Tissue collection

This retrospective study included 34 patients with histologically proven squamous cell vaginal cancer treated between 1994 and 2008 at the Department of Gynecology and Obstetrics at the University Hospital Vienna. Melanoma, sarcoma, adenocarcinoma, and other rare histologic types of vaginal cancer were excluded. The pretreatment evaluation included patient history; complete physical examination; routine laboratory studies; chest X-ray; computed tomography of the abdomen and pelvis; and, depending on the condition of the patient, cystoscopy and recto/sigmoidoscopy.

Patients were staged according to the International Federation of Gynecology and Obstetrics (FIGO) guidelines. Treatment was individualized and was based on the extent and location of the tumor and the medical status of the patient by surgery and radiation/chemotherapy therapy. All patients were followed up in 3 month intervals with visual examination of the vagina, vaginorectal palpation, and abdominal ultrasound for 5 years. In cases of equivocal findings, computed tomography was performed.

Following standard clinical guidelines, recurrent disease was either diagnosed clinically, with a biopsy, or suspected radiologically. Documentation of death and causes of death were performed using autopsy test results. Tissue specimens were obtained during surgery or by biopsy forceps of the suspected lesion, which may have appeared as a mass, a plaque, or an ulcer. Sections of 3 μm were cut from formalin-fixed, paraffin-embedded tissue blocks. One section from each tumor was stained with hematoxylin and eosin for histological confirmation of tumor presence. An experienced pathologist, who was blinded to patient outcome, reviewed all slides.

Clinical information, including follow-up data, was obtained from the database of the Department of Gynecology and Obstetrics of the University Hospital Vienna. The Institutional Review Board at the Medical University of Vienna approved this retrospective data analysis.

Immunohistochemistry (IHC)

The original hematoxylin- and eosin-stained sections were reviewed by an experienced pathologist. Three micrometer thick sections, showing representative areas of the primary tumor, were cut from the tissue blocks of the formalin-fixed and paraffin-embedded biopsies and surgical specimens.

Sections were stained with mouse monoclonal antibodies to EGFR (clone 111.6; Neo Markers for Lab Vision Corp, Fremont, CA) and to VEGF Ab-3 (JH121, Neo Markers for Lab Vision Corp). They were deparaffinized in xylene and rehydrated through graded alcohol concentrations to distilled water. To block endogenous peroxidase activity, the slides were immersed in a Coplin jar with 3% hydrogen peroxide in methanol for 15 minutes. Antigen retrieval was performed in EDTA buffer (pH 8) by microwaving the sections for 15 minutes at a subboiling temperature, followed by 30 minutes of cooling.

The mouse monoclonal antibody to VEGF Ab-3 (JH121) was diluted with antibody diluent (Dako, Glostrup, Denmark) at 1:50 and incubated overnight at 4°C and briefly rinsed in PBS (pH 7.4). Biotinylated antibodies (AB2; Dako Chem-Mate Denmark) were used as the secondary antibody and incubated at room temperature for 25 minutes. After washing in PBS, the slides were incubated in strepavidin peroxidase-horseradish peroxidase (Dako Chem-Mate) for 25 minutes at room temperature. After short rinses in PBS, the staining was developed using liquid DAB-substrate (Dako Chem-Mate) for 5 minutes. Excess substrate was removed by a short rinse in distilled water. Finally, the slides were lightly counterstained with Gill’s haematoxylin and mounted with Kaisers glycerine gelatine (Merck, Darmstadt, Germany).

VEGF expression was defined as positive if unequivocal granular staining of the cytoplasm was present in at least 10% of tumor cells. The intensity of immunoreactivity was scored semiquantitatively as weak (+), moderate (++), and strong (+++).

For EGFR, pretreatment of the sections was identical as described in previous text, including the blocking of peroxidase activity. Tissues were then treated with 0.2% protease type XXIV (Sigma Chemical, St Louis, MO) at 37°C for 10 minutes, and digestive activity was stopped by 2 rinses of cold PBS (pH 7.4) for 5 minutes each. The primary antibody to EGFR was diluted with antibody diluent (Dako) 1:100 and incubated overnight at 4°C.

The remaining steps are identical to the method described in previous text. For EGFR the following scoring approach of membranous staining was used: score −, no staining or unspecific staining of tumor cells; score +, weak and incomplete staining of more than 10% of tumor cells; score ++, moderate and complete staining of more than 10% of tumor cells; score +++, strong and complete staining of more than 10% of tumor cells. Evaluation was done following EFGR pharm Dx Tm Dako interpretation guidelines.

Cytoplasmic staining was considered as nonspecific. Appropriate human placenta tissue was used as positive control for the antibodies against EGFR and angoiosarcoma tissue for the antibodies against VEGF. The negative control slide was prepared from the same tissue block as the specimen. Instead of using a primary antibody, we used a nonimmune mouse serum (DAKO code X0942).

Statistical analysis

After testing for normality using the Kolmogorov-Smirnov test, values are given as mean (SD) or median (interquartile range [IQR]), where appropriate. Correlation between clinicopathologic parameters (clinical stage, tumor size, age, and recurrent disease) and EGFR and VEGF were tested using the χ ² test. Results are given as P value and odds ratio (95% confidence interval [CI]).

Parameters have been calculated as follows: EGFR status (negative, weak positive vs moderate strong positive and intense positive), VEGF status (negative, weak positive vs moderate strong positive and intense positive), tumor status (FIGO I-II vs III-IV), grading (G1 vs G2-G3), and tumor size (≤3 cm vs >3 cm). Survival probabilities were calculated by the product limit method of Kaplan and Meier.

Differences between groups were tested using the log-rank test. The results were analyzed for the endpoint of disease-free and overall survival. Survival times of patients without any evidence of recurrent disease or still alive at the time of last follow-up were censored with the last follow-up date. Univariate and multivariate Cox regression models for disease-free and overall survival were performed. A 2-sided P < .05 was considered statistically significant. The SPSS system (SPSS 11.0; SPSS Inc, Chicago, IL) was used for the calculations.