Nipple-areolar complex malposition in breast reconstruction after nipple-sparing mastectomy: a multi-institutional retrospective observational study in Japan

Background Position of the nipple-areolar complex (NAC) is an important factor in the aesthetic impression of the breast, and NAC malposition is often an issue in breast reconstruction after nipple-sparing mastectomy (NSM). The purpose of this study was to evaluate the degree of NAC malposition depending on several factors using data quanti�ed with the Mamma Balance application (Medic Engineering K.K., Kyoto, Japan). Methods Patients who underwent unilateral breast reconstruction after NSM at 8 hospitals in Japan between 2007 and 2020 were retrospectively investigated. Using Mamma Balance, NAC malposition was quanti�ed separately in horizontal and vertical directions using patient photographs from pre-operatively and 6–24 months post-operatively. The degree of malpositioning was then statistically compared using various factors.


Introduction
Nipple-sparing mastectomy (NSM) is an oncological surgical procedure that has been accepted for certain patient populations.This procedure allows patients to preserve the native nipple-areolar complex (NAC) and facilitates immediate, oncologically safe breast reconstruction [1].However, cranial and lateral NAC malposition is often reported as an issue with breast reconstruction after NSM [2][3][4].The position of the NAC represents an important factor in the aesthetic impression of the breast.The presence of the NAC is known to be associated with improved self-satisfaction and post-operative satisfaction in cases with serial mastectomy and breast reconstruction [5,6].
Many reports have described NAC malposition after NSM.While some reports have mentioned revision surgery for a malpositioned NAC, few have investigated the causes or risk factors.In addition, few reports have utilized quantitative, objective, unbiased measures to analyze NAC malposition in NSM [4,7].Almost all of those few studies have included relatively small cohorts [3,8,9] and many reports have used photographs to assess NAC malposition.The recently developed Mamma Balance application (Medic Engineering K.K., Kyoto, Japan) allows the position of the NAC to be quantitatively evaluated [10].We designed the present multi-institutional retrospective observational study of NAC position with breast reconstruction after NSM with photographs and the Mamma Balance application.The purpose of this study was to evaluate the degree of NAC malposition depending on the following factors: use of implants or aps; degree of ptosis; ap type; incision line; post-operative complications; use of the pull-down procedure; breast size; and post-operative radiation therapy.

Patients and Methods
We undertook a retrospective study of patients who had undergone unilateral breast reconstruction after NSM at the following facilities from 2007 to 2020: Tokyo Medical and Dental University Hospital, Okayama University Hospital, Hiroshima University Hospital, The Cancer Institute Hospital of JFCR, Yokohama City University Hospital, Japan Community Health Care Organization Saitama Medical Center, St. Luke's International Hospital and Mie University Hospital.Subjects in this study were patients with photographs from pre-operatively and 6-24 months post-operatively.Patients who had undergone contralateral revision, insertion of tissue expander before implant replacement on the affected side, or absence of the NAC on post-operative photographs were excluded.
With Mamma Balance, two triangles are automatically created with vertices of the sternal notch, the nipple on either side and a point on a vertical line from the sternal notch to the umbilicus or xiphoid process.Mamma Balance provides information on the area of the resulting triangle in pixel values (S) and the angle between the midline and the line between the sternal notch and nipple (θ).We de ned the following three lengths: from sternal notch to nipple (O); vertical line of the triangle (V); and horizontal line of the triangle (H).These values were calculated using the following formulas (Fig. 1): We assumed that lengths V and H on the unaffected side would not be changed post-operatively.We then calculated the post/pre-operative ratio of lengths V and H on the reconstructed side, corresponding to the degrees of post-operative craniocaudal and mediolateral malposition of the NAC.A post/pre-operative ratio for V less than 1 indicates cranial NAC malposition, while a ratio for H less than 1 re ects medial NAC malposition (Fig. 2).Differences in malpositioning due to the method of breast reconstruction, breast ptosis, ap type, incision type, breast amount, post-operative complications (hematoma, skin envelope necrosis, ap necrosis, circulation disorder of the NAC, and infection), pull-down xation of the NAC, post-operative radiation therapy and the correlation between transition of NAC position and amount of mastectomy (in grams) were evaluated statistically.Surgical complications were classi ed using the Clavien-Dindo classi cation.
Statistical analysis was carried out using Prism 9 software (GraphPad Software, San Diego, CA).Values of p < 0.05 were considered signi cant.The Mann-Whitney U test, Kruskal-Wallis test and Dunn's test with Bonferroni correction were applied for group comparisons.Spearman's rank correlation coe cient was used to examine correlations.
No signi cant difference was observed according to the degree of breast ptosis (Fig. 4).
Comparing by incision line, a signi cant difference in the frequency of lateral malposition was seen with aps (p = 0.0005), but not with implants (Fig. 5).
No signi cant difference was observed according to the presence or absence of the pull-down operation (Fig. 6).
Comparing the absence of post-operative complications and varying degrees of complications (hematoma, skin envelope necrosis, ap necrosis, infection, circulatory disorder of NAC), post/pre-operative ratio of V differed signi cantly between no infection and grade 3 infection (Table 2).Post/pre-operative ratio of V was signi cantly smaller with radiation therapy than without radiation therapy for both autologous and implant reconstruction ( ap †p = 0.022, implant ‡p = 0.014).
The correlation between amount of mastectomy (in grams) and NAC position transition is shown in the bottom row of the table as r (p-value).Spearman's rank correlation coe cient was applied.In cases with ap reconstruction, a very weak negative correlation was identi ed between lateral transition and amount of mastectomy (r=-0.214,p = 0.019).In cases with implant reconstruction, a very weak negative correlation existed between vertical transition and amount of mastectomy (r=-0.200,p = 0.034).

Flap
(n) V: ap V:implant (ratio) We sorted the degree of post-operative complications according to the Clavien-Dindo classi cation.
Grade 2: Requiring pharmacological treatment with drugs other than those allowed for grade 1 complications.Blood transfusions, antibiotics and total parenteral nutrition are also included.
With the use of radiation therapy, post/pre-operative ratio of V was signi cantly smaller than that in radiation-free cases with both autologous and implant reconstruction ( ap: 0.93, p = 0.022; implant: 0.87, p = 0.014) (Table 2).
In terms of the correlation between transition of NAC position and amount of mastectomy (grams), in ap reconstruction cases, correlation coe cients were − 0.21 for lateral transition (p = 0.019) and − 0.15 for vertical transition (p = 0.111).In cases with implant reconstruction, correlation coe cients were − 0.12 for lateral transition (p = 0.199) and − 0.20 for vertical transition (p = 0.034) (Table 2).

Discussion
This study revealed the following results: The NAC was deviated more cranially and medially with implant reconstruction than with ap reconstruction; LD ap cases showed lateral deviation more often than DIEP ap cases; With ap reconstruction, lateral incisions were associated with lateral malposition and peri-areolar incisions were associated with medial NAC malposition; The NAC tended to be deviated cranially more often with severe post-operative implant infection than with no infection; The NAC deviated cranially with radiation in both implant and ap groups; A very weak correlation was apparent between larger mastectomy volume and greater cranial malposition of the NAC in both ap and implant cases.
Various studies have examined NAC malposition in terms of grades [11], angles [9], subjective evaluations [12], vertical and horizontal balance [4], and more.Mamma Balance is original software that expresses NAC malposition in terms of area ratio.At this time, by replacing date from Mamma Balance with length, we were able to compare pre-and postoperative values in both vertical and Horizontal directions as continuous variables.This report is the rst to quantify the degree of NAC malposition and compare risk factors in detail for a large number of cases.
Previous studies have shown that both autologous and implant reconstructions carry a risk of NAC malposition (implant [3,13]; autologous [14,11]).In this study, implant reconstruction was seen to result in a signi cant risk of NAC cranial medial malposition.In contrast, the NAC tended to move more laterally with autologous reconstruction.These issues have been described in previous reports [7,10].With autologous reconstruction, LD aps tended to show more lateral deviation compared to DIEP aps.This may be because the origin of the LD muscle is in the upper arm, and contraction pulls the NAC laterally.The use of LD aps may have affected the difference in lateral malposition between autologous and implant cases.The procedure for dealing with LD muscle insertion (whether to cut the insertion or not) and thoracodorsal nerve differ depending on the facility, so one possibility is that NAC malposition occurred due to the outward pulling force of the musculature, particularly in facilities without disconnection of the LD muscle insertion.
These ndings are valuable because differences in NAC malposition among types of autologous ap have not previously been reported.
Past studies have identi ed severe breast ptosis and breast size as risk factors for NAC asymmetry [8,16].Many reports have stated that NSM is unsuitable for severely ptotic breasts [9,17], but no studies have examined associations between the degree of breast ptosis and NAC malposition.
This study found no signi cant associations between the degree of breast ptosis according to the Regnault classi cation and either implant reconstruction or autologous tissue reconstruction.However, the number of cases with severe ptosis was small, which may have been why no signi cant differences were identi ed.In addition, only very weak correlations were observed between the amount of mastectomy and NAC malposition with both ap and implant reconstructions.
One possibility that cannot be excluded is that some degree of selection bias may have existed in cases for which the weight of resected breast tissue was recorded, so the results of this retrospective study should be considered only as a reference.
According to a past report, a vertical or Wise pattern incision is a risk factor for NAC malposition [9].In our study, a signi cant difference was observed depending on the incision line used with ap reconstruction, but not with implant reconstruction.The NAC was thought to be pulled laterally by contraction of scar tissue forming at the incision line, but LD aps tended to result in lateral deviation that might have affected this result for incision lines in ap reconstruction.
Secondary healing after necrosis of the NAC or skin envelope is a risk factor of NAC malposition [17].However, Choi et al. reported that necrosis of the NAC is not an obvious risk factor for malposition [9].In this study, skin envelope necrosis and circulatory disorder of the NAC were not found to cause signi cant malpositioning of the NAC.This may have been a result of the small number of cases with these complications, but was still a new nding worth further investigation.The association between post-operative infection and NAC malposition has not yet been reported.
Radiation therapy has been reported to cause NAC malposition [8], but some studies have found no association between radiation therapy and severe NAC malposition warranting resection [17,18].In our study, both autologous and implant reconstructions caused signi cant NAC malposition in irradiated patients, supporting the nding that radiation therapy caused NAC malposition.
NAC xation by suture or the pull-down procedure using tape have been reported to prevent NAC malposition [19][20][21].In our cases, no signi cant difference was observed according to the presence or absence of the pull-down operation.However, particularly with implants, some cases show severe cranial NAC malposition without the pull-down operation.
Many reports have described revision operations for NAC malposition, but NAC xation or pull-down measures for breast mount reconstruction appear effective and should be actively adopted to reduce patient burdens, both physically and mentally.Although there was no statistically signi cant difference in this study, this does not mean that the pull-down operation is unnecessary.Further case collection and study are required.
These results should be considered when providing full information to patients making decisions on reconstructive methods.Patients who choose implant-based breast reconstruction should be aware of the risk of cranial malposition of the NAC during the post-operative course.The same applies for lateral malposition with LD ap reconstruction.Obviously, surgeons act to reduce post-operative complications and optimize the results of breast reconstruction.
Moreover, we should consider adopting maneuvers like the pull-down method concurrently with formation of the breast mount.
Various limitations of this study need to be kept in mind.First, almost all patients were Asian and we were unable to consider individuals from other ethnic backgrounds.Second, we assumed that the position of the NAC on the unaffected side before and after surgery would not change, and that imaging conditions would not be perfectly constant.Third, subcutaneous implant-based reconstruction was not considered because subpectoral implants are recommended by the guidelines used in Japan.Moreover, we could not consider the expansion speed of tissue expander cases, although early expansion reportedly reduces the risk of NAC malposition [10].Finally, we could not completely remove the possibility of selection bias.Since this study was conducted as a retrospective analysis of cases from multiple institutions, surgical techniques could not be standardized.

Conclusion
We investigated NAC malposition following breast reconstruction after NSM in a large number of cases.
Implant cases were more prone to cranial deviation than ap cases.In addition, we were able to obtain some insights into the tendencies and characteristics of NAC malposition.We should properly assess risks and proactively apply countermeasures to achieve the optimal results from NSM.
Procedure Assuming that V and H on the unaffected side are constant pre-and post-operatively, we calculated the preoperation/post-operation(pre/post-op) rate for V and H on the affected side: f: V ratio <1; g: V ratio >1; h: H ratio <1; i: H ratio >1."1" represents the same position as the unaffected side.
Flap vs implant, ap types for each item indicate mean and standard deviation.We applied the Mann-Whitney U test for the post/preoperative ratio of ap reconstruction (n=160) and implant reconstruction (n=196).Three fat graft and 1 latissimus ap + implant cases were excluded.The post/pre-operative ratio of vertical position (V) was signi cantly smaller with implants than with aps (p<0.0001,implant: 0.91, ap: 0.98).Moreover, post/pre-operative ratio for Horizontal position (H) was signi cantly smaller with implants than with aps (p=0.0209,implant: 0.98, ap: 1.01).
for using Mamma Balance and classifying NAC malposition Upper: a: A clinical photograph is taken b: The photograph is uploaded into Mamma Balance and the midlines of the clavicles, both nipples and the navel or xiphoid process are marked.c: The application makes two triangles: formed by midline of the clavicle, midline of the body and both nipples.The blue dashed line triangle represents the unaffected side, the red triangle represents the affected side.d: The application calculates four numbers: the angle formed by the nipple, midline of the clavicle and midline of the body; the areas of the triangle formed by these points; the common area of these triangles; and the common area of these triangles as a percentage of the total area within each triangle.In this picture, the blue line is a part of mirror-reversed dashed triangle of Figure The red line is a part of red triangle of Figure1-c.The green triangle represents the common area.e: Values of V and H are determined from θ and area S (pixel value) using the below formulae: O 2 =2S/cosθ×sinθ V=√(2S/cosθ×sinθ)×cosθ H=√(2S/cosθ×sinθ)×sinθ Lower: Regarding the type of ap, Kruskal-Wallis test and Dunn's test with Bonferroni correction were used.A signi cant difference was found for lateral malposition DIEP ap vs LD ap (p=0.0344).P-values are not shown for groups with no signi cant difference.Dashed lines indicate median and interquartile range in the violin plot.

Table 1
Patient data are given as the average (range for age, standard deviation for BMI and follow-up).

Table 2
Correlations between amount of mastectomy, post-operative complications and radiation therapy