The Associated Clinicopathologic Characteristics of Nipple-Areola Complex Involvement in Breast Cancer

Background. At present,the indication for nipple-sparing mastectomy (cid:0) NSM (cid:0) remains inconclusive,and the occult NAC involvement (cid:0) NI (cid:0) is one of the most considerable problem when carrying out NSM.We aimed at identifying the predictive factors of NAC involvement,to provide an implement of selecting suitable NSM candidates. Methods. The association between the NI and tumorous clinicopathologic characteristics of 272 mastectomies and 31 therapeutic NSM performed in our hospital from to 2018 (cid:0) were

from 5.6% to 58%,this disparity may result from differentiation of subjects the researches had choosed, histologic sampling methods and de nition of positive NAC margin. the correlated clinicopathologic characteristics were also proposed in several studies,such as tumor size,tumor location,tumor-Nipple distence(TND [11][12][13][14].
This retrospective study had analyzed the rates of NAC involvement in mastectomy and NSM as well as its correlation with each clinicopathologic characteristics. We expected to observe the outcome of NSM, identify the associated factors in mastectomy and proposed a predictive model that can be useful in the surgical planning.

Methods
Three hundred female patients diagnosed as breast carcinoma between 2016 and 2018 in our hospital were included in our study,270 patients received mastectomies,30 patients received therapeutic NSM and breast reconstruction.We conducted NSM for patients who had wish to have NAC preservation and breast reconstruction,exclude only women with imaging evidence of NAC involvement,locally advanced breast cancer,in ammatory breast cancer,paget's disease.
All the information were extracted from case management system of our hospital,such as age,clinical nipple signs(nipple signs would be deemed as abnormality if there present nipple discharge,bleeding,retraction, ulceration,palpable mass, skin thickened .Tumor size was recorded in both preoperative ultrasound and gross pathologic samples.Tumor location was categorized to central/retroareolar and peripheral tumors.The shortest distance between tumor and nipple was measured as tumor-nipple distance(TND) on pathologic samples.The tumor multicentricity/multifocality was de ned as more than one lesion of invasive carcinoma separated by benign tissue.
The identi cation of tumor cells in the nipple margin sections would be de ned as positive,and we deemed NAC to be involved if they contained invasive cancer,ductal carcinoma in situ (DCIS), lobular carcinoma in situ(LCIS),or Paget's disease.For mastectomy samples The pathologic examination was done by vertical section of the nipple at 2 mm intervals,the sections were then stained using H&E stain and immunohistochemical staining method.For NSM, a separately submitted retro-areolar tissue was sent to intraoperative frozen section examination as a nipple margin and followed by para n sectioning examination.
The patients who underwent NSM were followed up,getting regular physical examination and further imaging tests if necessary,including computed tomography of the chest or brain,bone scan, and liver ultrasonography. Patients who were absent from examination would be contacted by telephone to con rm whether they were alive.
Student's T-test was used to evaluated continuous variables, and the chisquare test or Fisher's exact test was used to evaluated categorical variables. Binary Logistic regression was then performed to form a predictive model of NAC involvement. Only P 0.05 was considered statistically signi cant. The relationship between the tumor ultrasound size and histopathologic size was analyzed by Pearson Correlation Coe cient .Statistical calculations were performed by SPSS software (version 26.0).

Results
A total of 272 standard mastectomies and 31 nipple-sparing mastectomies were conducted, 40 cases(13.2%) has NAC involvement. As showed in table 1,we compared the characteristics of patiens in mastectomy group and NSM group.Despite there was no signi cant difference in NAC involvement between two groups,the NSM group had a lower incidence of NAC involvement than mastectomy group.For all mastectomies and NSM,the frequency of NAC involvement was 14.0%,6.5% respectively P=0.373 .Moreover, the signi cant difference had been observed in age and tumor size. patients in NSM group were yonger and had smaller tumor size than those in mastectomy group (mean age was 47 60 respectively P 0.05.mean ultrasound size was 1.90cm,2.42cm respectively,P 0.05.mean histopathologic size was 2.18cm,2.88cm respectively,P 0.05).The difference of lymph node status,tumor location,and nipple signs were not statistically signi cant.However,compared to mastectomy group, patients underwent NSM tended to have negative lymph node,peripheral tumor and normal nipple signs.
Among 30 patients who were planing to receive NSM, 1 patient underwent bilateral NSM.2 patients are proved to have a positive nipple margin.One of them converted to mastectomy due to identifying tumor cells in intraoperative frozen section of the submitted nipple margin.Another patient whose pathological examination showed lobular carcinoma in situ involving lactiferous ducts choosed to preserve the NAC.Nonetheless,it turned out no evidence of involment in the removing NAC at the routine para n sections after surgery.All the results of intraoperative frozen section were consistent to nal para n sections.
In a median follow-up duration of 24 months (range 3-42 months),none of 29 patients occurred distant metastasis.Only one patient was diagnosed as metastasis of rib 13 months after the operation and received local excision. All patients were con rmed to be alive at the last follow-up.

Discussion
Mastectomy is inevitable for some breast cancer patients,due to tumor size,tumor multicentric,type of tumor,or other reasons.The NSM and breast reconstruction technique provide a viable choice for these patients.NSM retains the natural envelope of the skin and nipple areola complex,achieving excellent aesthetic results. Neverthless,the preservation of NAC has also raised concern about occult nipple involvement,which has been discussed in many studies.In our study,the incidence of NAC involvement of patients underwent mastectomy and NSM were 14.0% 6.5% respectively.The former was similar to the incidence reported in several studies [15][16][17][18][19]. The difference between this two groups may resulted from that our surgeons has the inclination to select patients with normal nipple,smaller tumor size,peripheral tumor,and negative lymph node.However,of merely 30 cases in our study,the NAC involvement incidence in NSM may be underestimated.
The local cancer recurrence rates(1.7-10.3%) after NSM have been reported in published studies [8,[20][21][22], but they had different patients inclusion criteria and duration of follow-up. Smith et al performed NSM for stages 0 to 3 breast cancer patients who had no clinical or imaging evidence of NAC involvement,the rate of local recurrence was 3.7%(median follow-up duration of 51 months).In a study by Kim et al,the incidence of local recurrence was only 2% during a median follow-up of 60 months.The indication of preserving NAC was any stage, any tumor size, and any tumor areola distance tumors with clinically normal nipple. Burdge et al found a local recurrence of 10.3% in patients with locally advanced disease (71.8 % lymph node status was positive,average tumor size was 3.8cm.),and the median follow-up duration was 18 months.During a median follow-up of 24 months in our research,we found the local recurrence was low(3.2%) after NSM,and all patients survived at the last follow-up.To some extent,we had validated the oncological safety of NSM in selected patients.
According to the recommendation of latest NCCN guidelines(2020.v3), the nipple-areolar complex (NAC) could be preserved in cancer patients selected by experienced multidisciplinary teams.Paget's disease, nipple discharge associated with cancer,and imaging evidence of the nipple or subareolar tissues involvement should be excluded,and nipple margin assessment is de nitely needed.However,the standard assessment of NSM candidates and indications are still unde ned and inconsistent.In our study,frozen section examination of nipple margin(retro-areolar tissue) were conducted in all NSM cases,whose results were all consistent to permanent para n sections.One patient sacri ced the NAC due to a positive nipple margin in NSM,but the removed NAC turned out no cancer involved.Smith et al used frozen sections rarely in the procedure of NSM,because they deemed that frozen section was less accurate than permanent pathologic examination,and has di culty in distinguishing benign atypia from ductal carcinoma in situ (DCIS) [21].In a study by D'Alonzo, a discoid specimen beneath the base of the nipple was obtained as nipple margin.They found that the sub-areolar/nipple tissue assessment had a sensitivity of 42.3% in predicting nipple involvement.Though the frozen sections showed a great concordance(92.1%) with nal para n-embedded sections,it should be careful to remove the NAC when DCIS was diagnosed in the frozen section examination [23]. Additionally, Ponzone suggested to perform double intra-operative assessment of subareolar ducts and proximal nipple ducts,to increase the sensitivity of intraoperative pathological assessment [24].
Meanwhile we observed that patients with abnormal nipple signs(nipple discharge,bleeding,retraction, ulceration,palpable mass, skin thickened) were more likely to have NAC involvement(P 0.001). Billar found abnormal nipple signs or symptoms had a 61% sensitivity,86% speci city, 45% positive predictive value (PPV), and 92% negative predictive value (NPV) for determining NAC involvement [17].Though as one of the most common symptoms in breast cancer, nipple discharge is not the contraindication of NAC preservation if there was no evidence of tumor invasion to nipple margin.Since nipple discharge is not necessarily the outcome of tumor invasion to NAC, and NAC only acts as a drain channel when breast cancer invades ducts far from the NAC [25].As reported TND,tumor size,tumor location,lymph node status,and HER2 overexpression showed signi cantly correlation with NAC involment [26][27][28],which also had been demonstrated in our study.Weidong et al reported that patients with TND≤3cm,pathological tumor size 2.5cm,central tumor, lymph node metastasis or HER2 overexpression were at higher risk of getting NAC involved [27].Additionally,they considered the lymphatic vascular invasion(LVI) as another crucial predictive factor,which had been mentioned in other studies as well [29].Gulben et al categorized patients into different risk groups by using tumor location, number of axillary lymph node status, and LVI.Patients with two or three risk factors were classi ed among the high-risk group and patients with no or one risk factors were classi ed among the low risk group.The possibility of NAC involvement was 11.4 times higher in high-risk group compared to low risk group [30].
Though the effects of tumor size and TND had been con rmed,it has not been clearly determined the exactly cut-off gures on imaging or pathology.Study showed that the possibility of NAC involvement was lower when TND 1cm measured by mammography or MRI.Especially for MRI,its NPV and PPV of predicting NAC involvement were 100%,39% respectively [26].Steen et al found TND 2cm,tumor size 2cm measured by MRI were signi cantly associated with nipple involvement [31].A study of patients with carcinoma in situ by Hwang et al reported that short TND and suspicion of nipple involvement on mammographic or MRI were predictive factors of NAC involvement,but ultrasound ndings were not signi cantly correlated with NAC involvement.However,Lim found the tumors contacting or invading the nipple on US was signi cantly correlated with nipple involvement [32].In our research,all patients accepted ultrasound test preoperatively,which described and recorded the tumor location and tumor size.We found the incidence of NAC involvement in patients with tumor ultrasound size 4cm ≤4cm were 36.7% 10.3% respectively(P 0.001). Tumor size measured by ultrasound was signi cant predictive factors of NAC involvement,and Pearson correlation coe cient analysis demonstrated preoperative ultrasound can accurately represent histopathologic tumor size.Therefore,we con rmed the functional role of preoperative ultrasound test and recommend to apply it in the management of NSM.
As for the predictive model A recent study produced a preoperative predictive model using seven factors,including MRI tumor size≥4cm, mammographic TND 1cm,MRI TND 1cm MRI nipple enhancement,central tumor,multicentric/multifocal, clinical node involvement.Each factor scores 0 or 1 point,and the total scores were categorized into low (0-3), intermediate (4), or high (5-7) risk group.Patients in high risk group were recommended to sacri ce the NAC,and patients in the intermediate risk group who hoped to preserve NAC could undertake the frozen section examination [33].Besides,another study by Wang at al proposed a model consisting of tumor location, nuclear grade,and HER2 expression [28].Schecter at al reported a formula for predicting NAC involvement based on tumor size,TND and stage,which was found to have a sensitivity of 92%, speci city of 77% [34].The models developed in these studies are basing on more costly image testing or preoperative biopsy.The factors including pathological TND,tumor size,number of metastatic lymph node,HER2 overexpression were generally known after surgery.Whereas,parameters of the predictive model in our study are much more easier to obtain.Basing on ultrasound tumor size,tumor location,and nipple signs,a predictive model were proposed to provide the possibility of NAC involvement.We found breast cancer patients with abnormal nipple signs,in central location, and tumor size 4cm,were most likely to develope NAC involvement (95.4%);patients with normal nipple signs,tumor in peripheral location,and tumor size ≤4cm had a lowest possibility(4.1%) of NAC involvement.

Conclusion
In summary,our study shows that the local recurrence rate after NSM is low 3.2% in our research NSM appears to be an oncologically safe procedure for selected patients with negative nipple margin.NAC involvement had a incidence of 14.0% in mastectomy patients and was associated with the nipple signs,tumor location,tumor size measured by ultrasound or gross pathologic samples,tumor-nipple distance,lymph node metastasis,HER2 overexpression.We developed a predictive model based on nipple signs,tumor size and tumor location,which aimed to improve the accuracy of selecting eligible patients for NSM,by combining with pathological examination of retro-areolar tissue,to conduct NSM safely.