Skin-sparing mastectomy (SSM) and nipple-sparing mastectomy (NSM) allowed for a noticeable improvement in reconstructive surgery aesthetic results [1]. SSM, in particular, is defined as the removal of the mammary gland and the nipple-areola complex (NAC) with breast skin sparing, whereas NSM saves both the NAC and the breast skin [1, 2, 3]. Both of these procedures aim to reduce the sense of mutilation after breast cancer surgery, particularly NSM, by avoiding NAC loss, which patients will not fully perceive until several weeks after surgery [2, 4, 5, 6, 7].

Candidates for SSM and NSM are generally subjected to an immediate reconstruction, which can be prosthetic or non-prosthetic, as well as in a single stage or two stages in the case of expander placement. Consequently, a further fundamental aspect to be taken into consideration when speaking of skin- or nipple-sparing mastectomy is that of perioperative complications, generally associated with reconstructive procedures, which can also significantly impact the patient’s well-being [8].

Women’s psychophysical wellness, which results strongly related to their individual perception of the aesthetic result, represents a critical target of breast reconstruction [2]. And actually, it is demonstrated by the current literature that psychophysical stress may correlate with bad oncological outcomes [9]. Therefore, it is imperative to achieve an optimal balance between oncological safety and aesthetic results [10].

Comparing SSM and NSM with the traditional modified total mastectomy (MTM), the current literature supports the oncological safety of these two procedures followed by immediate reconstruction, recognizes their evident improvement of cosmetic results to a high degree, and excludes any evidence of impairment of disease local control nor any increase of local or distant recurrence rates [11, 12, 13, 14], especially when combined with the use of radiation therapy and adjuvant chemotherapy, as indicated [15]. However, controversies still exist about the long-term oncological safety of these two procedures, and the data in the literature are not conclusive [16, 17, 18].

Our study aims to evaluate the oncological outcomes of women who underwent SSM and NSM in terms of overall survival (OS) and disease-free survival (DFS) and to analyze eventual risk factors for recurrence in patients treated by NSM, comparing them to MTM. The secondary aim was to compare the outcome of mastectomy to breast-conserving surgery (BCS).

Potentially eligible women treated because of invasive breast carcinoma during the study period were in 2036. Among these, we excluded 104 cases of TNM stage IV at diagnosis and 96 patients from the analysis because of incomplete data about axilla or breast surgery or follow-up. Finally, 1739 women were considered eligible for the present study, treated for 1836 invasive breast carcinomas.

In Table 1, we present the population characteristics. The 49% (899/1836) of the considered cancers were treated by BCS, 4.7% (87/1836) by NSM, 10.7% (196/1836) by SSM, and 35.6% (654/1836) by MTM. CALND was required in 49.6% (911/1836) of women, including cases of primary CALND and those of CALND secondary to a positive intraoperative SLNB. The median follow-up was 66 months (34–98), and the maximum follow-up was 120 months.

The mean age of the whole population was 59.14 years ($\pm$12.36). The youngest women were those who underwent NSM (mean age of 51.17 years ($\pm$10.21)) or SSM (mean age of 52.95 years ($\pm$10.27)) in comparison to the other two groups of BCS (mean age of 59.36 years ($\pm$11.00)) and MTM (mean age of 61.77 years ($\pm$13.80)) (p 0.05). In addition, women who underwent NSM had a significantly lower BMI in comparison to BCS (24.18 kg/m${}^2$ ($\pm$4.06) vs. 25.89 kg/m${}^2$ ($\pm$4.86)). Post-menopausal status prevalence was 77.8% (1429/1836) in the whole population but only 50.6% (44/87) in the NSM group.

Tables 2,3 show, respectively, tumor characteristics and TNM staging. The prevalence of Luminal A, Luminal Her, Her enriched, and Basal-like subtypes was similar in patients treated by BCS and NSM. Furthermore, women who underwent mastectomy (NSM, SSM, or MTM) had a higher prevalence of comedo-like necrosis, multifocality/multicentricity, extended intraductal component, and lymphovascular invasion than BCS. The prevalence of extra-capsular invasion of lymph node metastasis was similar between NSM and BCS. BCS had a significantly smaller tumor size and a more favorable nodal status than mastectomies (NSM, SSM, and MTM). However, the NSM group had a non-significant higher prevalence of N1, N2, N3, and G3 than the BCS group.

We analyzed in this population OS and DFS among the four groups of different breast surgery types (Figs. 1,2). The 5-year OS was 98.5% (95% confidence interval (CI), 97.6–99.4%) in the BCS group, 94.3% (95% CI, 88.2–100.0%) in NSM, 94.6% (95% CI, 91.1–98.1%) in SSM, and 95.1% (95% CI, 93.2–97.0%) in MTM. The 5-year DFS was 96.6% (95% CI, 95.2–98.0%) in BCS, 86.7 (95% CI, 76.7–98.0%) in NSM, 92.2% (95% CI, 88.2–96.4%) in SSM, and 90.4% (95% CI, 87.9–93.0%) in MTM. We found no significant differences in OS and DFS between SSM and MTM. At the same time, we found a significantly shorter DFS in NSM than MTM among TNM stage I cancers (p 0.05) (Fig. 2A).

Fig. 1.

Kaplan-Meier analysis of the entire population. Kaplan-Meier curves and cumulative events curves of the whole considered population (p-values in this figure refer to the Log-rank test). (A) Overall survival (OS) analyses (BCS had a significantly longer OS than NSM, SSM, or MTM p 0.05). (B) Disease-free survival (DFS) analyses (BCS had a significantly longer DFS than NSM, SSM, or MTM p 0.05). In particular, the difference between NSM and MTM was p = 0.269. (C) Cumulative local recurrence curve (BCS had a significantly lower prevalence of local recurrence than NSM, SSM, or MTM p 0.05). (D) Cumulative distant metastases recurrence curve (BCS had a significantly lower prevalence of distant metastases recurrence than SSM, or MTM p 0.05). In this case, we found no significant differences between BCS and NSM.

Fig. 2.

Kaplan-Meier analysis according to TNM stage. Kaplan-Meier curves of disease-free survival (DFS) analyses are subdivided by TNM stage (p-values in this figure refer to Log-rank test). (A) DFS analyses in TNM stage I. In this analysis, we found NSM to have a shorter DFS than MTM and BCS (p 0.05). We also observed a significantly longer DFS in BCS than in MTM (p 0.05). (B) DFS analyses in TNM stage II. This analysis found BCS to have a longer DFS than SSM and MTM (p 0.05). (C) DFS analyses in TNM stage III. This analysis found BCS to have a longer DFS than NSM (p 0.05).

Table 4 also shows an analysis using univariate and multivariate Cox proportional hazards regression models. In particular, in the multivariate analyses, we adjusted for the following factors: menopausal status, type of axilla surgery, neoadjuvant or adjuvant therapies, histological type, molecular sub-type, comedo-like necrosis, perivascular invasion, peritumoral inflammation, extra-capsular invasion of lymph node metastasis, bunched lymph nodes, TNM stage, and grading. After adjusting the multivariate analysis for these factors, we found BCS to influence DFS compared to MTM positively and NSM to have a significantly shorter DFS than MTM (p 0.05).

In particular, among the 87 cancers treated by NSM, we registered six local breast recurrences and, in three of these cases, also distant metastases. Among women experiencing any recurrence, we found a significantly lower BMI, higher prevalence of basal-like sub-type, extra-capsular invasion of lymph node metastasis, N2 status, and TNM stage III than the cases without recurrence. We also performed a univariate and multivariate logistic regression model analysis to define the most predictive factor for disease recurrence in women treated by NSM. In the initial logistic regression multivariate model, we included all factors showing in univariate analysis a p-value 0.300. After step-wise selection, the most predictive model had BMI (odds ratio (OR) 0.733, 95% CI, 0.533–1.008, p = 0.056), basal-like molecular sub-type (OR 28.932, 95% CI, 1.110–754.247, p 0.05), extended intraductal component (OR 11.160, 95% CI, 0.595–209.224, p = 0.107), and extra-capsular invasion of lymph node metastasis (OR 8.727, 95% CI, 0.792–96.165, p = 0.077). This final model held an AUC of 94.2% (95% CI, 88.8–99.6%) to predict recurrences in NSM.

In this study, we found NSM to have a significantly shorter DFS than MTM. Furthermore, low BMI, basal-like molecular sub-type, extended intraductal component, and extracapsular invasion of lymph node metastasis resulted in the most significant predictive factors for recurrence in women treated with NSM. We found no significant differences between SSM and MTM, while breast cancers treated by BCS had better OS and DFS than those who underwent a mastectomy.

The 5-year recurrence rate among women who underwent SSM and NSM in our population was respectively 7.8% and 13.3%. Despite the high variability in median follow-up lengths, in the literature recurrence rate of SSM and NSM range respectively from 0% to 14.3% [15, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36] and from 0% to 10.3% [25, 32, 34, 35, 37, 38, 39].

Some risk factors for recurrence among women who underwent NSM, such as basal-like subtype or extended intraductal component, are recognized risk factors for an unfavorable prognosis independently by surgical strategy [40, 41, 42]. By sparing the skin and, in some selected cases, the NAC, the risk of leaving some occult residual neoplastic foci and the recurrence rate increases [17]. Although NSM should be offered to women without any ascertained skin or NAC tumoral involvement, it is impossible to completely exclude the likelihood of the misdiagnosed presence of tumoral cells on the residual skin or NAC. For example, Mallon and colleagues [43] described an overall rate of occult nipple malignancy of 11.5%. Moreover, they saw the following primary tumor characteristics impacting occult nipple malignancy: tumor-nipple distance less than 2 cm, grade, lymph node metastasis, lymphovascular invasion, Her2/Neu overexpression, hormonal receptors negativity, tumor size greater than 5 cm, retro-areolar/central location, and multicentric tumors [43]. Among these factors, our receiver operating characteristic (ROC) curve confirmed the strong predictivity only of hormonal receptors negativity, as in our population, basal-like tumors treated by NSM were significantly more likely to experience local recurrence. However, the other risk factors for occult nipple involvement include many factors recognized nowadays as generally predictive of an unfavorable prognosis. Huang and coworkers [17] recently also found an overall rate of occult nipple malignancy of 12.6%. They also discovered that nipple signs, tumor location, tumor size, tumor-nipple distance, lymph node metastasis, and HER2 over-expression were all predictive of occult nipple malignancy [17]. Furthermore, other authors observed an increased probability of retro-areola positive histology directly correlated with the tumor size but not related to the nodal status [44]. However, these factors did not significantly predict recurrences after NSM in our study population.

For what concerns BMI, although a high BMI is recognized to be a risk factor for surgical complications after SSM and NSM [45, 46], the literature lacks its eventual role in the oncological outcomes. A confounding factor may be that low BMI has a probable direct correlation with small-volume breasts, which are more likely to undergo NSM in case of an unfavorable rate between tumor size and breast volume.

In agreement with the literature, non-surgical treatments had no significant influence on OS or DFS in our population of women who underwent SSM and NSM. In particular, many studies demonstrated that SSM with immediate reconstruction could be offered to patients who have been down-staged with neoadjuvant chemotherapy [14]. Moreover, even the performance of SSM and NSM did not significantly influence non-surgical treatment success. In particular, some authors remarked that SSM with immediate reconstruction prolonged overall cancer treatment time but did not impair oncological outcomes [33, 47].

In summary, we could explain the higher recurrence rate among women undergoing NSM through a possible occult nipple neoplastic involvement by negative intraoperative histological examination of subareolar tissue, as well as an occult skin involvement. NSM has often been offered to women with the multicentric disease, who were probably more likely to have occult neoplastic foci involving the NAC or residual skin. In this perspective, more attention should be paid to those patients who present one or more risk factors for recurrence after NSM, including basal-like molecular sub-type, extended intraductal component, and extracapsular invasion of lymph node metastasis. These patients should be cautiously counseled about the possibility of undergoing NSM and be made aware of their possible recurrence risk by the presence of one or more of the risk factors observed in this study to plan a more frequent better follow-up and promptly recognize any possible recurrence.

AJCC/UICC, American Joint Committee on Cancer/International Union Against Cancer; BCS, Breast-Conserving Surgery; BMI, Body Mass Index; CALND, Complete Axillary Lymph Node Dissection; DFS, Disease-Free Survival; EIC, Extensive Intraductal Component; ER, Estrogen Receptor; ExCp, Extracapsular Invasion; IBR, Immediate Breast Reconstruction; IRB, Internal Review Board; MTM, Modified Total Mastectomy; NAC, Nipple-Areola Complex; NSM, Nipple-Sparing Mastectomy; OS, Overall Survival; PR, Progesteron Receptor; PTI, Peritumoral Inflammation; PVI, Peritumoral Vascular Invasion; ROLL, Radio-Guided Occult Lesion Localization; SLNB, Sentinel Lymph Node Biopsy; SSM, Skin-Sparing Mastectomy; TNM, Tumor-Node-Metastasis; WLBB, Wire-Localized Breast Biopsy.

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Substantial contributions to conception and design (SB, APL, LM, PCP, AR, CC). Significant contributions to data acquisition (SB, LS, FP, SM, VM, MA, MO) or analysis and interpretation of data (SB, APL, CC). Drafting or revising the article critically for important intellectual content (SB, APL, LS, FP, SM, VM, MA, MO, LM, PCP, AR, CC). All authors have read and approved the final manuscript.

The present study was approved by the internal review board of the Department of Medical Area (University of Udine, approval number: #34/14), it was conducted in accordance with Helsinki Declaration, and it followed the dictates of the general authorization to process personal data for scientific research purposes by the Italian Data Protection Authority. According to national legislation, the need for informed consent was waived by the IRB listed above because this was a retrospective cohort study.

We would like to thank the whole staff for collaborating in article collection, selection, reading, and in paper writing and reviewing.

This research received no external funding.

The authors declare no conflict of interest. Serena Bertozzi is serving as one of the Guest editors of this journal. Ambrogio Pietro Londero is serving as one of the Editorial Board members and Guest editors of this journal. We declare that Serena Bertozzi and Ambrogio Pietro Londero had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to Felix Wong.