Purpose: To evaluate the prognostic significance of cyclooxygenase-2 (COX-2) in breast cancer patients treated with conservative surgery and radiation therapy (CS+RT).
Methods and materials: Between 1975 and 2003, we retrieved specimens from 504 breast cancer patients treated with CS+RT. The specimens were constructed into tissue microarrays processed and stained for estrogen receptor (ER), progesterone receptor, Her2/neu, and COX-2. Each core was scored as positive or negative. All data including demographics, clinical, pathologic, staging, and outcome variables were entered into a computerized database.
Results: Expression of COX-2 was positive in 58% of cases and correlated with younger age (p = 0.01) and larger tumor size (p = 0.001). Expression of COX-2 was predictive of local relapse (relative risk[RR], 3.248; 95% confidence interval [CI], 1.340-7.871; p = 0.0091), distant metastasis (RR, 2.21; 95% CI, 1.259-3.896; p = 0.0058), and decreased survival (RR, 2.321; 95% CI, 1.324-4.071; p = 0.0033). Among ER-positive patients, COX-2 expression was predictive of worse local control (85% vs. 93%, p = 0.04), distant metastasis (75% vs. 95%, p = 0.002) and worse survival (65% vs. 94%, p = 0.002). Among ER-negative tumors COX-2 expression was not significantly correlated with local control (87 vs. 95%, p = 0.12), distant metastasis (73% vs. 78%, p = 0.39), or survival (77% vs. 87%, p = 0.15).
Conclusions: In breast cancer patients treated with CS+RT, COX-2 expression is associated with younger age, larger tumor size, worse local control, distant metastasis, and worse overall survival. The significance is limited to hormone receptor-positive tumors, consistent with the known effect of COX-2/PGE2 on aromatase activity. Use of COX-2 inhibitors in estrogen-dependent breast cancers warrants further investigation.