Objectives: To validate the results of a previous study with the tissue microarray technology showing that cyclooxygenase 2 (COX-2) overexpression and absent caspase 3 expression are associated with poor disease-specific survival in univariate analysis.

Methods: The study group comprised 80 consecutive patients with vulva cancer treated in the period from 1999 to 2003 in a university hospital. A tissue microarray with 3 tumor tissue cores per patient was constructed and stained with antibodies against COX-2, caspase 3, epidermal growth factor receptor, p16^INK4, cyclin D1, and Ki-67. The impact of the expression of these protein markers and selected clinicopathologic variables on disease-specific as well as disease-free survival was measured. Cox proportional hazard model was used for both univariate and multivariate analyses.

Results: In multivariate analysis, lymph node metastases and strong COX-2 expression were related to disease-free (hazard ratio [HR], 8.33, 95% confidence interval [CI], 2.97-23.36; P < 0.001; and HR, 6.42; 95% CI, 2.33-17.72; P < 0.001) and disease-specific survival (HR, 6.04; 95% CI, 2.12-17.19; P = 0.001; and HR, 5.11; 95% CI, 1.82-14.36; P = 0.002). In the present series, no association was found between caspase 3 expression and survival.

Conclusion: The prognostic significance of COX-2 overexpression was confirmed. In contrast, in the present series, no relation was found between caspase 3 expression and survival.