Uterine bleeding in breast cancer patients: Insights from a 5-year institutional review

Diagnostic findings in uterine bleeding cases

Malignancies and premalignant lesions were detected in the majority of cases, accounting for 37.7% cases (n = 11). These included metastatic breast carcinoma (n = 2/29, 6.9%), Nonkeratinising squamous cell carcinoma of cervix (2/29, 6.9%), keratinising squamous cell carcinoma of cervix (1/29, 3.4%), adenocarcinoma of cervix (2/29, 6.9%), cervical intraepithelial neoplasia grade 3 (1/29, 3.4%), endometrial malignancies including endometrioid carcinoma (1/29, 3.4%), high grade serous carcinoma (1/29, 3.4%) and carcinosarcoma (1/29, 3.4%). Benign endometrial polyps were the second most frequent finding, accounting for 31% cases (n = 9/29). Endometrial hyperplasia was identified in 10.3% uterine bleeding cases (n = 3/29). Other causes of UB included adenomyosis (1/29, 3.4%), leiomyoma (1/29, 3.4%), and endocervical polyp (2/29, 6.9%). There were two patients who did not have any specific pathology on the endometrial biopsy.

PAML-COEIN classification

The PALM-COEIN classification system, endorsed by the International Federation of Gynaecology and Obstetrics, categorises the underlying causes of abnormal uterine bleeding into structural factors (PALM: polyp, adenomyosis, leiomyoma, malignancy, and hyperplasia) and non-structural factors (COEIN: coagulopathy, ovulatory dysfunction, endometrial, iatrogenic, and not yet classified).^[3] We utilised the system to classify the cases of uterine bleeding (including PMB) in our cohort. According to PAML-COEIN Classification, AUB predominantly occurred due to structural causes (27/29), which included polyps (P = 11/29), adenomyosis (A = 1/29), leiomyoma (L-1/29), and miscellaneous (malignancy and hyperplasia = 14/29). This has been represented in Figure 1b.

Radiology findings

Patients receiving tamoxifen therapy exhibited increased endometrial thickness (>8mm) and showed characteristic cystic changes suggestive of cystic endometrial hyperplasia (2/27) on ultrasonography. Endometrial polyps (2/27) were another common finding, appearing as focal echogenic or enhancing lesions, while the myometrium may reveal adenomyosis (1/27) with a bulky uterus and widened junctional zone, or leiomyomas (1/27) as well-defined hypointense masses with heterogeneous enhancement. Less commonly, imaging detected atypical hyperplasia or endometrial carcinoma (1/27), seen as irregular, markedly thickened endometrium with disruption of the endo-myometrial interface and restricted diffusion on magnetic resonance imaging, sometimes with shallow or deep myometrial invasion. Also, cervical changes, such as enhancing T2 intermediate lesions suggestive of cervicitis or polyps (9/27), were also seen in a subset of patients. No imaging was available in the two metastatic cases. However, the patient with metastatic lobular carcinoma exhibited no tracer uptake on a positron emission tomography scan.

Histopathology findings

The primary breast malignancy encompassed luminal A (7/28,25%), luminal B (13/28, 46.4%), HER2-enriched (4/28, 14.3%), and basal (4/28, 14.3%) molecular subtypes. Notably, one patient with hormone-positive in-situ breast carcinoma, undergoing adjuvant Tamoxifen therapy, presented with uterine bleeding.

Uterine bleeding in breast cancer patients was predominantly linked to primary gynaecological malignancies, with metastatic disease being an infrequent contributor. There were only two patients who presented with metastasis to the endometrium and cervix. One of these patients with lobular breast carcinoma exhibited dyscohesive signet ring cells infiltrating the cervical stroma [Figures 2a-b]. The tumour cells were immunopositive for Cytokeratin as shown in Figure 2c, GATA3 as shown in Figure 2d, TRPS1, oestrogen receptor (ER), and progesterone receptor (PR).

Close

Figure 2: (a-d) Histopathological findings (a) Lobular carcinoma cells admixed with inflammatory infiltrate infiltrating the cervical stroma (H&E, 100x), (b) Signet ring cells were noted on higher magnification (H&E, 400x); Immunohistochemical findings (c) Strong and diffuse positivity for cytokeratin in tumour cells (IHC, 200x), (d) Diffuse nuclear staining for GATA3 in tumour cells (IHC, 200x). H&E: Haematoxylin and eosin stain, IHC: Immunohistochemistry, GATA3: GATA-binding protein 3.

Export to PPT

Another patient with ductal breast carcinoma on Tamoxifen. presenting with AUB, exhibited atypical, glandular proliferation on endometrial biopsy [Figure 2e]. The differential, considering the clinical setting, was Endometrioid carcinoma. The tumour cells were diffusely positive for ER as shown in Figure 2f and PR but were consistently negative for vimentin [Figure 2g]. The vimentin negativity prompted us to explore the possibility of a metastatic disease. On subsequent immunohistochemistry, the tumour cells exhibited diffuse GATA3 expression as shown in Figure 2h, confirming metastasis. Thorough histopathological workup with ancillary testing was employed in diagnosing other malignant lesions. The results have been summarised in Table 1. Synchronous dual primary malignancies and metastases were more frequent in elderly patients.

Close

Figure 2: (e-h) Endometrial curettage (e) The tumour cells exhibited marked nuclear pleomorphism and brisk mitotic activity (H&E, 400x), (f) Strong and diffuse nuclear expression for oestrogen receptor (IHC, 200x), (g) Tumour cells were completely negative for vimentin (IHC, 200x), (h) Diffuse nuclear staining for GATA3 in tumour cells (IHC, 200x). H&E: Haematoxylin and eosin stain, IHC: Immunohistochemistry, GATA3: GATA-binding protein 3.

Export to PPT

Endometrial polyp was the predominant histological finding in patients with tamoxifen exposure. These polyps exhibited staghorn-shaped glands with periglandular stromal condensation. Mucinous metaplasia was also noted. Patients with tamoxifen exposure also exhibited endometrial hyperplasia with an increase in the gland: stroma ratio (>3:1). Endometrioid carcinoma and Carcinosarcoma were also detected in two patients, developing metachronously subsequent to Tamoxifen exposure.

Treatment

In the 29 patients with AUB, 26 (89%) had non-metastatic breast cancers, with the majority (65%) locally advanced, while 3 patients had metastatic disease. Neoadjuvant chemotherapy was given to all cases with locally advanced breast cancer, and anti-HER2 therapy was added for HER2-enriched tumours. Out of the 26 cases, breast conservation surgery was done for 11 (42%) patients, while 58% underwent mastectomy. Among the 26 cases that had hormone receptor-positive breast cancer, adjuvant hormone therapy with Tamoxifen was given to 20 patients, while Letrozole was given to 6 patients. There were 3 patients diagnosed with metastatic breast cancer, all of whom were hormone receptor positive and HER2 negative. Palliative chemotherapy and endocrine therapy +/- CDK 4/6 inhibitor were administered. For the 5 cases detected to have a second primary in the cervix, radical chemoradiotherapy was given for 4 patients and radical radiotherapy alone in 1 patient. There was 1 patient who had metastatic recurrence on follow-up and received palliative chemotherapy. Second malignancy, endometrial carcinoma, was diagnosed in 3 cases, with 1 patient planned for observation after staging surgery, 1 patient receiving neoadjuvant chemotherapy, while another had metastatic disease for which palliative chemotherapy was administered. Out of the 8 patients who were detected with a second cervical or uterine malignancy, 3 had advanced or metastatic disease at diagnosis.

Tamoxifen-induced uterine bleeding: The predominant mechanism

Tamoxifen is a selective oestrogen receptor modulator utilised in the adjuvant management of ER-positive breast carcinoma. The administration of tamoxifen at a dosage of 20 mg daily for five years in the adjuvant setting is associated with an approximate 40% reduction in breast cancer recurrence during the first decade of follow-up.^[6] Routine ultrasonographic surveillance of asymptomatic breast cancer patients on tamoxifen therapy has not been found to be a clinically beneficial approach, as it demonstrates suboptimal specificity and positive predictive value.^[7] While its primary therapeutic benefit comes from its antiestrogenic effects, tamoxifen also exhibits mild estrogenic activity. At standard doses, in the endometrium specifically, tamoxifen acts as an oestrogen receptor (ER) agonist, promoting cell proliferation and increasing the risk of various uterine abnormalities, such as typical and atypical glandular hyperplasia, endometrial polyps, uterine fibroids, adenocarcinoma, and sarcoma.^[4,5] Endometrial polyp is the most common endometrial pathology associated with Tamoxifen exposure.^[8] Tamoxifen-induced polyps display larger size, increased proliferative activity, stromal condensation, and aberrant glandular differentiation, which can disrupt vascular architecture and predispose to bleeding. Tamoxifen’s estrogenic effect can also stimulate the growth of leiomyomas, particularly in postmenopausal women, by binding to oestrogen receptors in myometrial and stromal cells. The combination of increased cellular proliferation, reduced apoptosis, abnormal stroma, and angiogenesis makes the endometrium more susceptible to abnormal, heavy, or prolonged bleeding episodes in tamoxifen-treated patients.^[9]

Endometrial carcinoma arising in Tamoxifen users is associated with an aggressive morphological subtype, leading to poorer outcomes.^[10,11] Long-term Tamoxifen usage has also been associated with the development of carcinosarcoma, particularly in postmenopausal women.^[12] In the present study, we had two cases, one of endometrial carcinoma and the other of carcinosarcoma, both developing metachronously with tamoxifen therapy. Even though the number of cases is not significant, regular monitoring for abnormal uterine bleeding and prompt investigation with endometrial assessment is recommended in this population.

Metastatic disease

Uterine bleeding may rarely occur due to uterine metastases from retrograde lymphatic dissemination of ovarian metastases or hematogenous spread from a primary breast neoplasm.^[13] Metastatic breast carcinoma rarely involves the uterus, with myometrial involvement being more common than endometrial.^[14] It is essential to differentiate primary gynaecological malignancies and potential metastasis from breast carcinoma, particularly given the minority of patients demonstrating metastatic spread. A literature review by Huo et al. identified 13 cases of uterine metastases from breast cancer, noting that ER/PR-positive tumours, especially invasive lobular carcinoma, are more prone to endometrial spread than invasive ductal carcinoma.^[15,16] Ductal metastases are exceptionally uncommon, though Cift et al. reported a case of myometrial metastasis from ductal carcinoma in a patient on Tamoxifen therapy.^[17] In rare instances, patients may present with metastatic breast carcinoma affecting tamoxifen-induced endometrial polyps, necessitating meticulous examination of these lesions.^[18] Furthermore, a few cases of breast metastases to the cervix have been documented; Razia et al. reported a rare instance of invasive lobular carcinoma simultaneously metastasising to an endometrial polyp, cervix, and leiomyoma.^[19-21] In our study, there were only two patients who presented with metastasis to the endometrium and cervix. These findings strongly suggest that uterine bleeding in breast carcinoma is not only a primary complaint but also a potential harbinger for secondary metastatic involvement in the gynaecological tract.

Uterine surveillance

Women with a history of breast cancer may have an increased risk of endometrial cancer due to inherited genetic conditions like Lynch syndrome or, more rarely, Cowden syndrome.^[22] The association of BRCA mutations with endometrial cancer is controversial, few studies suggest that BRCA1-positive women have a higher likelihood of developing serous or serous-like endometrial carcinoma.^[23] Therefore, it is advisable to conduct thorough surveillance to identify any concurrent malignancies and assess endometrial health prior to initiating tamoxifen therapy.^[24] Current evidence points to the existence of distinct high- and low-risk categories for atypical hyperplasia development during tamoxifen treatment in postmenopausal women, differentiated by the presence or absence of benign endometrial polyps before therapy commencement. Accordingly, pretreatment screening, potentially involving transvaginal ultrasonography and sonohysterography as necessary, may be beneficial for postmenopausal women prior to starting tamoxifen treatment.^[25]

Cervical carcinoma screening

The high prevalence of cervical carcinoma underscores the potential efficacy of routine PAP smear screening. For women diagnosed with breast cancer, routine cervical PAP smears are crucial, as their risk of developing cervical cancer is independent of their breast cancer status. Furthermore, treatments such as chemotherapy or immunosuppressive therapies can compromise the immune system, heightening susceptibility to infections like HPV, the principal etiological agent of cervical cancer. Consequently, consistent screening is particularly vital for this population.^[26] The presence of carcinosarcoma, high-grade serous carcinoma, and endometrioid carcinoma in this cohort highlights the importance of thorough pathological examination. These aggressive malignancies, though rare, have significant prognostic implications. In this study, we observed that despite patients developing secondary malignancies being on regular follow-up, none had received periodic ultrasonographic evaluations or PAP smear examinations.

Uterine surveillance by radiology

Uterine surveillance by radiology in breast cancer patients on tamoxifen therapy has several limitations because tamoxifen induces multiple, often confounding, changes in the uterus. This affects the accuracy of ultrasound and imaging interpretation, particularly regarding endometrial thickness and pathology. Benign tamoxifen changes (e.g., thickened endometrium, cystic changes) mimic malignant patterns, reducing the specificity of ultrasound. Many patients with thickened, irregular, or cystic endometrium do not have cancer but will still be referred for biopsy or hysteroscopy. There is no universally accepted endometrial thickness cutoff in tamoxifen users that reliably distinguishes benign from malignant findings. Some prefer an 8 mm cut-off for asymptomatic women, but findings above or below this can be both benign or malignant. Hence, routine annual imaging is not universally recommended for all asymptomatic women on tamoxifen due to limited specificity. Symptomatic women (with abnormal bleeding or pain) should always be investigated further, regardless of endometrial thickness. In symptomatic women, hysteroscopy with directed biopsy is the gold standard for diagnosis when imaging is suspicious. MRI may be needed for complex cases, especially if adenomyosis or deeply invasive disease is suspected.^[9,27]

However, despite these acknowledged limitations, there is a compelling case for developing new evidence-based guidelines that support routine ultrasonographic surveillance in tamoxifen users. The current approach of symptom-based evaluation alone may miss critical opportunities for early detection in a well-defined high-risk population. Modern technological advances, including high-resolution transvaginal ultrasound, 3D imaging, Doppler studies, and emerging AI-assisted interpretation, have significantly improved diagnostic accuracy compared to earlier studies that formed the basis of current recommendations against routine screening.^[28,29] A risk-stratified surveillance approach combining patient age, therapy duration, and standardised ultrasound findings could address the specificity concerns while optimising early detection benefits. Furthermore, the cost-effectiveness of preventing advanced disease through early detection, combined with patient preference for proactive monitoring and anxiety reduction, supports a paradigm shift toward systematic surveillance. Rather than abandoning ultrasound due to its limitations, we propose that standardised protocols with clear follow-up algorithms, continuous quality improvement measures, and integration of emerging technologies represent a more effective approach to reducing endometrial cancer mortality in this vulnerable population while serving as a platform for ongoing research and protocol optimisation.

Treatment

Tamoxifen-associated endometrial polyps are treated by hysteroscopic polypectomy for diagnosis and definitive treatment.^[30] Endometrial hyperplasia is treated with progestins.^[31] For atypical endometrial hyperplasia/endometrial intraepithelial neoplasia, the definitive treatment is hysterectomy for patients who have completed childbearing; selected fertility-sparing cases require continuous progestin with close surveillance.^[32] Endometrial carcinoma treatment depends upon stage and histological type.^[33] Cervical pathology should follow guideline-concordant care, stage-based surgery, or chemoradiation.^[34] Rare metastatic breast carcinoma to the gynaecologic tract requires systemic therapy; local measures (curettage, palliative radiotherapy) can palliate bleeding.^[35] In postmenopausal bleeding, tissue diagnosis should precede conservative management, given the risk of malignancy. Supportive care includes correction of anaemia (iron supplementation) and acute haemostasis with high-dose progestins. Short-course GnRH analogues can help in refractory bleeding.^[36]