Molecular heterogeneity of breast cancer: implications for treatment and clinical trial design

Abstract

Comprehensive molecular analysis of breast cancer indicates that breast carcinoma is not a single disease with heterogeneous morphology and variable hormone receptor expression, but a collection of molecularly distinct neoplastic diseases of the breast. The magnitude of molecular differences that are seen between estrogen receptor (ER)-positive and ER-negative cancers at the mRNA expression or the DNA copy number level are similar to the extent of molecular differences that distinguish acute myeloid leukemias from lymphoid leukemias. The clinical differences between ER-negative and ER-positive cancers have long been recognised and the recent genomic data provided further evidence that these cancers represent different diseases. Furthermore, two major groups within the ER-positive cancers can also be easily recognised at the molecular level: one that corresponds to high-grade, highly proliferative tumors; and the other to lower-grade cancers with low proliferation rate. These two molecular subsets have very different prognosis: the former group is less sensitive to endocrine therapy and has poor prognosis but is more sensitive to cytotoxic drugs, and therefore adjuvant chemotherapy may improve outcome. The latter group has excellent prognosis with endocrine therapy alone, and does not appear to benefit much from adjuvant chemotherapy. Several different molecular assays can assist in distinguishing between these different prognostic subsets of ER-positive cancers; the most commonly used commercial test in the United States is Oncotype DX; however, the MammaPrint and GenomicGrade molecular assays are also able to identify ER-positive patients with excellent prognosis with endocrine therapy. In the absence of the statistically more accurate multivariate molecular prediction models, the histologic grade, HER2 status and Ki67 may be used as a poor man's alternative to estimate prognosis and chemotherapy sensitivity of ER-positive cancers. The recognition that breast cancer is not a single disease has important consequences for clinical trialists and academic investigators. Different diseases require separate clinical trials, and different biomarkers may be needed to predict response or prognosis accurately in the different disease subsets. Different clinical issues are pressing for the different types of breast cancers, better systemic therapies are needed for triple receptor-negative cancers, proper sequencing and length of adjuvant endocrine therapy is an issue yet to be clarified for endocrine-sensitive ER-positive cancers, and optimal use of chemotherapy and better drugs are needed for the poor-prognosis ER-positive cancers. It is also apparent that different biomarkers are needed for the different cancer subsets. For example, proliferation-related markers are strongly prognostic and also predictive of chemotherapy sensitivity among ER-positive cancers but these same markers are not predictive in ER-negative cancers. Traditional breast cancer studies where all patients with cancer of the breast are eligible for the same therapy will soon be regarded as naive as a clinical trial proposal to treat all acute and chronic leukemias with the same drug in a single trial and perform subset analysis for the various cytologic types at the end.