For many individuals with longstanding gender dysphoria, surgical modification of their primary and secondary sex characteristics is essential to alleviate symptoms and establish congruity with their gender identity.1,2 Benefits of sex reassignment surgery include improvements in well-being, sexual function, and cosmesis.3–6 For female-to-male patients, creation of an aesthetically pleasing male chest facilitates living in the male gender role before genital surgery. Subcutaneous mastectomy is commonly the first operative step toward sex reassignment for phenotypic female patients with gender dysphoria.7–12
The goal of subcutaneous mastectomy and male chest contouring is to produce a normal appearing male chest. This is achieved by removal of the breast parenchyma, obliteration of the inframammary fold, contouring of the mastectomy flaps to produce lateral and inferior pectoral definition, and appropriate sizing and positioning the nipple-areola complex. Scars are minimized and nipple-areola complex and chest wall sensation is preserved if possible.10–12 Ideally, all of these tasks should be performed in a single stage to minimize psychological stress and optimize recovery.
Many traditional techniques used to treat gynecomastia have been used for subcutaneous mastectomy in these patients.7,12–19 However, traditional techniques often fall short of achieving all the goals of subcutaneous mastectomy and male chest contouring—a procedure that is more challenging than most gynecomastia procedures. Adding to the complexity is the common practice of breast binding, which can produce increased ptosis and excess skin of poor quality.
Currently, there is no consensus regarding which technique produces the best aesthetic appearance with the lowest complication and revision rates. Monstrey et al.12 described results using five techniques: semicircular,19 transareolar,18 concentric circular,15 extended concentric circular,10 and a breast amputation with a free nipple graft.14 The semicircular and transareolar techniques minimize scarring20,21 but do not adequately address the excess skin often present after removal of glandular tissue. Furthermore, removing the inframammary fold and contouring the chest wall is difficult with these approaches. For these reasons, many authors favor techniques using longer incisions rather than risking a suboptimal result bearing resemblance to the previous female form.
The concentric circular and the breast amputation with free nipple graft techniques have proven to reliably and adequately address all of the goals of female-to-male chest contouring.22 This article describes a single surgeon’s experience with 101 consecutive female-to-male transgender patients undergoing subcutaneous mastectomy using either the concentric circular or free nipple graft technique—the two techniques exclusively used at our institution. Our objective was to compare clinical and aesthetic outcomes of concentric circular and free nipple graft patients and identify anatomical characteristics that may assist with selection of the appropriate chest contouring technique for transgender patients.
PATIENTS AND METHODS
Female-to-male transgender patients undergoing bilateral subcutaneous mastectomy performed by the senior surgeon (C.C.B.) were identified from the University of British Columbia Sex Reassignment Program between 2006 and 2011. Patients were excluded if they presented for a revision of a subcutaneous mastectomy performed elsewhere. Approval for this project was obtained from the University of British Columbia Clinical Research Ethics Board.
Data were collected retrospectively from electronic medical records following institutional ethics review board approval. Each breast was treated as an independent observation. Data collected included demographic information, comorbidities, preoperative breast measurements, surgical details, postoperative complications, need for additional operations, and aesthetic outcome. Age and body mass index were analyzed as continuous variables. Skin elasticity was categorized into lax or thin, moderate, and good to excellent. Preoperative measurements included sternal notch–to-nipple distance (in centimeters), nipple-to–inframammary fold distance (in centimeters), base width (in centimeters), and nipple-areola complex diameter (in millimeters) (Fig. 1). Ptosis was categorized into none, grade 1, grade 2, grade 3, and glandular ptosis according to the Regnault and Brink classification systems.23,24 The primary independent variable was surgical procedure (concentric circular versus free nipple grafting). Dependent variables included postoperative complications and revisions, which were treated as binary variables. The presence of one or more complication or revision was considered an event and was counted once per breast for our calculation of complication and revision rates. Rates of each individual complication were also recorded. Postsurgical nipple-areola complex sensation was categorized into none, decreased, or no change based on patient-reported ratings.
Surgical Protocol and Operative Details
Patients were selected to undergo either the concentric circular or free nipple graft technique based on subjective assessment of breast volume, degree of ptosis, and skin elasticity. Those with larger, more ptotic breasts or borderline cases with poor skin elasticity were offered only the free nipple graft technique, as it was felt that the goals of adequate skin excision and ideal nipple-areola complex placement would be unattainable with the concentric circular technique. Patients with smaller breast size, minimal ptosis, and good skin elasticity were offered the concentric circular technique because of its advantages of reduced scar burden, a more natural appearing nipple-areola complex, and the potential for maintenance of nipple-areola complex sensation.
The operative sequence for the free nipple graft technique is shown in schematic format in Figure 2. The operative sequence for the concentric circular technique is shown in Figure 3. In both techniques, the inframammary fold was removed en bloc with the mastectomy specimen. Postoperative care for both techniques includes closed suction drainage, and a compression vest for 4 to 6 weeks.
The following complications were recorded: infection, mastectomy flap or nipple-areola complex necrosis, hematoma, and wound dehiscence. Major complications were defined as those requiring operative management.
Revision surgery to improve aesthetic outcome included surgery to improve scars, surgery to improve chest contour (including liposuction, skin reduction, dog-ear corrections, and fat grafting), and nipple-areola complex revisions (including reduction, areolar reshaping, nipple reconstruction, and nipple tattooing). Several patients underwent more than one revision at the time of surgery.
Standardized postoperative photographs were obtained before any revision surgery. Four staff breast surgeons and four senior residents reviewed the photographs and scored results using a three-item, five-point Likert scale, where 1 represents a poor outcome and 5 represents an excellent outcome. The scale items included chest contour (i.e., appearance of male chest, lateral pectoral definition, resemblance to previous female form, and overall symmetry), scar (i.e., position of scar, widening of scar, and symmetry), and nipple-areola complex (i.e., size, position, symmetry, and overall aesthetics). Before rating, sample photographs of both free nipple graft and concentric circular technique results that included poor to excellent outcomes were used for rater training and calibration. All of the original postoperative photographs were reviewed in two separate sessions to determine intrarater reliability.
Means and frequencies were calculated for continuous variables and categorical data, respectively; t tests were used to compare means between groups, and the Pearson chi-square test was used to compare binary variables where applicable. Logistic regression analysis was used to evaluate the relationship between procedure type (i.e., concentric circular and free nipple graft) and complications/revision surgery. Adjusted odds ratios and 95 percent confidence intervals were calculated. Classification and regression tree models were constructed using the patient’s characteristics and preoperative measurements.25 Ten-fold cross-validation was used to find the best tree that most accurately predicted technique selection. Intrarater and interrater reliability were calculated for each of the aesthetic rating subscales and the total sum of the subscales using intraclass correlation.26 Average aesthetic scores were calculated and compared between the two procedures. A two-tailed value of p < 0.05 was selected to indicate statistical significance. Analyses were performed using the Stata 11.0 (StataCorp, College Station, Texas) software package and CART software version 6.6 (Salford Systems, San Diego, Calif.).
A total of 101 patients underwent 202 subcutaneous mastectomies (concentric circular, n = 92; free nipple graft, n = 110). The groups were comparable, with the exception of older age and higher body mass index in the free nipple graft group (Table 1). In keeping with our selection algorithm, concentric circular patients had smaller, less ptotic breasts with better skin elasticity. The majority of patients (90 percent) self-reported on their nipple-areola complex sensation such that all breasts in the free nipple graft group lost sensation, whereas 46.7 percent of the concentric circular group maintained full or partial sensation.
A greater proportion of total complications occurred in the concentric circular group compared with the free nipple graft group (33.7 percent versus 10.9 percent; p < 0.001). A total of 36 breasts (17.8 percent) had one or more minor complication, with a greater number of complications in the concentric circular group (27.2 percent versus 10.0 percent; p = 0.001). Major complications requiring a return to the operating room occurred in 13 breasts (6.4 percent), with a greater number occurring in the concentric circular group (13.0 percent versus 0.91 percent; p < 0.001). Table 2 demonstrates the breakdown of major and minor complications and shows that major complications in the concentric circular group usually involved the nipple-areola complex (necrosis and delayed healing/dehiscence). After adjusting for body mass index, age, smoking, and ptosis, the odds ratio of total complications remained statistically significant, with 3.0 times the odds of complications in the concentric circular group compared with the free nipple graft group (p = 0.03). Increased age was a significant predictor of a worse outcome in either group, with smokers having 2.48 times the odds of a complication (p = 0.043).
A total of 48 breasts (23.7 percent) required revisions to improve the aesthetic outcome, 34 (37.0 percent) in the concentric circular group and 14 (12.7 percent) in the free nipple graft group (p < 0.001). There were 4.0 times the odds of revision surgery in the concentric circular group compared with the free nipple graft group (p < 0.001). Two revision procedures were performed simultaneously in 26 breasts (12.9 percent). Two breasts (1 percent) required three simultaneous procedures. The indications for revision are listed in Table 3. The concentric circular group had the largest proportion of revisions, with nipple- areola complex revision procedures making up the majority (32 breasts) (p < 0.001). Conversely, only one patient in the free nipple graft group underwent a nipple-areola complex revision.
A total of 33 concentric circular patients (71.7 percent) and 43 free nipple graft patients (78.2 percent) had follow-up photographs available. Table 4 lists the intrarater and interrater agreement and mean rating scores between the two groups (for information on interpretation of intraclass correlation, see Fleiss27 and Rosner28). The nipple-areola complex mean rating was not significantly different between groups (3.34 versus 3.14; p = 0.113). The mean scar rating in the free nipple graft group was significantly lower than the mean rating in the concentric circular group (2.62 versus 3.39; p < 0.001). Similarly, free nipple graft contour was rated significantly lower than concentric circular contour (3.34 versus 3.82; p < 0.001).
As part of classification and regression tree analysis, multiple factors believed a priori to be predictors of the best technique for mastectomy/skin reduction were analyzed. Nipple-to– inframammary fold distance and body mass index provided the best model with which to predict technique based on patient factors. By using the criteria of nipple-to–inframammary fold distance and body mass index, 93 percent of patients were correctly sorted into either the concentric circular or free nipple graft group (Fig. 4 and Table 5). Of the eight patients who, based on classification and regression tree analysis, should have received the alternative procedure, five experienced complications and one required revisions (Table 6).
Enhanced awareness and increasing access to transsexual health services in British Columbia have resulted in an increased demand for sex reassignment surgery. In our opinion, the concentric circular and free nipple graft techniques adequately address all of the goals of female-to-male chest contouring and limit the number of complications and required revisions. Although minimal scar approaches, such as the semicircular technique based on Webster’s approach to gynecomastia,19 provide good results in individuals with very small breasts,20,21 they do not allow for adequate skin excision or areola reduction in the vast majority of patients. Other techniques17 that leave breast tissue behind as a pedicle for the nipple-areola complex will have an overly full contour, leaving an unacceptable resemblance to the female form.
Technique selection can be challenging. Preoperative considerations include chest wall anatomy, body habitus, skin quality, and preoperative breast size and shape.11 Previously, authors have categorized patients using subjective estimates of breast size, ptosis, or elasticity.10,12,20 However, objective data, such as standard preoperative breast measurements, would be useful to guide surgeons in technique selection. Using classification and regression tree analysis (Fig. 4), we were able to identify specific values of well-known and easily attainable preoperative measurements that matched the greatest number of patients with the appropriate operative technique. Classification and regression trees “take a more inductive approach to building a model than is done in traditional regression, allowing the data greater role in suggesting the correct relationships between variables rather than imposing them a priori.”29 This analysis can identify useful predictors of an outcome and discover interactions between predictors without the need to anticipate and specify these in advance. These trees can then provide a predictive model for forecasting outcomes as an aid to clinical decision-making. The objective of using classification and regression tree analysis in our study is to identify specific predictor variables and determine the optimum cut point for each variable that will correctly place individual patients in the correct treatment arm. After extensive computer modeling, nipple-to–inframammary fold distance and body mass index were identified as the two predictor variables that most accurately sorted all patients into either the concentric circular or free nipple graft categories.
The free nipple graft technique (Fig. 5) provides the surgeon with the most control over chest contouring and nipple-areola complex placement. However, this results in an unnatural appearing and insensate nipple-areola complex and longer scars. The disadvantage of a longer scar can be partially diminished by placing the scar along the inferolateral border of the pectoralis major muscle, as it is more harmonious with the resulting masculine contour.
The concentric circular technique (Fig. 6) addresses the goals of female-to-male chest contouring and results in far less scar burden and a natural appearing and potentially sensate nipple-areola complex. However, contouring of the chest is more challenging, as the surgeon is working through a small window. In addition, the concentric circular technique is limited by the amount of skin that can be removed, and if attempted in larger breasted patients, it would cause an unacceptable amount of puckering or tethering. Based on results of classification and regression tree analysis, we recommend that a patient with large ptotic breasts (nipple-to–inframammary fold distance >7 cm) or a smaller breasted patient (nipple-to–inframammary fold distance <7 cm) who has an increased body mass index (body mass index >27 kg/m2) should undergo the free nipple graft technique.
The concentric circular technique was associated with higher complication and revision rates. Even when controlling for smoking, body mass index, age, ptosis, skin elasticity, and breast binding, the increase in complications and revision surgery in the concentric circular group persists. This suggests that despite having better preoperative characteristics, this technique imparts an increased risk of complications that may require additional surgery. Most complications and revisions in this group were nipple-areola complex issues. We believe that this is a result of a significant reduction in blood supply to the nipple-areola complex, which increases the risk of delayed wound healing, widening of the nipple-areola complex, and poor scarring. Strategies to preserve nipple-areola complex blood supply are outlined in Table 7.
Problems with the concentric circular technique are somewhat mitigated by the observation that the concentric circular group has a preferable aesthetic outcome, even before undergoing revision surgery. Interestingly, increased complications with the nipple-areola complex in the concentric circular group did not correlate with an aesthetic difference between the nipple-areola complex in the concentric circular and free nipple graft groups. The concentric circular technique is limited by the amount of skin that can be removed, and if attempted in larger breasted patients, it would cause an unacceptable amount of puckering or tethering. We did not specifically measure the final skin excision lengths for each patient, although all were under 8 cm. If some concentric circular patients required inappropriately large incisions resulting in stretched-out, puckered, or misshapen areolae and poor scars, we would expect higher rates of revision in the scar category and nipple-areola complex. Perhaps negative outcomes associated with nipple-areola complex complications are offset by a more natural appearance and a resulting location and symmetry that are almost always excellent. The free nipple graft nipple-areola complex has an artificial appearance that may include a “stuck-on” appearance, pigmentation changes, and effacement of the nipple itself.
The concentric circular group was also superior for the aesthetic appearance of chest wall scars. Unexpectedly, the concentric circular group also demonstrated superior chest contour. This is surprising, as the free nipple graft technique allows the most control over chest wall shaping. It is possible that the poor preoperative characteristics of the free nipple graft patients (increased body mass index with a poorer body habitus) reflect a more displeasing body contour overall—thus negatively influencing the aesthetic scores.
Standardized photographs were used for our aesthetic analysis. We have attempted to minimize bias and demonstrate reliability by using multiple expert raters, providing rater training, using multiple blinded rating sessions, rating photographs before revision surgery, and performing interrater and intrarater analysis. Limitations inherent in the use of photographs include the inability of raters to assess skin quality and compliance. Also, because follow-up photographs were not available for all patients, there is a potential risk for selection bias. Canada is geographically very large with a relatively small population. There are only two gender reassignment programs in the entire nation, meaning that gender dysphoria patients often travel very long distances to have surgery. If the initial postoperative course is uneventful, it can be very challenging to convince them to return for regular follow-up. However, we would expect self-selection to be weighted toward poorer outcomes, because those with problems would be more likely to return to plan and receive revisions, although it is possible that some patients may have moved or visited other surgeons. Despite a higher revision rate in the concentric circular group, our follow-up rates were similar.
One limitation is the variation in dates of follow-up photographs. Ideally, a prospective study with 1-year follow-up would be the best way to examine these patients. Because our aesthetic assessment is relatively early in the postoperative course, some raters could have a less favorable impression of a result because of scar immaturity or early postoperative swelling influencing contour. Obtaining additional photographs of the long-term result (i.e., >1 year) for comparison would be beneficial; however, we are constrained by the logistics of coordinating long-term follow-up in this patient population. Furthermore, many of these patients would have had revisions that may bias the aesthetic analysis. Rather than publish results based on a much smaller proportion of patients with long-term follow-up, our goal was to assess at a standardized postoperative time point, capturing a larger proportion of patients. To compare results fairly, we felt it was important that this initial comparison be made at a standardized time point before any patient underwent revision surgery.
Another limitation was that the scale for scoring the aesthetic analysis did not have specific anchors defining the appearance of contour, nipple-areola complex, scars, or symmetry. In the absence of a well-known aesthetic chest surgery scoring rubric that has been previously validated in this subpopulation of transgender patients, we elected to use a traditional Likert scale ranging from poor (1) to excellent (5). What defines an aesthetic result is inherently subjective and thus this interpretation may differ between individual raters. Our intrarater agreement was good, suggesting that each rater is consistent in their own interpretation; however, interrater agreement was fair, suggesting that raters’ conception of an aesthetic result may vary between individual surgeons.
Finally, nipple-areola complex sensation was self-reported rather than assessed by use of objective measures. We do not routinely use monofilament or two-point discrimination for nipple-areola complex sensation but acknowledge that these objective measures could be used to supplement patient self-reported sensation.
Because of the wide range of techniques, and varying definitions of complications, comparison between studies is difficult. Of the other series, only Monstrey at the Universitair Ziekenhuis Gent in Belgium used both techniques found in our study. When comparing the free nipple graft and the concentric circular groups (including their extended concentric circular group), our complication rates are similar, with the exception of increased wound dehiscence seen in our concentric circular group (Table 8). However, this may simply reflect a difference in the threshold to include delayed wound healing or a partial dehiscence as a complication. Our overall revision rate was 23.8 percent, which is comparable to the rate of 32.6 percent reported by Monstrey et al. (Table 9). Despite the large number of techniques available in the literature, our experience suggests exclusive use of the concentric circular and the free nipple graft techniques will accomplish the goals of subcutaneous mastectomy and male chest contouring for any patient.
This study represents a large single-surgeon series of female-to-male transgender patients undergoing subcutaneous mastectomy using two chest wall contouring techniques. In patients who meet selection criteria, the concentric circular technique is preferred at our institution because of fewer scars, improved aesthetic contour, and potential for retained nipple sensation. These patients must be counseled regarding the higher rate of complications and increased likelihood of additional surgery. Not all patients are candidates for the concentric circular technique. Smokers or those with a nipple-to–inframammary fold distance greater than 7 cm or a nipple-to–inframammary fold distance less than 7 cm and a body mass index greater than 27 kg/m2 should undergo the free nipple graft technique. Although there is no substitute for sound clinical judgment, we have attempted to provide objective criteria based on specific values of well-known and easily attainable preoperative measurements to provide a convenient adjunct to assist surgeons with preoperative planning.
1. Hage JJ, Karim RB. Ought GIDNOS get nought? Treatment options for nontranssexual gender dysphoria. Plast Reconstr Surg. 2000;105:12221227.
2. Coleman E, Bockting W, Botzer M, et al. Standards of care for the health of transsexual, transgender, and gender-nonconforming people, version 7. Int J Transgend. 2012;13:165232.
3. De Cuypere G, T’Sjoen G, Beerten R, et al. Sexual and physical health after sex reassignment surgery. Arch Sex Behav. 2005;34:679690.
4. Gijs L, Brewaeys A. Surgical treatment of gender dysphoria in adults and adolescents: Recent developments, effectiveness, and challenges. Annu Rev Sex Res. 2007;18:178224.
5. Klien C, Gorzalka B. Sexual functioning in transsexuals following hormone therapy and genital surgery: A review (CME). J Sex Med. 2009;6:29222939.
6. Pfafflin F, Junge A. Sex Reassignment: Thirty Years of International Follow-Up Studies SRS. A Comprehensive Review, 1961–1991. 1998.Dusseldorf: Symposium Publishing.
7. Lindsay WR. Creation of a male chest in female transsexuals. Ann Plast Surg. 1979;3:3946.
8. Eicher W. Transsexualismus. 1992.Stuttgart: Fisher Verlag.
9. Pryor JP, Gill H. Surgical considerations in 29 female-to-male transsexual patients. Gender Dysphoria 1991;1:44–46.
10. Hage JJ, Bloem JJ. Chest wall contouring for female-to-male transsexuals: Amsterdam experience. Ann Plast Surg. 1995;34:5966.
11. Hage JJ, van Kesteren PJ. Chest-wall contouring in female-to-male transsexuals: Basic considerations and review of the literature. Plast Reconstr Surg. 1995;96:386391.
12. Monstrey S, Selvaggi G, Ceulemans P, et al. Chest-wall contouring surgery in female-to-male transsexuals: A new algorithm. Plast Reconstr Surg. 2008;121:849859.
13. Colić MM, Colić MM. Circumareolar mastectomy in female-to-male transsexuals and large gynecomastias: A personal approach. Aesthet Plast Surg. 2000;24:450454.
14. Conway H. Mammaplasty; analysis of 110 consecutive cases with end-results. Plast Reconstr Surg (1946) 1952;10:303315.
15. Davidson BA. Concentric circle operation for massive gynecomastia to excise the redundant skin. Plast Reconstr Surg. 1979;63:350354.
16. Maillard GF. A Z-mammaplasty with minimal scarring. Plast Reconstr Surg. 1986;77:6676.
17. McKissock PK. Reduction mammaplasty with a vertical dermal flap. Plast Reconstr Surg. 1972;49:245252.
18. Pitanguy I. Transareolar incision for gynecomastia. Plast Reconstr Surg. 1966;38:414419.
19. Webster JP. Mastectomy for gynecomastia through a semicircular intra-areolar incision. Ann Surg. 1946;124:557575.
20. Takayanagi S, Nakagawa C. Chest wall contouring for female-to-male transsexuals. Aesthetic Plast Surg. 2006;30:206212; discussion 213.
21. Antoszewski B, Bratoś R, Sitek A, Fijałkowska M. Long-term results of breast reduction in female-to-male transsexuals. Pol Przegl Chir. 2012;84:144151.
22. Monstrey S. Personal communication, 2012.
23. Brink RR. Management of true ptosis of the breast. Plast Reconstr Surg. 1993;91:657662.
24. Regnault P. Breast ptosis: Definition and treatment. Clin Plast Surg. 1976;3:193203.
25. Breiman L, Friedman JH, Olshen RA, Stone CI. Classification and Regression Trees. 1984.Belmont, Calif: Wadsworth.
26. Shrout PE, Fleiss JL. Intraclass correlations: Uses in assessing rater reliability. Psychol Bull. 1979;86:420428.
27. Fleiss JL. Statistical Methods for Rates and Proportions. 1981.2nd ed. New York: Wiley.
28. Rosner B. Fundamentals of Biostatistics. 2005.Belmont, Calif: Duxbury Press.
29. King MW, Resick PA. Data mining in psychological treatment research: A primer on classification and regression trees. J Consult Clin Psychol. 2014;82:895905.