associates evaluated 26 BRCA 1 or 2 mutation carriers (18 known deleterious mutations, eight variants of uncertain significance) who underwent risk-reducing mastectomy, of which 23 were subcutaneous. There were no new breast cancer diagnoses at a median of 13.4 years of follow-up [46].

Based on these data, there is increasing acceptance of therapeutic or risk-reducing NSM in BRCA mutation carriers. However, appropriate counseling and compliance with long-term clinical follow-up are required, and long-term follow-up data are necessary to confirm the safety of this option for BRCA mutation carriers.

The rationale for more liberalized oncologic criteria is that, while there are strong associations between certain clinicopathologic features and occult nipple involvement, there is, nonetheless, a high false positive rate for many of these risk factors. A small study (n = 58) of patients who underwent total mastectomy found that clinical criteria of tumor size <4 cm and tumor-to-nipple distance of ≥2 cm had a false positive rate of 44.4 % for predicting nipple involvement on final pathology [47]. A larger study of a prospectively maintained database of total, skin-sparing, and nipple-sparing mastectomies (n = 392) found that, although tumor size and tumor-to-nipple distance were associated with nipple involvement on pathology, on multivariate analysis only clinical or imaging evidence of NAC involvement persisted as independent predictors of nipple involvement on pathology [5].

In recent years, a number of institutions have reported their experience with more liberalized oncologic criteria for NSM candidacy (Table 3.2). Six institutions’ experience is reviewed here. The first two analyses summarize outcomes of NSM after liberalized criteria were adopted [3, 4]. Two others report on broadening inclusion criteria and parallel changes in patient outcomes over time [7, 48]. The final two report on outcomes of NSM in a subgroup of patients with locally advanced disease [10, 49].

Paepke and associates report on the outcomes of 109 NSM in 96 patients; 94 NSM were performed for malignancy. Exclusions from NSM were for extensive tumor involvement of the skin, inflammatory breast cancer, a clinically suspicious nipple, or imaging suspicion of NAC involvement. In 14 of 109 breasts (12.8 %), intraoperative frozen section of the subareolar margin was positive for tumor involvement, and in 13 of these cases the NAC was excised, converting to a skin-sparing mastectomy. Although most instances (81.8 %) of NAC excision for positive histology were in patients with retroareolar tumors , nipple conservation was achieved in the majority of patients with retroareolar tumors. At a median follow up of 34 months, there were no recurrences in the NAC. One of 97 conserved NACs experienced necrosis requiring surgical intervention [3].

Similarly, Fortunato and colleagues reported their experience with 138 NSM (122 for malignancy) in 121 patients. Of these, 124 cases were performed after 2009, when a liberal policy of considering NSM for all patients with a minimal tumor-to-NAC distance of 1 cm, no bloody nipple discharge, no large or ptotic breasts, and no absolute contraindications (clinical involvement of the NAC, Paget’s disease, inflammatory breast cancer) was adopted. Sixteen percent (n = 19) of patients underwent NAC excision based on retroareolar margin <2 mm on pathology; of these 19 excised NACs, 11 (58 %) had residual cancer on pathology. At a median follow up of 26 months, there was one local recurrence, which was outside of the NAC. There were 6 cases (4 %) of total NAC necrosis, 14 cases (10 %) of mild NAC desquamation, and 6 cases (4 %) of partial mastectomy skin flap necrosis [4]. Of note, this is the only report of liberalized oncologic criteria reviewed here that retained a tumor-to-NAC distance criterion as a contraindication to NSM.

Two studies examined how institutional criteria for NSM candidacy and outcomes after NSM evolved in parallel over time. Both report improved technical outcomes over time and no NAC recurrences despite broadening indications for NSM [7, 48].

Coopey and colleagues assessed trends in eligibility and outcomes over time for 645 NSM performed in 370 patients at the Massachusetts General Hospital from 2007 to 2012. Three hundred fifteen NSM were for malignancy (48.8 %). Initial criteria for therapeutic NSM were tumor size <2 cm and tumor-to-nipple distance ≥2 cm. However, with time these exclusions were eliminated, and final exclusion criteria included only clinical or imaging evidence of NAC involvement, locally advanced breast cancer with skin involvement, inflammatory breast cancer, or bloody nipple discharge. Despite this liberalization of criteria and the fact that more NSM in the later years were done for cancer, the authors noted a decreasing incidence of positive subareolar margins over time (6.5 % in 2007–2010 vs. 2.7 % in 2011–2012, p = 0.027), which they attributed to improved patient selection with increasing surgeon experience. At a mean follow-up of 22.1 months, local recurrence occurred in 2.6 % of breasts operated on for cancer, and no recurrences involved the NAC. Total nipple necrosis occurred in 1.7 % of breasts, and this rate was stable across time periods (p = 0.45) despite higher body mass index (BMI) and higher breast volumes in the later time period [48].

Krajewski and associates at the Mayo Clinic reported on their experience over time with 341 patients who underwent 566 NSM (218 NSM were for malignancy). Early criteria for NSM included risk reduction or cancers 2 cm or less in diameter and located over 2 cm from the NAC; inflammatory breast cancer and cancer extension into the NAC were exclusion criteria. Obesity, large/ptotic breasts, and prior radiation were initially relative contraindications. Over time (2009–2010 vs. 2013–2016/2014), there were significant increases in mean BMI and proportion with bra cup size C or larger, prior radiation, tumor within 2 cm of the NAC, and neoadjuvant chemotherapy . Of 19 nipple excisions in patients with a preoperative diagnosis of cancer, six were for positive pathology (five atypia, one DCIS). No factors, including tumor-to-nipple distance, multifocality/multicentricity, or performance of preoperative MRI were associated with nipple excision. At a median 19-month follow-up, there were five locoregional recurrences, and none were within the NAC. Between the earlier and later time periods, there were declining rates of 30-day postoperative complications (41.4 % vs. 18.7 %, p < 0.001), postoperative complications requiring treatment (15 % vs. 3 %, p = 0.0001), and a trend towards decreased nipple loss (5.7 % vs. 2.2 %, p = 0.10) [7].

Finally, two studies report their outcomes after NSM for a subgroup of patients with locally advanced disease. Again, the oncologic outcomes were excellent, with no cases of NAC recurrence [10, 49].

Burdge and associates at the University of Arkansas retrospectively reviewed skin-sparing and nipple-sparing mastectomy procedures performed from 2001 to 2012 for locally advanced disease. Exclusions were for skin involvement, inflammatory breast cancer, collagen-vascular disease, and smoking within the previous 6 months. Twenty-one patients who underwent skin-sparing mastectomy and 39 patients who underwent NSM were included. Of 39 patients treated with NSM, 38 received neoadjuvant chemotherapy, and all underwent post-mastectomy radiation to the affected breast. Mean pretreatment tumor size for patients undergoing NSM was 3.4 cm (±2.2 cm). In the skin-sparing and nipple-sparing groups combined, wound/tissue necrosis occurred in 16.7 %, implant removal occurred in 5 %, and capsular contracture occurred in 10 %. Over a mean follow-up of 25.3 ± 18.8 months, 10.3 % of patients treated with NSM experienced locoregional recurrence, and no recurrences were within the NAC [49].

Peled and associates at the University of California, San Francisco reported on prospectively recorded outcomes of 139 NSM performed from 2005 to 2013 for locally advanced disease (stage IIB, 18 %; stage III, 82 %). This group’s selection criteria have evolved over time, and current criteria require MRI only for tumors that are close to the NAC on mammogram or clinical exam, and exclude patients from NSM based only on NAC involvement on exam or imaging, significant ptosis, or large breast size [9, 10]. Patients with initial skin involvement that responds to neoadjuvant chemotherapy are offered NSM. In this study, 97 % of included patients received chemotherapy (77 % neoadjuvant, 20 % adjuvant), and 63 % received postmastectomy radiation therapy. Reconstruction was with expander-implant (92 %), pedicled TRAM flap (7 %), or DIEP flap (1 %). With a mean follow-up of 41 months, 5 % of patients had locoregional recurrences, 15 % experienced distant recurrence, and 2 % had simultaneous local and distant recurrences. There were no recurrences in the NAC. Surgical complications included NAC necrosis (1.4 %), mastectomy skin flap necrosis (3.6 %), implant loss (11 % of expander-implant reconstructions), and severe infection requiring IV antibiotics (12 %). Mastectomy skin flap necrosis and implant loss were seen only in patients who received postmastectomy radiation [10]. In a separate publication, the same group reported that axillary lymph node dissection was an independent risk factor for implant loss, probably due to impaired lymphatic drainage of the breast, and suggested that patients who require axillary lymph node dissection be encouraged to undergo lumpectomy or mastectomy with autologous reconstruction whenever feasible [50].

In summary, despite the broadening indications for which NSM has been offered, including patients with locally advanced disease, tumor proximity to the NAC on preoperative imaging, and deleterious BRCA 1/2 mutations, the rate of NAC recurrence remains exceedingly low. There is increasing evidence that the NAC can be treated as any other skin margin, and can be safely retained in patients without clinical, imaging, or pathology evidence of NAC involvement with very low risk of local recurrence.

Whereas oncologic criteria have been developed to minimize NAC excisions and NAC recurrences, technical considerations should be designed to minimize complications such as skin flap necrosis, nipple necrosis, implant loss, wound infection, and poor cosmetic outcome.