Gene-expression profiling in breast cancer: bespoke cancer therapy or more fiction than science?

Abstract

Breast cancer is the most common life threatening malignancy in women. Despite advances in molecular therapeutics, many patients with apparently early stage and curable disease still develop progressive and fatal systemic disease. Haffty et al. [1] demonstrated that despite conservative surgery and radiation followed by administration of chemotherapy in early-stage (T1 tumours) breast cancer patients, the distant metastasis-free survival rate was 71% at 5 years in patients with triple negative disease and 83% at 5 years in non-triple negative. These patients have potentially been under-staged by conventional staging, and therefore may not have received optimum therapy due to their misleading clinical disease profile. Conversely, patients receiving systemic therapy may not derive any additional benefit. Breast cancer is an intricate genetic milieu characterised by numerous molecular alterations that prevents standardising therapeutic strategies to all patients. Until recently, it has been difficult to accurately profile and predict the biological properties and clinical behaviour of breast tumours, however, with the development of DNA microarray technology, it is now possible to analyse the RNA expression of several thousands of genes simultaneously. The clinical application of this technology has the potential to signal a paradigm shift in patient selection for systemic therapy. First, it provided a new molecular classification of breast cancer. This classification was proposed by Perou et al. [2], and subdivided breast cancers into four classes. Luminal-A cancers are ER-positive and mostly low grade. Luminal-B cancers are also ER-positive but tend to be high grade. Basal-like cancers are triple negative and Her-2-like cancers are Her-2 positive and usually ER negative. Currently, there are a number of validated commercially available multigene assays that identify expression of distinct sets of genes in breast cancer: Mammaprint (Agendia), Oncotype DX (Genomic Health), MapQuant and H/I (AvariaDX) (see Table 1 for comparison of assays). Mammaprint uses DNA microarray technology, whereas Oncotype DX and H/I use real time PCR based assays. These assays accurately predict the potential for disease recurrence and enhance patient selection for chemotherapeutic strategies. Mammaprint uses a 70-gene predictor and is involved in the MINDACT (Microarray In Node negative Disease may Avoid ChemoTherapy) trial [3]. The objective is to administer chemotherapy to patients deemed at high risk of recurrence, and hormonal therapy only to low risk patients, as determined by the Mammaprint gene profile. Having recruited large numbers of patients who are stratified based on the clinical and pathological risks, it is hoped that this study will provide important information regarding the prognostic capabilities of the Mammaprint test and also validate its use in identifying women associated with improved survival. Previously, Buyse et al. [4] demonstrated that Mammaprint outperformed Adjuvant Online in predicting survival in an independent group of patients. Unfortunately, however, a major disadvantage of Mammaprint is that it requires fresh tissue for analysis, which may make its clinical integration difficult. The Oncotype DX and H/I assay, however, can use archival tissue (tissue from original biopsy) and may be more readily applied to the clinical setting. By analysing the expression of 21 known genes, Oncotype DX accurately determined the prognostic outcome in hormone receptor positive cancers post adjuvant tamoxifen in a retrospective cohort of 668 M. Barry M. R. Kell (&) Eccles Breast Screening Unit, Mater Misericordiae University Hospital, Eccles St, Dublin 7, Ireland e-mail: malcolm.kell@breastcheck.ie