Biological mechanisms that trigger breast cancer (BC) tumor progression are molecular subtype dependent.

Sotiriou C, Desmedt C, Haibe-Kains B, Harris A, Larsimont D, Buyse M, Wirapati P, Delorenzi M, Bontempi G, Piccart MJ

Background: We have recently developed several gene expression indices related to hallmarks of breast cancer involving various biological processes such as tumor invasion, impairment of immune response, sustained angiogenesis, evasion of apoptosis and self-sufficiency in growth signal, and investigated their impact on clinical outcome. Here, we aim to refine our biological understanding and the prognostic impact of these indices according to the previously described molecular subtypes based on the estrogen (ER) and ERBB2 receptors.

Methods: Each of these indices were developed in a series of 581 BC samples and then computed on several publicly available microarray studies totaling over 2100 BC patients. Multivariate analyses were used to study the dependency patterns between these indices, the molecular subtypes and their impact on survival.

Results: ER- /ERBB2- and ERBB2+ subgroups were significantly associated with high expression levels of the proliferation, tumor invasion, angiogenesis and immune response indices. Multivariate analysis showed that in the ER+/ERBB2- subgroup, only tumor size and the proliferation and tumor invasion indices were significantly associated with clinical outcome, with the proliferation index having the largest HR and most significant p-value (HR 3.25; CI 2.31-4.56; p=1.2 10-11). In contrast, in the ER-/ERBB2- subgroup, only tumor size (HR 2.08; CI 1.14-3.81; p=0.01) and immune response index (HR 0.66; CI 0.46-0.95; p=0.02) were associated with prognosis whereas in the ERBB2+ tumors only nodal status (HR 3.40; CI 0.96-12.10; p=0.05) and tumor invasion index (HR 3.03; CI 1.32-6.95; p=0.009) showed significant association with survival. Of interest, proliferation index lost its significance as almost all ER-/ERBB2- and ERBB2 + tumors showed high proliferation levels.

Conclusions: Although proliferation seems to be the strongest parameter predicting clinical outcome in ER+/ERBB2- subtype, immune response and tumor invasion appear to be the main molecular mechanisms associated with prognosis in the ER-/ERBB2- and ERBB2+ subgroups respectively. Defining these clinico-genomic models in the specific molecular subgroups will be the key to success for personalized medicine.