Łukasz Kuźbicki
The frequency of COX-2 expression in breast cancer is high but heterogeneous. COX-2 detected in epithelial cancer cells is a poorly effective prognostic and predictive marker. The discrepancies in the results may have both biological and methodological causes.
INTRODUCTION: Induced Cyclooxygenase-2 (COX-2) can support cancer progression through various mechanisms and may be useful in predicting outcomes. COX-2 inhibitors are considered to have chemopreventive and therapeutic potential. This review aimed to verify the frequency of COX-2 immunohistochemical expression in breast cancer and its prognostic significance.
METHODS: The results for a total of 13,923 breast carcinomas from 60 publications were analyzed. A meta-analysis of COX-2 expression and overexpression frequency, an analysis of correlations with prognostic factors, and a cluster analysis of studies with similar results were conducted.
RESULTS: The median and mean values of the fraction of COX-2-positive cases across 34 studies, regardless of expression intensity, were 81% and 77±20%, respectively. COX-2 overexpression was demonstrated by 37 researchers, with median and mean frequencies of 49% and 49±15%, respectively. However, the results were highly heterogeneous (I2>90%, P < 0.0001). COX-2 was generally uncorrelated with clinical, pathological, and molecular prognostic factors and was associated with shorter patient survival in approximately half of the studies.
CONCLUSION: The frequency of COX-2 expression in breast cancer is high but heterogeneous. COX-2 detected in epithelial cancer cells is a poorly effective prognostic and predictive marker. The discrepancies in the results may have both biological and methodological causes.