A total of 97 cases of CCA treated between 1984 and 2007 at the National Defense Medical College Hospital, (Tokorozawa, Japan) were enrolled in the present study. Of the 97 CCAs, a consecutive series of 38 CCAs synchronous with endometriosis (EM-related CCAs) and 21 CCAs adjacent to CCAF component (CCAF-related CCAs) were identified, according to the histopathological criteria described previously (18). Of those, 31 non-atypical endometrioses, 38 atypical endometrioses, 20 benign CCAFs and 21 borderline CCAFs were identified. A total of 18 cases with solitary endometriosis that had no CCA were used as controls. All patients provided written informed consent for the present study.

Two core specimens, 1.5 mm in diameter, for each case were obtained from cancer tissue blocks and transferred to recipient blocks using a Tissue Microarrayer (Beecher Instrument, Silver Spring, MD, USA). All specimens were cut into 4-µm-thick slices to make tissue sections for IHC staining. The tissue sections were deparaffinized and boiled in an autoclave at 121°C for 15 min in 0.01 mol/l citrate buffer (pH 6.0) and were then allowed to cool at room temperature. Endogenous peroxidase activity was blocked using methanol added to 0.3% hydrogen peroxidase. The slides were incubated at 4°C overnight with mouse monoclonal primary antibody against BAF250a (cat. no. sc-32761; dilution, 1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA). Following incubation, the samples were reacted with dextran polymer reagent combined with secondary antibodies and peroxidase (cat. no. Z0420; 1:200; Dako A/S, Glostrup, Denmark) for 1 h at room temperature. Specific antigen-antibody reactions were visualized with 0.2% diaminobenzidine tetrahydrochrolide and hydrogen peroxidase, and counterstaining was performed using Mayer's hematoxylin. Non-neoplastic cells, including fibroblasts and lymphocytes, served as positive internal controls. As negative controls, tissue sections without the primary antibody were used.

For BAF250a detection, the presence of nuclear immunoreaction was taken into account for assignment of immuno-positivity. The lesions were considered to be positive for BAF250a if 50% or more of tumor cells in the area of interest showed equal to or more strong immunoreactive intensity compared with the positive controls (BAF250a-retained cases). If no detectable nuclear staining of tumor cells or <50% of tumor cells in the area of interest showed less immunoreactive intensity compared with the positive controls, they were defined as having a loss of BAF250a expression (BAF250a-deficient cases). The lesions were assessed independently by two observers (Masafumi Kato and Morikazu Miyamoto) in a blinded manner and any discrepancies between the two observers were resolved by conferring over a multi-viewer microscope.

Patient background, including age, concurrence of endometriosis, co-existence of CCAF, international federation of gynecology and obstetrics (FIGO) stage, residual tumor in primary surgery and chemotherapy regimen were compared, according to the BAF250a expression status. In addition, frequencies of loss of BAF250a expression in EM-related CCAs and CCAF-related CCAs were examined, as well as precursors (non-atypical endometrioses, atypical endometrioses, benign CCAFs and borderline CCAFs). Tumor response to adjuvant chemotherapy in evaluable cases, progression-free survival and overall survival were analyzed according to BAF250a expression status. Multivariate analyses for overall survival and progression-free survival were performed.

Statistical analyses were performed using Stat Mate IV software (ATMS, Tokyo, Japan) and Statview version 5 software (SAS Institute Japan, Ltd., Tokyo, Japan). Student's t-test and χ² test were used to compare patient characteristics of two groups. For survival analyses, Kaplan-Meier curves and the log rank test were used. Prognostic significance was analyzed using the Cox proportional hazard model using variables as follows: Age (continuous variable), concurrence of endometriosis (yes vs. no), FIGO stage (I/II vs. III/IV), residual tumor (≤1 cm vs. >1 cm) and BAF250a status (BAF250a-deficient vs. BAF250a-retained). P<0.05 was considered to indicate a statistically significant difference.

The characteristics of the patients were assessed according to the BAF250a expression status, and this is shown in Table I. BAF250a-deficient expression was observed in 30% (29/97) of all cases. No differences were observed in age, FIGO stage, residual tumor in primary surgery and chemotherapy regimen between the two groups. Concurrence of endometriosis was observed more frequently in BAF250a-deficient cases compared with in BAF250a-retained cases (P<0.05). By contrast, co-existence of CCAF was significantly more frequent in BAF250a-retained cases compared with that in BAF250a-deficient cases (P=0.04).

The frequencies of BAF250a-deficient expression were 19% (6/31) in non-atypical endometriosis, 26% (10/38) in atypical endometriosis, 5% (1/20) in benign CCAF, 5% (1/21) in borderline CCAF, 39% (15/38) in EM-related CCA and 10% (2/21) in CCAF-related CCA (Fig. 1). In solitary endometriosis, loss of BAF250a expression was detected in 6% (1/18) of cases. In comparison with the frequency of BAF250a-deficient expression between EM-related CCAs and CCAF-related CCAs, a significant difference was observed between the two groups (P=0.02).

The response rate of chemotherapy in evaluable cases is shown in Table II. No significant difference was observed in response rate of primary chemotherapy between the two groups. A total of 50% (13/26) in BAF250a-retained cases and 30% (3/10) in BAF250a-deficient cases (P=0.48) was observed.

Kaplan-Meier survival curves of all patients are shown in Fig. 2. BAF250a-deficient expression status was not significantly correlated with progression-free and overall survival of CCA in all enrolled cases. The 5-year progression-free survival rates were 68% in BAF250a-retained cases and 59% in BAF250a-deficient cases (P=0.24). Additionally, the 5-year overall survival rates were 67% in retained expression cases and 49% in deficient cases (P=0.43).

Kaplan-Meier survival curves of stage I/II patients (Fig. 3A and B) and stage III/IV patients (Fig. 3C and D), according to BAF250a expression, were also shown. Additionally, no significant differences were observed in the progression-free and overall survival in patients with stage I/II cases. The 5-year progression-free survival was 86% in BAF250a-retained cases and 91% in BAF250a-deficient cases (P=0.82). The 5-year overall survival rate was 85% in BAF250a-retained cases and 78% in BAF250a-deficient cases (P=0.38). In cases with stage III/IV disease, no differences were observed between the two groups. The 5-year progression-free survival was 23% in BAF250a-retained cases and 13% in BAF250a-deficient cases (P=0.23). The 5-year overall survival rate was 25% in BAF250a-retained cases and 16% in BAF250a-deficient cases (P=0.21).

In the multivariate analysis for progression-free survival, BAF250a expression status was not identified as an independent prognostic factor (P=0.47; Table III). Residual tumor diameter was identified as an independent factor for progression-free survival. In the multivariate analysis for overall survival, age (P=0.01), FIGO stage (P<0.01) and residual tumor diameter (P=0.02) were prognostic factors; however, the BAF250a expression status was not identified as an independent prognostic factor (P=0.56) in the present cases.