Oncoplastic Breast Consortium consensus conference on nipple-sparing mastectomy

The first set of questions aimed at setting the stage before individually assessing each question in depth in the following categories (Fig. 1 and Supplementary Appendix 4). The panel reached consensus when identifying the need for standardizing indications, contraindications and outcome assessment, while a clear majority recommended standardization of surgical technique, type and timing of reconstruction in the setting of adjuvant radiotherapy and radiological follow-up.

A clear majority of the panel felt that the oncological safety of NSM is comparable to conventional mastectomy without reconstruction and to BCS, and the panel reached consensus that it is comparable to SSM if cases are selected appropriately (Fig. 1 and Supplementary Appendix 4). Residual breast tissue left behind underneath the skin envelope in NSM and SSM has raised concerns about the completeness of mastectomy [16]. The available evidence on oncological safety of NSM is based on observational studies of low overall quality [17]. One of the pioneering prospective single-center studies of 216 NSM patients reported loco-regional recurrence (LRR) rates of 8.5% among irradiated and 28.4% among non-irradiated patients at a median follow-up of 13 years [18]. The high rate of LRR was primarily attributed to the non-radical surgical technique of subcutaneous mastectomy that was used at the time of recruitment between 1988 and 1994. In addition, the patient population included many large, node-positive tumors. However, more than 60 NSM series have been published since 2000; a selection of which is shown in Supplementary Appendix 2. The vast majority were single-center studies. They almost all showed acceptable rates of recurrence after NSM. A recent analysis of the Surveillance, Epidemiology, and End Results (SEER) database identified 2440 breast cancer patients who received NSM from 1998 to 2013. The 5- and 10-year cancer-specific and overall survival rates were 96.9% and 94.9%, and 94.1% and 88.0%, respectively, very much in keeping with standard mastectomy techniques [19].

While a slim majority of the panel, with one-third abstaining, believed that the oncological safety of NSM is not compromised by the use of immediate fat grafting, there was clear consensus that safety is not compromised when timing is delayed (Fig. 1 and Supplementary Appendix 4). The oncological safety of autologous fat grafting after NSM and immediate reconstruction has been widely debated [20]. Fat grafting can improve esthetic outcomes of breast reconstruction when compromised by radiotherapy, for example, which may translate into increased patient satisfaction and psychosocial wellbeing [21, 22]. Preclinical studies, however, indicated that fat grafting may promote breast cancer growth and metastasis [23]. Several reviews and contemporary patient series show that fat grafting does not increase the risk of recurrence when applied as a delayed procedure after autologous reconstruction [24‐27]. Finally, a slim majority of the panel felt that the oncological safety of NSM did not depend on the use of systemic therapy.

There was consensus that NSM can be performed for any tumor size that does not involve the skin or nipple areola complex (NAC) independent of axillary status (Fig. 1 and Supplementary Appendix 4). However, the panel was divided when asked if NSM could be offered to patients with locally advanced breast cancer (LABC) without the use of successful neoadjuvant chemotherapy (NACT). Several groups have broadened the indication for NSM to include patients with LABC, who have been successfully down-staged with neoadjuvant systemic therapies [2, 6]. The evidence base for use of NSM in this setting is currently poor and more studies with longer follow-up are required.

The panel recommended NSM for early breast cancer and ductal carcinoma in situ (DCIS), and, unanimously, in the risk reducing setting. This latter indication is now well-established in clinical practice [4, 5, 28].

The panel strongly felt that only specialized surgeons with high-volume training should perform NSM. This claim has been made repeatedly in the past by specialized breast and plastic surgeons, and should be supported by volume-outcome research with caseload as predictor and rate of complications and local recurrence as outcomes. NSM certainly is technically challenging and surgeons experience greater physical symptoms, mental strain, and fatigue with NSM than SSM [29].

The panel strongly considered the location of the incision to be a risk factor for severe mastectomy flap necrosis requiring re-operation (Fig. 2 and Supplementary Appendix 5). This is supported by results of a single-center retrospective review of 500 NSM procedures that showed a dramatic increase in the risk of mastectomy flap necrosis by the use of peri-areolar incisions, while the inframammary approach was associated with a decreased risk [30]. Another retrospective single-center study could not confirm this relationship [31].

The panel did not reach consensus or even a majority agreement when asked if the risk of flap necrosis depends on the surgical technique used for mastectomy flap dissection. This disagreement is mirrored by discussions in the literature. A prospective observational study at Memorial Sloan-Kettering Cancer Center assessed risk factors for skin necrosis in patients undergoing uni- or bi-lateral mastectomy and reconstruction. They found that the use of sharp dissection versus cautery significantly increased the odds of any skin flap necrosis in multivariable analysis (odds ratio 5.94, 95% CI 2.16–16.34, p < 0.001) [32]. Another single-center study did not confirm this association [31]. By contrast, the panel reached consensus that the amount and duration of pressure applied by retractors during surgery play an important role.

The panel considered skin flap thickness to be associated with risk of skin necrosis, which is supported by the limited published evidence [33]. In this single-center retrospective review, 10 NSMs with ischemic complications had significantly thinner NSM flaps as measured by post-operative magnetic resonance imaging (MRI) compared with 50 NSMs without ischemic complications (7.3 mm vs. 9.0 mm, p = 0.0280).

The panel was divided on the potential benefit of intra-operative skin flap viability assessment: almost half abstained, suggesting it is rarely used clinically and the other half was divided on its clinical value. One single-center study from Japan showed a dramatic decrease in the rate of full-thickness skin necrosis by the use of indocyanine green angiography-guided skin trimming from 17.8 to 4.8% (p < 0.05) [34]. During the iterative question review-and-adjust process preceding the conference, a plastic surgeon added a question addressing the concept of tissue pre-conditioning, for example by the local application of nitroglycerine. More than half of the panel abstained from a vote about the value of nitroglycerine, suggesting little experience in its use and very few were in favor. However, a randomized controlled trial (RCT) performed at the University of British Columbia was stopped for early success after 165 patients showed a difference in the rate of mastectomy flap necrosis of 18.5 percent in favor of nitroglycerine (p = 0.006; 95% CI 5.3–31.0%) [35]. The panel did not feel that early use of compression dressings or bras has any influence on the development or the prevention of flap necrosis. There was strong consensus that the risk of flap necrosis depends on the expertise and experience of the surgeon. The organization and promotion of high-quality surgical training is one of the key missions of the OPBC.

There was a strong consensus that the thickness of the skin flaps should be determined by the place and depth of the superficial fascia and that dissection should follow this plane (Fig. 2 and Supplementary Appendix 5). This is in line with the conclusion of a recent literature review [36]. It showed that the variable thickness of the subcutis precludes adoption of a single specific universal thickness for mastectomy skin flaps.

The panel did not recommend the use of an inframammary fold incision independent of breast size and shape, and was equally divided when asked if it can be used for tumors < 1 cm from the nipple on imaging (Fig. 2 and Supplementary Appendix 5). The inframammary fold incision seems to be the most popular and commonly used approach today [37, 38]. A retrospective single-center study demonstrated that refinements of surgical techniques, such as use of the inframammary fold incisions, can dramatically lower the rate of complications after NSM [39]. New and less invasive approaches are currently being evaluated. For example, a single-axillary-incision endoscopic hybrid technique was safe and associated with low morbidity and high patient satisfaction in an early patient series from Taiwan [40]. This hidden incision type can also be used for robotic assisted techniques that have the potential to further improve patient satisfaction although these techniques are not widely practiced at present [41, 42].

A bare majority felt that the inframammary approach should not be used for tumors in the upper inner quadrants in large breasts to avoid compromising oncological safety due to limited access that may increase the risk of positive margins. A retrospective analysis of a prospectively maintained single-center database showed that the majority of relapses (12 of 14; 85.7%) developed in the subcutaneous tissue of the quadrant where the primary tumor was located [43]. A large majority of the panel was of the opinion that the inframammary incision can be used independent of tumor size and location or breast size as long as the whole breast tissue is safely reached and removed and operative visibility is maintained. Finally, a majority of the panel discouraged a separate periareolar incision for retro-areolar frozen section to exclude nipple involvement before the inframammary fold is incised.

A slim majority voted that cT4b and cT4c breast cancers limited to a 1 × 1 cm area of skin ulceration should be a contra-indication to skin preservation (Fig. 3 and Supplementary Appendix 6). However, other investigators have suggested that careful assessment of pathology and treatment response may identify clinical T4 patients appropriate for conservation of the skin envelope [44]. A clear majority of the panel felt that edema of the skin with enhancement on imaging should be a contraindication even without redness. The panel reached consensus that inflammatory breast cancer is an absolute contraindication even with complete remission of all skin changes after NACT. Half of the panel believed that the absence of any distance between tumor and skin on pre-operative imaging should be a contraindication, even in the absence of clinical signs of skin infiltration. A slim majority of panelists felt that a histological margin of less than 1 mm is no contraindication for the preservation of the skin flap, while 37% of panelists considered it a contraindication. During the discussion, it became obvious that there are wide international variations in clinical practice concerning positive or close margins after NSM, ranging from further surgery to post-mastectomy radiotherapy to no treatment. A large patient series at Mayo Clinic showed that the overall 5-year risk of local recurrence was higher in patients with a margin ≤ 2 mm compared to a margin > 2 mm (11.2% vs. 3.1%), with the proximity of the final margin being an independent significant risk factor for local recurrence [45]. However, in a retrospective single-center study of 64 SSM procedures with a positive mastectomy margin towards the skin in the area of the primary tumor, only 13 (20%) had residual cancer in simultaneous re-excisions of the subcutaneous tissue, suggesting a high incidence of false-positive margins [16].

There was strong consensus that clinical signs of nipple involvement and any R1 resection at the nipple margin are contra-indications to nipple preservation (Fig. 3 and Supplementary Appendix 6). When the panel was asked if the nipple can be excised without the areola in patients with positive retroareolar margins, no consensus was reached. A majority felt that radiotherapy without excision should not be offered. Finally, no consensus was achieved concerning atypia in the margin on frozen section that is not 100% consistent with the diagnosis of DCIS or invasive cancer.

Even though the clinical relevance of a positive retroareolar margin is not known in the individual patient, the involvement of the NAC has traditionally been considered the main contraindication to NSM [46]. Local recurrence in the nipple can occur as a rare and late event after NSM. In a retrospective analysis of 861 patients, seven nipple recurrences were diagnosed on average 32 months after surgery [47]. In clinical practice, there are many ways to address a positive nipple margin: More surgery, radiotherapy (± boost), or observation [48]. A retrospective analysis of 217 mastectomy patients revealed that despite a high frequency of malignant nipple involvement in 10.6%, less than 1% had involvement of the areola [49]. A center that routinely performed further surgery in 46 patients with positive nipple margins showed that surgical practice differed considerably, with 51% of patients having only the nipple and 49% having the entire NAC excised [50].

The panel was generous in offering to attempt nipple preservation to patients with an increased risk of occult nipple involvement before surgery, such as patients with a tumor-nipple distance ≤ 1 cm or tumors > 3 cm in diameter. Tumor size, tumor-nipple distance, extensive DCIS component and multicentricity are consistent predictors of nipple involvement [51, 52]. Several studies have investigated the role of imaging, particularly MRI and ultrasound, in predicting the risk of nipple involvement, and suggest a minimum distance of 1–2 cm [53‐55]. However, patient selection based on clinicopathologic characteristics is controversial, since a negative retroareolar margin may exclude occult nipple involvement with a high negative predictive value even in patients at high risk [51]. This is in support of the panel recommendations to treat the nipple margin like any other margin.

While a slim majority of the panel agreed that bloody nipple discharge is considered an oncological contraindication to nipple preservation, a clear majority felt that this does not apply to clear nipple discharge, since the presence of nipple discharge is not equivalent to NAC involvement [56].

There was a consensus recommendation that skin reduction techniques with NAC pedicles or free nipple grafting can be offered to women with large and ptotic breasts as part of NSM irrespective of the type of reconstruction. A strong majority felt that a delayed procedure after pre-shaping the breast by reduction mammoplasty is a good approach. Indeed, several techniques have been described to reduce large and ptotic breasts during NSM, and the concept of delayed NSM after reduction mammoplasty is well supported by the literature as well [57‐59].

The panel did not routinely recommend sentinel lymph node biopsy (SLNB) during risk reducing mastectomy, in line with a recent review of this topic (Fig. 4 and Supplementary Appendix 7) [60]. A majority felt that a pre-operative MRI without abnormal finding is a clear reason to omit SLNB. There was consensus not to remove the inframammary fold and the soft tissue all the way to the clavicle in both the risk reducing and therapeutic setting; rather, the breast tissue should be removed according to its individual anatomical borders. However, several panelists cautioned that personalizing the extent of surgery may have an impact on the effectiveness of the procedure. While the panel reached consensus that the fascia of the pectoral major muscle should not be removed during risk reducing NSM, only half of the panel felt that this also applies to therapeutic NSM. Similarly, while there was consensus against routine retroareolar frozen section in the risk reducing setting, only a bare majority felt that this recommendation also applies to the therapeutic setting.

While a slim majority of panelists did not use synthetic meshes, the panel was divided when asked if they use acellular dermal matrix (ADM; Fig. 5 and Supplementary Appendix 8). The preferred method of reconstruction after NSM in a patient with small breasts and no planned radiotherapy varied widely. In fact, of eight different techniques, ranging from autologous to implant-based reconstruction with different timings and positioning of the implants, every option was chosen by at least one panelist. Interestingly, one-third of the panel chose immediate one-stage sub-pectoral implant-based reconstruction without synthetic mesh or ADM. The extensive literature on breast reconstruction after NSM has been summarized in Supplementary Appendix 3. The wide variation in clinical reconstruction practice that is mirrored in the literature calls for RCTs to guide treatment.

There was a clear consensus that patients undergoing NSM should be included in national and/or international prospective registries (Fig. 6 and Supplementary Appendix 9). One such registry that has been recommended during the discussion is the international NSM registry INSPIRE [61]. Indeed, a recent review of 11 observational cohort studies evaluating 7018 NSM, SSM and traditional mastectomy procedures concluded that these studies were of low quality [17]. Hence, there is a clear need for high-quality multi-center prospective studies to assess the efficacy and safety of NSM [62].

The panel was unanimous in recommending the use of pre- and post-operative pictures as standard tool for objective outcome assessment after NSM (Fig. 6 and Supplementary Appendix 9), thereby recognizing the extent of such a commitment at high-volume centers. Since the association between objective esthetic outcomes and QoL is complex, the panel also endorsed the routine evaluation of PROs. The European Organisation for Research and Treatment of Cancer QLQ-BRECON23 questionnaire was proposed by one of the panelists as an internationally validated tool for standardized outcome assessment in patients undergoing breast reconstruction [63]. Almost half of the panelists abstained (presumably, because this is a new and not yet well-known questionnaire) and the other half was divided when asked if this tool should be used for outcome assessment. When asked about the well-established BREAST-Q reconstruction module, a majority of the panel recommended not using all scales for feasibility reasons due to its size, but a large majority voted for selected scales of the questionnaire as a standard tool for outcome assessment [10, 22, 64, 65].