Background: Serum HER-2/neu antigen concentrations have been reported to correlate with increased tumor volume in patients with breast cancer. We measured serum CA 15-3, a surrogate marker of disease burden, and correlated serum CA 15-3 with serum HER-2/neu and analyzed the association of both markers with clinical outcomes.
Methods: Pretreatment serum samples from 566 patients were retrospectively analyzed from 2 phase III clinical trials of estrogen receptor-positive (ER(+)), ER(-)/progesterone receptor-positive, or ER status unknown metastatic breast cancer patients randomized in two similar studies to receive second-line hormone therapy with either megestrol acetate or an aromatase inhibitor (fadrozole). The extracellular domain of the HER-2/neu (c-erbB-2) oncogene and serum CA 15-3 were measured by ELISA on the Bayer Immuno 1.
Results: Serum HER-2/neu protein was increased in 168 patients (30%), and CA 15-3 was increased in 337 (60%) patients. Serum CA 15-3 and HER-2/neu were weakly correlated (r = 0.39; P <0.0001). The clinical benefit (complete responses plus partial responses plus stable disease) of endocrine therapy was significantly lower in patients with increased serum HER-2/neu. When adjusted for serum HER-2/neu, serum CA 15-3 was not predictive of response rates. The median time to progression was shorter in patients with increased serum HER-2/neu (89 days) compared with patients with normal serum HER-2/neu (176 days). Survival was significantly shorter in patients with increased serum HER-2/neu (513 vs 869 days; P <0.0001) or increased serum CA 15-3 (689 vs 939 days; P <0.0001). This observation was confirmed by multivariate analysis.
Conclusions: Serum HER-2/neu is a significant independent predictive and prognostic factor in hormone receptor-positive metastatic breast cancer, even when adjusted for tumor burden as measured by CA 15-3. The combination of increased serum HER-2/neu and increased serum CA 15-3 predicts a worse prognosis than does increased CA 15-3 alone.