Utility of a simplified molecular classification of tumors for predicting survival of patients with invasive ductal breast carcinoma.

BACKGROUND In a recent report it was shown that molecular subgroups of early (pT1) breast tumors belonging exclusively to the most common histological variant, ductal-infiltrating carcinoma, showed significantly different clinical and biological features. OBJECTIVE To test in a series of patients with ductal-infiltrating carcinoma encompassing all stages of the disease if the above-mentioned biological differences already detected at the earliest stage are ultimately reflected in survival differences. PATIENTS AND METHODS All patients with ductal infiltrating carcinoma operated upon at Hospital de Móstoles, Madrid, Spain, between 1997 and 2002 were included into the study, to allow for at least five years of follow-up for survivors. Of 242 studied patients, according to the International Federation of Gynecology and Obstetrics (FIGO) classification, 37.6% were in stage I, 47.2% in stage II, 11.0% in stage III, and 4.2% in stage IV. According to the simplified molecular classification of Carey et al., 116 tumors (47.9%) expressed estrogen receptors, and did not express c-erb-B2 (Luminal A), 67 (27.7%) coexpressed hormone receptors (either estrogen receptors, progesterone receptors or both) and c-erb-B2 (Luminal B), 33 (13.3%) expressed c-erb-B2 in the absence of hormone receptors (HER-2), and 26 (10.7%) neither expressed hormone receptors, nor c-erb-B2 (triple-negative, basal). RESULTS In a univariate model, both disease-free survival and overall survival of the patients were significantly associated with stage (p=0.0003; p<0.0001), histological grade (p<0.0001; p<0.0001), lymphatic vascular space invasion (p=0005; p=0.0044), menopausal status (p=0.04; p=0.034) and molecular subgrouping (p=0.037: p=0.01). In a multivariate model, only stage (p=0.013), grade (p<0.0001), and menopausal status (p=0.007) retained their prognostic power for predicting disease-free survival, and just stage (p<0.0001) and grade (p<0.0001) for predicting overall survival. The molecular classification of the tumors almost reached statistical significance for predicting overall survival (p=0.06).