The frequency of skin-sparing and nipple-sparing mastectomies (SSM and NSM, respectively) has increased significantly over the past years due to a number of reasons. For example, growing awareness of familial and genetic factors predisposing to breast cancer has led to a higher number of risk-reducing mastectomies.1 In addition, continuously improved diagnostic procedures increase the detection rate of widespread ductal carcinoma in situ (DCIS) components and multicentric tumors that do not require conventional mastectomy. Skin-sparing mastectomy and NSM are meanwhile considered as established and safe surgical procedures with superior cosmetic outcome compared with conventional mastectomies.2,3,4,5,6 While SSM was already introduced in 1991, NSM has been being performed since 2001.7,8,9,10,11

Both mastectomy techniques were designed to remove the entire breast tissue under concomitant preservation of the skin envelope and the inframammary fold, thereby enabling immediate breast reconstruction. Because the breast maintains its natural shape and contour, cosmetic results surpass the results of conventional mastectomies. Several retrospective and some prospective studies reported a satisfying oncological outcome regarding local recurrence after SSM and NSM, which is comparable to the outcome of conventional mastectomy, and even NSM without consecutive radiotherapy has been shown to be safe.2,3,12,13,14,15,16,17,18,19,20,21,22

Superior cosmetic results, increasing surgical experience and promising safety data, suggest that the frequency of SSM and NSM will increase. However, the surgical challenge of removing a maximum amount of breast tissue while achieving low morbidity and the best possible cosmetic result via an adequate skin envelope will remain. Both in oncological and risk-reducing surgery, residual breast tissue (RBT) can be considered as a potential risk factor for breast cancer recurrence or increased breast cancer incidence, respectively.

Thus, there is a need to investigate prospectively the radicality of SSM and NSM with regard to the amount of histologically detected RBT or residual disease under the skin envelope. To our knowledge, three trials using three different methods investigated hitherto the radicality of SSM, yielding conflicting results. In 42 patients, Torresan et al. performed a modified SSM and removed the skin flap for subsequent assessment of RBT and residual disease. In this study, RBT was reported in 60% and residual disease in 10% of the mastectomies.23 Dreadin et al. investigated skin specimens that were removed during SSM for nononcological reasons and immediate reconstruction in 66 patients. They reported RBT in 6% of mastectomies but no residual disease.24 Slavin et al. assessed 114 biopsies taken from the remaining native skin flap edges after periareolar SSM in 32 patients and reported no RBT.25 In a related study, Reynolds et al. evaluated the prophylactic and therapeutic value of NSM and examined histologically for RBT and residual disease a total of 62 excised nipple-areolar-complexes (NAC) from 33 BRCA mutation carriers. They found RBT within the NAC in 24% of all cases and residual disease in 3% of the 29 therapeutic cases.26 In the present prospective trial, we assessed the radicality of SSM and NSM by quantifying the presence of RBT and of residual disease. Importantly, we assessed patient- and therapy-related factors that might be associated with RBT and evaluated the surface of the mastectomy specimens with respect to the distance to the breast tissue.

Methods

Before mastectomy, each patient was informed about the study scope and design and gave written, informed consent. The research protocol (KEK-ZH-Nr. 2015-0565) was approved by the local ethics committee of the Canton of Zurich, according to the national and international ethics guidelines. Mastectomies were performed at the Brust-Zentrum Zürich (n = 152) and the Brust-Zentrum St. Anna Hospital in Lucerne (n = 8). Only surgeons with high surgical volumes (i.e., at least 50 surgeries per year) participated in the trial.

In cases of curative indication, the decision to perform SSM or NSM was primarily based on tumor characteristics, such as tumor size, tumor location, and DCIS-component, and on patient characteristics, such as breast shape, breast size, and previous surgery. Patients` preference also was considered whenever it did not contradict the medical indication. NSM was the standard procedure in cases of risk-reducing indication.

To assess the presence of RBT after SSM and NSM, biopsies were taken from the remaining skin envelope immediately after breast removal. The biopsy procedure was performed according to a predefined pattern of radial localizations: 10 biopsies per SSM, and 14 biopsies per NSM. Thus, in the case of NSM four additional biopsies were taken under the peripheral part of the areola (Fig. 1). A mastectomy was defined as RBT-positive whenever RBT was detected in one or more biopsies.

Fig. 1
figure 1

Pattern of biopsy points in SSM (A–K) and NSM (A–O)

Full size image

Five additional incisions (A2–E2) were marked on the excised mastectomy specimens. Their localization corresponded to biopsy points A–E of the skin envelope. Pathologists were instructed to measure the distance between specimen surface and glandular tissue at these marked incision points (Fig. 2, points A2–E2). Low distance was defined as < 1 mm in each of the points A2–E2, large distance was defined as ≥1 mm in at least one of the points A2–E2.

Fig. 2
figure 2

Anatomical visualization of biopsy points in SSM and NSM in 12 o`clock position (A1, F, L) and incision points on the removed specimen in opposite (A2)

Full size image

Participating pathology institutes are qualified partners of the certified breast centres. Involved pathologists were trained on the study-specific assessment procedures. Surgeons remained blinded to the pathology results of the biopsies and to the surface distances until the end of the trial. Blinding was lifted whenever tumour tissue was detected in the biopsies. In these cases, surgeons and patients were informed and further therapeutic steps were decided upon at a multidisciplinary postoperative tumour board. The number of mastectomies per surgeon ranged between 6 and 65 (3.8% and 40.6% of all mastectomies, respectively).

Patient- and therapy-related data were collected to assess the relation of RBT to such factors as age, side, body mass index, indication for surgery, kind of axillary surgery, breast weight, surgeon, type of surgery, type of incision, as well as evidence and severity of mastectomy skin flap necrosis due to perfusion disorders (as defined by the SKIN Score).27 Evidence for and severity of mastectomy skin flap necrosis was assessed with the SKIN Score, which was determined by the plastic surgeon during the postoperative assessment.27 The SKIN Score system incorporates both depth and surface area of skin flap necrosis. “No necrosis” is defined as SKIN Score A1 (whereby “A” stands for no evidence for necrosis regarding depth and “1” stands for no evidence for necrosis regarding surface area). Anything beyond A1 (e.g., B1 or A2) is defined as evidence for necrosis.

Statistics

Statistical analyses were performed with the Statistical Package for the Social Sciences (IBM® SPSS® Statistics, Version 21, IBM Corp., Armonk, NY). Comparisons of patient- and therapy-related variables between SSM and NSM were done with χ2 and Student’s t tests for equal or unequal variances, wherever applicable. Logistic regression was used to assess the influence of putative risk factors on RBT, which served as binary dependent variable (0: no RBT; 1: at least one probe with positive breast tissue detection). Age, BMI, and breast weight served as continuous independent variables. Side of mastectomy (left vs. right), type of mastectomy (SSM vs. NSM), and skin flap necrosis (no necrosis vs. evidence for necrosis) served as binary independent variables. Indication for mastectomy, axilla surgery, preoperative treatment, surgeon, and incision type served as categorical independent variables (simple contrast scheme used). Results presented herein correspond to the final, concordant solution of the forward and backward selection. At each step, variables were included at nominal P < 0.05 and excluded at P > 0.1. For all tests, the level of significance was 0.05.

Results

Between April 2016 and August 2017, 124 patients with an indication for SSM or NSM were included in the trial. Thirty-six of these patients underwent bilateral mastectomy, resulting in 160 mastectomies (99 SSM, 61 NSM) in total (Table 1). The mastectomies were performed by seven surgeons. Indications for mastectomy were either risk-reducing (n = 51) or curative for histologically confirmed invasive (n = 83) or in situ breast cancer (n = 26).

Table 1 Patients and mastectomy characteristics
Full size table

Patients receiving NSM were significantly younger than patients receiving SSM (SSM: 51.6 ± 11.2 years; NSM: 46.1 ± 11.1 years, P = 0.011). BMI was significantly lower in patients who received NSM (P = 0.001; Table 1). Fifteen of 99 SSMs (15.1%) and 36 of 61 NSMs (59%) were performed for risk-reducing reasons.

Residual Breast Tissue Detection

Residual breast tissue was detected in 82 (51.3%) of 160 mastectomies. The median RBT percentage per breast was 7.1%. RBT-positive biopsies were homogeneously distributed between peripheral points A–K (3–11%) and were found in significantly more cases in the central biopsy sites L-0 behind the areola (18–28%, P < 0.001, Person’s χ2 test; Fig. 3). Residual tumour tissue was detected in one SSM (point F) and in one NSM (points M and N), which corresponds to 0.25% of 1,190 biopsies that were taken from curative mastectomies. Reexcision was performed in both cases.

Fig. 3
figure 3

Incidence of RBT in different biopsy points

Full size image

Residual Breast Tissue-Associated Factors

In the logistic regression, only the factors type of surgery, i.e., NSM vs. SSM, and surgeon contributed significantly to RBT (P < 0.001 for both variables; Table 2). The proportion of RBT in NSM was significantly higher than in SSM (68.9% vs. 40.4%, P < 0.001). Further analysis of the factor surgeon revealed that surgeon #1, who performed most mastectomies (n = 65), had a significantly lower rate of RBT (26.2%, P < 0.001) compared with the other surgeons. The number of mastectomies per surgeon was highly variable and ranged between 6 and 65 mastectomies. All surgeons performed SSMs (ranging between 4 and 33 mastectomies per surgeon) as well as NSMs (ranging between 1 and 32 mastectomies per surgeon) (Supplemental Table 1). Between-surgeon RBT rate ranged between 26.2% and 100%. However, the latter result (100% RBT rate) should be taken with caution because of the low overall number of mastectomies (n = 8) of the particular surgeon (Supplemental Fig. 1).

Table 2 Logistic regression model used to analyze factors related to RBT (biopsies included: A–O)
Full size table

With the exception of indication for mastectomy, categorical independent variables (i.e., axilla surgery, preoperative treatment, surgeon, and incision type) contained a relatively high number of low-frequency (≤ 5%; Table 1) subcategories. To test the stability of the logistic regression, we re-ran the analysis by using a binary coding scheme for the respective variables (i.e., no axilla surgery vs. rest, no preoperative treatment vs. rest, surgeon #1 vs. rest, periareolar incision vs. rest). As in the first analysis, only the factors type of surgery (i.e., NSM vs. SSM) and surgeon (i.e., surgeon #1 vs. rest) contributed significantly to RBT (P < 0.001 for both variables). Because two of the included factors (i.e., age and BMI) were patient- rather than mastectomy-specific, we repeated the analysis by considering only a subset of 124 mastectomies (i.e., 1 mastectomy per patient). In the case of patients undergoing bilateral surgery, the respective mastectomy was selected randomly. Again, only the factors type of surgery and surgeon contributed significantly to RBT (P < 0.001 and P = 0.005, respectively).

Skin Flap Necrosis

Skin flap necrosis was not related to the presence of RBT (P = 0.207; Table 2). Of the additional patient- and therapy-related factors, only type of surgery had a borderline significant influence on the incidence of skin flap necrosis, with higher incidence detectable when the nipple was preserved (mastectomies with detectable skin flap necrosis: 17/61 NSM vs. 15/99 SSM, P = 0.051).

Distance Between Tissue and Specimen Surface

In an additional step, pathologists assessed the distance between breast glandular tissue and the circumferential/anterior surface of the removed specimen at five defined points (Fig. 2). Distances ranged between 0 mm to > 10 mm with a high between-point variability. Left breast side, high BMI, and RBT in points A–K were significantly associated with lower distance between specimen surface and breast glandular tissue (linear regression, P = 0.004, P = 0.003, and P = 0.045, respectively). Conversely, an increased distance between breast tissue and circumferential/anterior surface of the mastectomy specimen correlated with low RBT rate (Supplemental Fig. 2).

Discussion

We detected at least one RBT-positive biopsy in 51.3% of the mastectomies in the present study. Previous trials examining the rate of histologically detected RBT after SSM or NSM reported RBT rates ranging between 0% and 60%.23,24,25,26 This high variability is probably caused by the different methods used to assess RBT. For example, Torresan et al. performed a modified SSM and removed the skin flap for subsequent assessment of RBT and residual disease. While the number of patients was low (n = 42), a large number (n = 80) of glass slides per skin flap was examined under an optical microscope. The authors reported RBT in 60% and residual disease in 10% of the cases.23 However, these results cannot be easily translated to the actual SSM situation, because the examined skin flaps were removed. In another study, Slavin et al. took 114 biopsies from the remaining native skin flap edges after periareolar SSM in 32 patients and reported no RBT. However, the number of biopsies per mastectomy varied from one to four, and biopsies were taken only at the edges of the skin flap and not at other points under the skin.25 Therefore, it is questionable whether the results of this study are generalizable.

Few additional attempts to measure RBT after SSM and NSM do exist, albeit each following a different methodology.24,26 The rather limited number of attempts to measure RBT so far along with the considerable variation in methodology renders the assessment of RBT after SSM and NSM challenging. Nevertheless, there is a need for reliable data on the presence of RBT and remaining tumour tissue after SSM and NSM with regard to oncologic safety of risk-reducing and curative mastectomies.

In our trial, we used a sensitive method with a higher number of mastectomies (n = 160) and biopsies (n = 1,844) to obtain representative and robustly quantifiable results. Moreover, our study adopted a prospective design, as biopsies were taken under the remaining skin envelope during routinely performed SSM or NSM. Importantly, biopsies were taken using a predefined scheme in each of the 160 mastectomies throughout the whole skin envelope after breast removal, resulting in a total of 1,844 biopsies.

Based on the high number of biopsies (n = 1,844) and the standardized biopsy scheme, we used the median RBT percentage per breast (7.1%) as an approximation of RBT-positivity in the 160 mastectomies. We acknowledge that this approximation might be limited given that not the entire area under the skin envelope was tested. However, an even denser pattern of biopsies to reach a more accurate conclusion would not have been feasible. The low percentage of residual tumour tissue (i.e., 3 of 1,190 biopsies taken from 109 curative mastectomies) confirms the efficacy of SSMs and NSMs regarding complete resection of invasive and in situ lesions.

We would like to stress that our study focused on factors influencing RBT detection specifically after SSM and NSM, because these techniques represent the vast majority of mastectomies performed in the participating centres. Thus, a comparison with conventional mastectomy is not provided.

With regard to patient- and therapy-related factors possibly influencing the occurrence of RBT, the assumption that high breast weight or BMI is linked to increased RBT was not confirmed. This also was the case for the factors age, indication of surgery, and type of incision. We note that type of incision was tightly linked to type of mastectomy. For example, a periareolar incision was almost exclusively used in SSM. Both factors entered the stepwise logistic regression model, which revealed that type of mastectomy was strongly associated with RBT. No additional variance regarding RBT was explained by the factor incision type.

Unlike Dreadin et al., we observed that the factor surgeon contributed highly significantly to the percentage of RBT-positivity.24 One senior surgeon with the highest mastectomy performance rate in the trial and a long-standing experience in SSM and NSM in general performed the mastectomies with a significantly lower number of RBT-positive biopsies.

Regarding the type of surgery, we confirmed a significant higher rate of RBT under the NAC. This expected finding stresses that the indication for NSM has to be taken carefully in selected patients with sufficient tumour distance to the nipple and less tumour aggressiveness.

Another aspect that should be considered in the attempt to optimize surgical techniques to avoid RBT is the composition of the breast glandular surface. In this trial, pathologists measured the distance between surface of mastectomy specimens to breast tissue at five points and found inconsistent distances within the same breast between 0 mm and > 10 mm. This finding supports the reports of Beer et al., who assessed the superficial layer (SL) that can be found between fat and breast tissue. They reported absence of SL in 56% of 62 examined breast specimens and the presence of irregular fascia with various glandular islands in 42% of the specimens.28 Notably, Nickell and Skelton argued that there is no visible separation between fat and breast tissue.29 With regard to the RBT-positive biopsies, however, we found that narrow distance (< 1 mm) between specimen surface and specimen breast tissue was significantly associated with higher RBT rates in the skin envelope.

Skin flap necrosis due to microvascular disorders (as defined by the Skin Score) is one of the most severe complications of SSM and NSM.27 In our study, the hypothesis that a more radical surgery (e.g., complete absence of histologically detected RBT) leads to a higher rate of skin flap necrosis was not confirmed.

In conclusion, this trial demonstrated that RBT underneath the skin envelope after SSM and NSM is a frequent phenomenon, especially behind the nipple. However, we observed that highly experienced surgeons can perform SSM and NSM more radically under concomitant preservation of a low rate of skin flap necrosis. This suggests that intensive training of surgeons less experienced in SSM and NSM may lead to more radical surgical techniques that are not associated with increased rates of skin flap necrosis. However, the variable surface character of the breast gland poses anatomical limitations to the removal of the entire breast tissue.28,29 Thus, the interplay of surgical technique and patient-specific anatomic conditions is central to the frequently observed phenomenon of RBT after SSM and NSM. We argue that future prospective studies should include additional preoperatively assessed anatomical factors, such as subcutaneous flap properties, that might influence the presence of RBT after SSM or NSM.

Notably, the clinical relevance of RBT after SSM and NSM remains elusive, given previous findings suggesting that local recurrence might be more strongly associated with tumour biology and tumour aggressiveness than with surgical technique.30,31 It is reasonable to assume that patients with aggressive tumours, high risk of local recurrence, or genetic predisposition might benefit from intensive postmastectomy follow-up that includes RBT detection. As shown in recent studies, the presence of RBT can be reliably assessed by postoperative MRI of the reconstructed breast.32,33,34,35 For example, Woitek et al. detected RBT in up to 50% of all breasts after NSM and 13% after SSM.32 These considerations and our results of the RBT-frequency after SSM and NSM led us to adapt the current wording from “prophylactic mastectomy” to “risk-reducing mastectomy” in the communication with our patients to prevent unrealistic perceptions concerning the efficacy of the surgical procedures. Furthermore, long-term, follow-up studies are warranted to reliably judge the true risk of cancer incidence or recurrence in cases with RBT.