Borderline Brenner Tumor: A Review of the Literature

Borderline Brenner tumors occur in a wide range of ages from 30 to 84 years, but are more common in older patients, with more than 80% of the patients older than 50 years (34 of 42 patients with age information).^1–4,6–9  The most frequent clinical presentation of borderline Brenner tumor is abdominal fullness or abdominal mass, followed by abdominal pain and postmenopausal bleeding.^1–4,6–9  Other abdominal/pelvic symptoms include nausea, vomiting, back pain, and bowel obstruction.³  Tumor-induced anorexia and weight loss can also be seen.³  Some patients have no clinical symptoms.³  The abdominal and pelvic symptoms could be explained by the increased pressure to the adjacent organs/tissues caused by the growth of the tumor. Borderline Brenner tumors are not functional, hence postmenopausal bleeding indicates that some of the borderline Brenner tumors may contain hormone-secreting elements or there could be coexistence of other diseases.

Most borderline Brenner tumors are unilateral with variable size from less than 5 cm to greater than 30 cm.^1–4,6–9  Most have a smooth outer surface and are uniloculated or multiloculated with mixed cystic and solid areas (Figure 1). The cystic areas often have a smooth inner lining with focal thickening and contain watery or slightly mucinous fluid. Some borderline tumors can show nodular areas as well as papillary excrescences in the cystic areas.^2,6  Pure solid borderline Brenner tumors are very rare, with only 1 reported in the literature.³ 

Borderline Brenner tumors have areas of atypia coexistent with benign components. The cytologic findings that qualify a Brenner tumor as proliferative or borderline are as follows: mucinous or squamous metaplasia in the transitional epithelium, presence of small papillary processes, complex glandular formation, nuclear atypia including hyperchromatic nuclei, coarse chromatin clumping, prominent nucleoli, and increased mitotic activity^1–4,6–9  (Figures 2 and 3). Focal areas resembling noninvasive papillary transitional carcinoma or papillary squamous carcinoma in situ have also been reported in borderline Brenner tumors.⁶  Some borderline Brenner tumors contain localized calcification, which can occur in the epithelial nests or in the stroma.³  Focal necrosis was reported in 5 cases.⁶  In one report, tumors with microinvasion of less than 3 mm were also classified as borderline or tumor of low malignant potential.⁷  Features of benign Brenner tumor are usually also found in the background.

Cancer antigen 125 (CA 125) and carcinoembryonic antigen are nonspecific biomarkers for ovarian tumors, which are usually used to monitor the treatment effect and recurrence. In the literature we reviewed, 4 patients were tested for these 2 markers and 2 of them had elevated CA 125; however, there is no correlation between serial levels of CA 125 and borderline Brenner tumors.^8,9 

Borderline Brenner tumors stain positively for Ras (rat sarcoma viral oncogene), epidermal growth factor receptor (EGFR), cytokeratin (CK) 5/6, CK7, P63, CA 125, epithelial membrane antigen, GATA-binding protein 3 (GATA3), and S100P, and are negative for p16, Rb (retinoblastoma protein), p53, CK20, and CDX2 (caudal type homeobox 2).^7–11  Cyclin D1 shows weak immunoreactivity.⁷  The combination of EGFR, Ras, cyclin D1, p16, Rb, and p53 can differentiate borderline/malignant Brenner tumors from transitional carcinoma of ovary: Brenner tumors are positive for EGFR, Ras, and cyclin D1, but negative or weakly positive for p16, Rb, and p53, whereas transitional carcinoma of ovary has the opposite immunoreactivity.⁷  CK20 can help to make the distinction between Brenner tumor and metastatic urothelial carcinoma when it is necessary.

The differential diagnosis of borderline or malignant Brenner tumors includes poorly differentiated serous or endometrioid adenocarcinoma, which sometimes can have epithelial cells closely resembling urothelial cells. Transitional carcinoma of ovary does not have an associated benign Brenner component. Appropriate extensive sampling of the tumor and immunohistochemical staining can help to make the distinction.

Surgical resection is the choice of treatment for borderline Brenner tumors. The staging system for borderline Brenner tumors is the same as for malignant ovarian tumors.^12,13  Most reported borderline Brenner tumors are stage I. Staging for ovarian malignancy generally includes total abdominal hysterectomy and bilateral salpingo-oophorectomy (TAH-BSO) with peritoneal washing, cytology of the diaphragm, omentectomy, and resection of grossly visible metastases.^12,13  In practice, clinicians do not always perform full staging for borderline tumors.¹⁴  In the reported 44 cases with treatment information we reviewed, TAH-BSO was performed in 29 patients (66%), unilateral salpingectomy in 12 patients (27%), and bilateral salpingectomy in 3 patients (7%).^1–3,6–9  Among these 44 patients, 23 patients have follow-up information. The average no-evidence-of-disease period was 4.3 years for TAH-BSO (14 patients), and 4.9 years for unilateral salpingectomy (8 patients).^2,3,6–9  There seems to be no apparent difference between the outcomes for these 2 procedures.

The prognosis of borderline Brenner tumor is good in general. In the cases we reviewed, only 1 lethal recurrence was reported. The patient had 2 recurrences and died from the last recurrence without histologic confirmation.⁸  The authors attributed her first recurrence to incomplete resection.⁸  There is also another report of a death of an unrelated leukemia.⁶ 

The histogenesis of Brenner tumor is unclear. The most widely accepted hypothesis is that the tumor develops from Walthard cell nests.¹⁵  Walthard nests are metaplastic transitional epithelium usually seen in the paratubal region. Seidman and Khedmati¹⁶  explored the coexistence of Walthard nest and Brenner tumors by sectioning the peritoneal surface. Although Walthard nests were found in 50% (25 of 51 patients) of patients with Brenner tumors, which was significantly higher than the 28% (76 of 272 patients) found in controls, the authors attributed the increase to more extensive sectioning and concluded that the association of Brenner tumors with Walthard nests was weak.¹⁶  Immunohistochemical staining has also been used to study the relationship between Walthard nests and Brenner tumors. Roma and Masand¹⁷  showed that Brenner tumors and Walthard nests share a similar immunohistochemical staining profile: GATA3 positive, but PAX8 (paired box 8) and PAX2 (pair box 2) negative. A study from other investigators demonstrated similar findings by using a different group of biomarkers: AKR1C3 (aldo-keto reductase family 1 member C3) and androgen receptor showed positivity, but calretinin showed negativity for these 2 entities.¹¹  In the latter study, the authors also showed the presence of cilia in Walthard nests and benign/borderline Brenner tumors, supporting the fallopian tube origin of Brenner tumors.¹¹  The above studies provide supportive evidence for a histogenic association between Brenner tumors and Walthard nests; however, none of the above biomarkers or a combination thereof qualifies as a unique signature for Walthard nests. Further study is still needed to clarify the histologic origin of this tumor.

Borderline Brenner tumor is thought to arise from benign Brenner tumor. Recent studies on EGFR-Ras-Raf signaling pathway and p16 gene in borderline and malignant Brenner tumors support this hypothesis.

Cuatrecasas et al⁷  did a comprehensive immunohistochemical and molecular genetic study on 13 Brenner tumors (5 benign, 7 borderline, and 1 malignant). The authors demonstrated moderate to strong immunoreactivity for EGFR and Ras in borderline and malignant Brenner tumors as opposed to benign Brenner tumors, suggesting that the increased expression of these proliferative markers may contribute to malignant transformation. The underlying mechanisms for the deregulation of EGFR-Ras-Raf signaling pathway were also explored in several studies,^9,11  including the abovementioned study,⁷  but the results are not conclusive.

Decrease or loss of p16 protein expression in borderline and malignant Brenner tumors has consistently been shown in multiple studies, indicating the loss of p16 may play a role in the progression from benign to borderline/malignant tumors in addition to the deregulated EGFR-Ras-Raf signaling pathway.^7,9,11  The mechanisms for the loss or decreased p16 expression are controversial. Cuatrecasas et al⁷  demonstrated loss of heterozygosity and hypermethylation of the p16 gene in the 1 malignant Brenner tumor included in their study by using DNA extracted from a fresh frozen specimen. One borderline Brenner tumor study by De Cecio et al⁹  failed to detect hypermethylation of the p16 gene. Kuhn et al¹¹  analyzed the presence of CDKN2A (cyclin-dependent kinase inhibitor 2A, encoding gene of p16) by using fluorescence in situ hybridization and reported homozygous deletion in the epithelial components in all 7 atypical proliferative (borderline) Brenner tumors but not in the benign tumors. The discrepancies among these studies could be due to the methods used. Brenner tumors have 2 distinct components and these 2 parts may harbor different molecular aberrations. Targeting a specific component may yield a different result than using the whole tumor tissue.

Borderline Brenner tumors are rare and usually have a good prognosis with appropriate treatment. Exclusion of malignant epithelial ovarian tumor mimicking Brenner tumor is crucial because of the different prognoses and management approaches. Recent studies on the pathogenesis of Brenner tumors have led to the identification of biomarkers for the distinction of borderline and malignant Brenner tumors. Pathogenesis of borderline Brenner tumors still remains uncertain and controversial.