Histologic Findings in Gynecologic Tissue From Transmasculine Individuals Undergoing Gender-Affirming Surgery

This is an institutional review board–approved retrospective study that included transmasculine individuals undergoing hysterectomy, salpingo-oophorectomy, and/or vaginectomy as gender-affirming surgery from January 2015 to June 2020 at New York University Langone Medical Center. Clinical data were retrieved from electronic health records, including age, body mass index (BMI), comorbidities, hormonal therapy, and pelvic ultrasound characteristics. All available surgical pathology and cytology slides were reviewed by 2 different pathologists. In addition, cervical cytology reports and human papillomavirus (HPV) testing were retrieved from medical charts.

Endometrium was classified as inactive, proliferative, or secretory.⁹  Inactive endometrium was characterized by presence of round to tubular glands with a single layer of cuboidal to columnar cells without mitotic activity. Proliferative endometrium was defined by round to tubular glands composed of pseudostratified columnar cells with presence of mitotic figures. Secretory endometrium was characterized by convoluted irregularly shaped glands with a single layer of columnar cells. Presence of nuclear vacuoles, intraluminal secretions, and stromal edema are features typically seen in the secretory phase. Follicular cysts were defined when the ovarian cysts were larger than 3 cm, while cystic follicles were classified when the largest diameter was between 0.5 cm and 3 cm. Ovarian follicle maturation was characterized as follows: primordial follicle (oocyte with a single layer of flattened granulosa cells); primary follicle (oocyte with a single layer of cuboidal granulosa cells); secondary follicle (oocyte with 2 layers of cuboidal and intact second layer of cubical granulosa cells); and antral follicle (oocyte with more than 2 layers of cuboidal granulosa cells and presence of an antrum).¹⁰  Cervical cytology specimens were evaluated according to the Bethesda classification.¹¹ 

Fifty-five patients were identified, and the clinical characteristics of the studied population are summarized in Table 1. The median age was 27 years (range, 18–56) and median BMI was 25.4 kg/m² (range, 18.9–43.4). Most patients were white (29 of 55), were receiving testosterone for at least 1 year (50 of 53), and had a previous gender-affirming chest surgery (52 of 55). The most common comorbidities were depression and anxiety (17 of 55), followed by asthma (7 of 55) and other psychiatric disorders (5 of 55). Only 1 patient had a history of cancer, which was non–hormonal-dependent. The histologic samples consisted of 40 uteri with bilateral fallopian tubes, 35 bilateral ovaries, and 15 vaginectomy specimens. The overall histologic and cytologic findings are summarized in Table 2.

Seventy-five percent of endometria were inactive (30 of 40) with significant stromal fibrosis in 80% (32 of 40) (Figure 1, A). Some patients showed evidence of cycling endometrium with proliferative (7 of 40) and secretory (3 of 40) patterns. All endometria, regardless of cycling status, had decreased thickness (Figure 1, B). There were no cases of endometrial hyperplasia or carcinoma. Myometrial atrophy (Figure 1, C), defined as prominent decrease of smooth muscle cytoplasm resulting in nuclear crowding without atypia or increased mitoses, was identified in 5% (2 of 40) of the patients. Common uterine pathology found in cisgender women was also identified in this cohort, such as leiomyomata (5 of 40) and endometrial polyps (3 of 40).

Cytohistologic findings in the cervix are shown in Figure 2, A through F. The most common finding in the cervix (17 of 40) and vagina (15 of 15) was transitional cell metaplasia, which showed negative p16 immunohistochemistry and low Ki-67 proliferation index (Figure 2, A through C). Prostatic-type glands in the cervix and vaginal tissue were identified in 20% (8 of 40) and 67% (10 of 15) of patients, respectively. Prostatic-type glands were characterized by glandular formations composed of luminal and basal cells, similar to prostatic tissue, arising in the lower portion of squamous epithelium (Figure 3, A through D; Figure 4, A and D). Immunohistochemistry was performed in 4 cases to confirm the immunoprofile of prostatic-type glands. The luminal cells in prostatic-type glands were positive for NKX3.1 and negative for PAX-8, which is the opposite immunoprofile of endocervical glands (Figure 4, B, C, and E). The basal cells were positive for P63 (Figure 4, F). Interestingly, most cases of prostatic-type glands arose in a background of transitional cell metaplasia, and in the cervix, they were in close proximity to the squamous-columnar junction.

Twenty-four cervical cytology reports were identified with the following results: 2 unsatisfactory (8%), 2 atypical squamous cells of undetermined significance (ASC-US) (8%), and 20 negative for intraepithelial lesion or malignancy (NILM) (84%). Seven cases had slides available for review with concordance with the original cytopathology diagnosis. All reviewed cases showed atrophic pattern with predominance of sheets of basal and parabasal cells associated with nuclear enlargement and mild hyperchromasia (Figure 2, D). One case showed presence of glycogenated cells with perinuclear halo and yellow pigment (Figure 2, F). A total of 8 HPV tests were available: 7 had negative findings (2 ASC-US and 5 NILM on cytology) and 1 was positive for high-risk HPV (NILM on cytology).

The most common findings in the ovaries were the presence of multiple bilateral cystic follicles (20 of 35) (Figure 5, A), stromal hyperplasia (5 of 35), and corpora lutea (5 of 35). In 57% (20 of 35) of cases, there was follicular maturation with histologic evidence of secondary and antral follicles (Figure 5, B), while 37% (13 of 35) of cases showed only primordial/primary follicles. Two cases (6%) had significantly decreased oocyte reserve, in which follicles were rarely identified; those 2 patients were 43 and 56 years old. Four cases of paratubal epididymis-like mesonephric remnant hypertrophy were identified (Figure 5, C), which represents 10% (4 of 40) of all cases and 50% (4 of 8) of cases in which mesonephric remnants were identified (Figure 5, D).

In this study, we compiled microscopic findings in uteri, fallopian tubes, ovaries, vaginal tissue, and cervical cytology results from 55 transmasculine individuals undergoing gender-affirming surgery. Most of the histologic changes in our series are not commonly found in cisgender women in their reproductive years and are mostly believed to be related to exposure to testosterone. The recognition and acknowledgment of these histologic variations can assist the pathology community in developing a standardized approach to these tissue samples, ensuring appropriate patient care and management.

The effect of testosterone in inducing endometrial atrophy has been reported in various previous studies^12–15  and was confirmed in the present work with the predominance of inactive endometria in 75% (30 of 40) of the patients. We did not identify any cases of endometrial hyperplasia or carcinoma, similarly to previous reports,^8,12,14,16  which is a reassuring finding and consistent with only rarely reported cases in the literature of endometrial hyperplasia and carcinoma in transmasculine individuals receiving testosterone.^13,17  The study by Grimstad et al⁸  found that most patients receiving testosterone had proliferative endometrium documented on pathology reports, while Loverro et al¹⁶  showed predominantly active endometrium with low Ki-67 proliferative index. In the present study, 25% (10 of 40) of patients showed cycling activity with proliferative and secretory endometrium and presence of corpora lutea. These findings are concordant with the reports in the literature indicating long-term testosterone users can still experience ovulatory events¹⁸  and unplanned pregnancies can occur in transmasculine individuals receiving testosterone; therefore, testosterone should not be used as a contraceptive method.¹⁹ 

Androgens have been shown to be important regulators of folliculogenesis and also play a role in polycystic ovary syndrome.²⁰  Several reports, including the present one, identified that a high percentage of transmasculine patients receiving testosterone can develop bilateral multiple cystic follicles in the ovaries, which has a striking similarity to polycystic ovary syndrome.^13,14,16  However, this finding was uncommon in some studies.^7,21  The presence of maturing follicles and corpora lutea confirms previous findings¹⁰  and reinforces the importance of fertility-preservation counseling in transmasculine individuals undergoing gender-affirming surgery. The studies by De Roo et al¹⁰  and Lierman et al²²  showed that oocytes collected from transmasculine patients had normal spindle structure and preserved oocyte maturation after undergoing assisted reproduction techniques.

Similar to the study by Khalifa et al,¹⁴  we found a high percentage of patients with transitional cell metaplasia in the lower genital tract. The importance of recognizing this finding lies in its morphologic overlap with and possible misinterpretation as squamous high-grade intraepithelial lesion (HSIL), due to the increased nuclear crowding and absence of typical maturation of squamous epithelium.^14,15  The use of immunohistochemical stains with p16 and Ki-67 can be helpful in the differential diagnosis if morphology is equivocal. These histologic findings correlate with the high percentage of ASC-US and unsatisfactory cervical cytologies in the present study. Previous reports showed higher percentage of unsatisfactory and abnormal cervical cytologies among transmasculine individuals receiving testosterone than among cisgender women.^23–25  Similar to histopathologic evaluation, cytologic changes in atrophic epithelium can mimic HSIL (Figure 2, E) and can represent a diagnostic pitfall without the proper history of testosterone exposure. Presence of glycogenated cells has been reported in cytology specimens with androgenic atrophy; this finding was rarely seen in this study, but could represent a potential mimicker of low-grade squamous intraepithelial neoplasia.²⁶ 

In this study, we also identified findings in transmasculine individuals that recapitulate histologic structures seen in cisgender men, such as prostatic-type glands in squamous epithelium of the lower genital tract and paratubal epididymis-like mesonephric remnant hypertrophy, which seems to be related to prolonged testosterone exposure. These phenomena were first described by Singh et al²⁷  in a series of 3 cases identified in transmasculine individuals. Epididymis-like mesonephric remnant hypertrophy seems to be exceedingly rare, as to our knowledge only 2 cases were published in the English literature, one in a transmasculine individual in the study by Singh et al²⁷  and the other in a postmenopausal cisgender woman.²⁸  Anderson et al²⁹  published a series involving patients with endogenous and exogenous androgen exposure, which included 7 transmasculine individuals presenting with prostatic-type glands arising in squamous epithelium of the lower genital tract. Our series is the first to study the frequency of prostate-type glands in a cohort of transmasculine individuals, which were found in 20% (8 of 40) of cervical and 67% (10 of 15) of vaginal tissue when systematic examination was performed. Owing to the benign appearance of these findings, they might be commonly overlooked during routine histologic examination. The presence of prostatic-type glands in the cervical stroma was previously described in cisgender women without history of androgen therapy, which seems to be exceedingly rare³⁰ ; owing to their distribution in the lamina propria and absence of epithelial involvement, the origin of these glands appears to differ from the prostatic-type glands in individuals receiving testosterone.

There is a wide range of variation in the frequency and type of histologic findings in the published literature of transmasculine individuals. This is possibly due to each cohort's patient characteristics, such as age, BMI and length, type, and dose of androgenic therapy, but can also be due to the different criteria used for the identification of specific histologic features. In addition, some studies did not include slide review and their results were based only on the initial pathology report, which, depending on local practices, can have large variations in the diagnosis of certain findings (eg, weakly proliferative versus inactive endometrium) and in the reporting of benign findings on patients' reports, which might not always be included (eg, prostatic-type glands, transitional cell metaplasia, hypertrophic mesonephric remnants).

One of the limitations of this study is a relatively small sample size. In addition, the absence of mechanistic and experimental studies does not allow definitive conclusions about the effect of testosterone in specific tissues. However, this is a comprehensive and systematic microscopic review of multiple tissue sites by subspecialized gynecologic and genitourinary pathologists, which provides a panorama of the findings in gynecologic tissue from transmasculine individuals.

Specific histologic features can be identified in tissue from transmasculine individuals undergoing gender-affirming gynecologic surgery. The recognition of these findings increases understanding of the gender-transitioning process and may contribute to optimal patient care.