Adenocarcinoma of Bartholin gland is a very rare disease and its molecular pathological features are poorly delineated. A 92-year-old woman with mucinous adenocarcinoma of Bartholin gland with metastasis to the right inguinal lymph node was treated with radiation therapy alone. Despite intensive radiation therapy, the tumor was locally recurrent and the patient died 10 months after radiation therapy. We searched for the presence of human papillomavirus 16 and 18 DNA and the expression of p53 protein, CA19-9, CEA and MIB-1 antigen. Immunohistochemical study showed that mucinous adenocarcinoma of Bartholin gland produced CA19-9 and CEA.
Received August 15, 2000; accepted January 31, 2001.
Bartholin gland carcinomas are an uncommon disease accounting for 2–7% of all vulvar carcinomas and 0.1% of all female genital malignant neoplasms (1–3). Histologically, they are classified as squamous cell carcinoma, adenocarcinoma, adenoid cystic carcinoma and others. Adenocarcinoma reportedly composes 10–43% of all cases of Bartholin gland carcinoma (1,4,5). Because of the rare incidence, its pathogenesis, biological behavior and radiation response are not well known. We report a case of mucinous adenocarcinoma of Bartholin gland treated with radiation therapy alone. Additionally, we examined for the presence of human papillomavirus (HPV) 16 and 18 DNA and the expression of p53 protein, CA19-9, CEA and MIB-1 antigen immunohistochemically.
A 92-year-old woman was admitted to our hospital on April 4, 1996, with complaints of a mass in the right vulvar vestibule. Pelvic examination revealed a 5.0 × 4.5 cm irregular solid mass in the right vulva, which was ulcerating with mucinous exudate from its surface. The tumor had invaded the right vaginal wall but no definite tumor lesion was observed in the uterine cervix. A right inguinal lymph node of 4 × 3.5 cm in diameter was palpable, but no other obvious lymph node metastases were recognized by pelvic computed topography (CT). No obvious distant metastases were recognized by chest and abdominal CT and bone scan. Blood examination revealed a high CA19-9 level of 480 U/ml (normal: <37 U/ml). The serum CEA level was within the normal range. The patient underwent radiation therapy because of age and advanced disease.
External irradiation with 10 MV X-rays was administered to the pelvic area including the pelvic lymph node chain and primary lesion with 25 fractions of 2 Gy per day to a total dose of 50 Gy. Along with the external irradiation, intracavitary irradiation with a remote afterloading system using 60Co was performed at a total dose of 15 Gy in three fractions for vaginal primary tumor. Because the radiation response was worse than usual for vulvar squamous cell carcinoma, the patient additionally received electron beam treatment for the residual tumor at a total dose of 15 Gy in five fractions. Consequently, a total dose of 80 Gy was distributed to the primary tumor. After radiation therapy, the patient was discharged in July 1996 and was followed up on an out-patient basis. The tumors gradually decreased and clinically disappeared 2 months after radiation therapy. The serum CA19-9 level became normalized at that time. However, 4 months after radiation therapy, the serum CA19-9 level became elevated again and the tumor locally relapsed. The patient refused additional treatment other than pain control. She gradually became weak because of appetite loss and died on February 2, 1997, 10 months after the initiation of radiation therapy. No autopsy was performed.
Corresponding with the site of Bartholin gland in the right vulva, an irregular solid tumor with ulceration and mucinous exudate from the surface existed (Fig. 1). Biopsy was carried out before radiation therapy and the cut surface was defined, solid, grayish and moist. Microscopically, it was diagnosed as mucinous adenocarcinoma. The squamous mucosal surface was partially broken by the tumor that extended from the submucosal tissue (Fig. 2a). The tumor cells had abundant cytoplasm and prejudiced nuclei floated loosely or fence-like in a mucinous lake. Formation of cancer lumen was poor. Some of the tumor cells had crescent-shaped nuclei, i.e. signet ring-type tumor cells (Fig. 2b). Mucin was positive for periodic acid–Schiff (PAS) staining. Normal Bartholin gland tissue was not found anywhere.
Immunohistochemical study was performed using the streptavidin–biotin method and primary antibodies against p53 protein (Dako, Glostrup, Denmark), CA19-9 (Zymed, CA, USA), CEA (Nichirei, Tokyo, Japan) and MIB-1 antigen (Immunotech, Marseilles, France). The sections were heated in 0.01 M citrate buffer, pH 6.0, using a microwave processor at 95°C for 20 min for MIB-1 immunostaining and at 90°C for 10 min for p53 protein. However, for CEA and CA19-9, heating pretreatment was not applied. CA19-9 was expressed both in the cytoplasm and on the cell membrane of tumor cells and was distributed in some parts of the tumor cells that invaded the stroma (Fig. 3). The expression of CEA was present in the cytoplasm of almost all the tumor cells but little or no CEA was found in the stroma (Fig. 4). Signet ring-type tumor cells expressed both CA19-9 and CEA in the cytoplasm. The cancer cells with nuclei positive for p53 and MIB-1 were determined as positive cells. p53 protein was not detected in the normal mucosa and MIB-1 antigen was expressed in parabasal cells of the squamous mucosa. Ten random fields were chosen and the numbers of positive tumor cells in each field were summed and the percentage in the tumor cells was calculated as the labeling index. The labeling indices of p53 and MIB-1 in the tumor cells were 45 and 23%, respectively. Interestingly, signet ring-type tumor cells were positive for p53 staining and negative for MIB-1 (Figs 5 and 6).
Polymerase Chain Reaction (PCR) Analysis
A tumor specimen was examined for the presence of HPV DNA by PCR (Perkin-Elmer Cetus, Norwalk, CT, USA). A DNA sample was extracted from the formalin-fixed and paraffin-embedded specimen and two type-specific primers for HPV 16 and 18 DNA were used (6). PCR analysis did not detect HPV 16 and 18 DNA in the tumor.
Diagnostic criteria for Bartholin gland carcinoma were described by Chamlian and Taylor (7); the areas of transition from normal to neoplastic elements must be present, the tumors should involve the area of Bartholin gland, the tumor is histologically compatible with origin from Bartholin gland and there is no evidence of a primary tumor elsewhere. In our patient, three of the four criteria were met, but it was difficult to detect the areas of transition from normal to neoplastic elements by punch biopsy only before radiation therapy. Moreover, in the case of a large Bartholin gland carcinoma, the tumor could obliterate the residual normal gland. Copeland et al. (4) reported that the transition was observed in 16 (44%) of 36 patients mainly treated by surgical excision. This indicated that the transition is indeed the most reliable criterion, but this is not always the case, even when surgical resection is performed. We agree with Copeland et al. (4) that vulvar tumors occurring in the anatomical location of the Bartholin gland can be regarded as possible Bartholin gland neoplasms unless there is an area of transition indicating the origin to be the vulvar epithelium or vaginal or rectal mucosa.
Recently, the association between HPV and carcinogenesis of the genital tract has become well established (8). Felix et al. (5) identified HPV type 16 DNA in six of seven patients with Bartholin gland squamous cell carcinoma. In their report, however, HPV DNA was not detected in a patient with adenocarcinoma and in the present adenocarcinoma case, HPV 16 and 18 DNA were also not detected. The sensitivity of HPV DNA detection by PCR methods is known to be lower for formalin-fixed specimens than fresh specimens. Mucinous adenocarcinomas of the uterine cervix demonstrated a very high prevalence of HPV DNA, similar to that reported for cervical squamous cell carcinoma (9). Additional work with large numbers of adenocarcinomas is needed to confirm whether or not HPV infection is associated with carcinogenesis in Bartholin gland.
Expression of p53 protein indicates p53 gene activation or mutations, since mutant p53 protein is predominantly recognizable by immunohistochemistry because of its longer half-life (10). The present study showed strong p53 overexpression, suggesting that p53 mutation may contribute to carcinogenesis of Bartholin gland. Milde-Langosch et al. (11) reported that in cervical carcinoma HPV was persistent and p53 mutations absent, whereas in vulvar carcinoma, HPV was mostly absent or not persistent and the p53 mutation rate was very high. Flowers et al. (12) reported that p53 mutations with associated 17p13.1 LOH were more common in HPV-negative vulvar carcinoma. The correlation between the p53 protein expression and clinical outcome after radiation therapy is controversial. In uterine cervical cancer, Ebara et al. (13) showed no prognostic value of p53 expression, although Nakano et al. (14) reported its prognostic value. Further investigations should be performed to clarify the role of p53 protein in Bartholin gland carcinoma.
In general, mucinous adenocarcinoma is a rare histological subtype in the colon, pancreas, breast and ovary. As for Bartholin gland, the nature of mucinous adenocarcinoma is unknown. Mossler et al. (2) reported that three of five patients with adenocarcinoma were the mucinous type, that all three underwent surgery and that the extension of disease affected prognosis more than histological subtype. The stage of the present case was also advanced at the time of diagnosis. Enhanced expression of CA19-9 was observed in both serum and tissue in the present case. From the immunohistological findings, we considered that the mucinous adenocarcinoma cells excessively produced CA19-9 in the cytoplasm and cell membrane and secreted into the stroma, leading to the elevation of serum CA19-9. On the other hand, the expression of CEA was present in the cytoplasm of almost all tumor cells but was not elevated in the serum. In mucinous gastric carcinoma, the relationship between the serum level of CA19-9 and lymph node metastasis was recognized and enhanced expression of CA19-9 in cancer cells of invaded stroma was observed (15). We consider that the differences of overexpression in stroma between CA19-9 and CEA may contribute to the elevation in serum.
Bartholin gland carcinoma, and also vulvar squamous cell carcinoma, commonly metastasize to the inguinal–femoral and pelvic lymph nodes as the disease progresses (1,4). The literature recommends radical vulvectomy and inguinal–femoral node dissection as the first choice of treatment (1,2,4). Local recurrence is very common (4). Therefore, postoperative radiation therapy is performed as a routine strategy for patients with high risk of local and regional recurrence (i.e. large tumor, close margins and positive nodes) (4). Prognostic factors that have been suggested are tumor size, differentiation, extent of local invasion and lymph node involvement (1,2). Although the 5-year survival rate was only 18% in patients with multiple positive node metastases, when only one inguinal node had microscopic metastasis it was more than 71% (1). As far as we know, there is no documented information as to the radiation response of Bartholin gland adenocarcinoma. In our case, the tumor treated with radiation therapy alone showed radiation resistance in spite of the fact that a high dosage was given. Yamada et al. reported that interstitial brachytherapy was useful in the treatment of Bartholin gland carcinoma (16). We consider that it is difficult to cure mucinous adenocarcinoma of Bartholin gland only by conventional radiation therapy and further intensive treatment that locally concentrates the radiation dosage is needed in inoperable cases.
This study was supported by a grant for the Charged Particle Therapy Project (National Institute of Radiological Sciences, NIRS, Chiba, Japan). The authors gratefully thank Professor H. Niibe of the Department of Radiology and Radiation Oncology, Gunma University School of Medicine, for his helpful suggestions and significant scientific support, Dr J. Hirato and Mrs M. Yokota of the First Department of Pathology, Gunma University School of Medicine for their technical assistance and Dr R. Imai of NIRS for helpful assistance.
For reprints and all correspondence: Tatsuya Ohno, Department of Radiology, Takasaki National Hospital, 36 Takamatsu-Cho, Takasaki, Gunma 370-0829, Japan. E-mail: firstname.lastname@example.org
1Division of Radiation Medicine, National Institute of Radiological Sciences, NIRS, Chiba, 2Second Department of Pathology, Gunma University School of Medicine, Maebashi and 3Department of Pathology, Mito Saiseikai General Hospital, Mito, Ibaraki, Japan