Impact of Prosigna test on adjuvant treatment decision in lymph node-negative early breast cancer-a prospective national multicentre study (EMIT-1).

BACKGROUND: EMIT-1 is a national, observational, single-arm trial designed to assess the value of the Prosigna, Prediction Analysis of Microarray using the 50 gene classifier (PAM50)/Risk of Recurrence (ROR), test as a routine diagnostic tool, examining its impact on adjuvant treatment decisions, clinical outcomes, side-effects and cost-effectiveness. Here we present the impact on treatment decisions. PATIENTS AND METHODS: Patients with hormone receptor-positive, human epidermal growth factor receptor 2-negative pT1-pT2 lymph node-negative early breast cancer (EBC) were included. The Prosigna test and standard histopathology assessments were carried out. Clinicians' treatment decisions were recorded before (pre-Prosigna) and after (post-Prosigna) the Prosigna test results were disclosed. RESULTS: Of 2217 patients included, 2178 had conclusive Prosigna results. The pre-Prosigna treatment decisions were: no systemic treatment (NT) in 27% of patients, endocrine treatment alone (ET) in 38% and chemotherapy (CT) followed by ET (CT + ET) in 35%. Post-Prosigna treatment decisions were 25% NT, 51% ET and 24% CT + ET, respectively. Adjuvant treatment changed in 28% of patients, including 21% change in CT use. Among patients assigned to CT + ET pre-Prosigna, 45% were de-escalated to ET post-Prosigna. Of patients assigned to ET, 12% were escalated to CT + ET and 8% were de-escalated to NT; of those assigned to NT, 18% were escalated to ET/CT + ET. CT was more frequently recommended for patients aged ≤50 years. In the subgroup with pT1c-pT2 G2 and intermediate Ki67 (0.5-1.5× local laboratory median Ki67 score), the pre-Prosigna CT treatment decision varied widely across hospitals (3%-51%). Post-Prosigna, the variability of CT use was markedly reduced (8%-24%). The correlation between Ki67 and ROR score within this subgroup was poor (r = 0.25-0.39). The median ROR score increased by increasing histological grade, but the ROR score ranges were wide (for G1 0-79, G2 0-90, G3 16-94). CONCLUSION: The Prosigna test result changed adjuvant treatment decisions in all EBC clinical risk groups, markedly decreased the CT use for patients categorized as higher clinical risk pre-Prosigna and reduced treatment decision discrepancies between hospitals.

Focal amplification and oncogene dependency of GAB2 in breast cancer.

DNA amplifications in breast cancer are frequent on chromosome 11q, in which multiple driver oncogenes likely reside in addition to cyclin D1 (CCND1). One such candidate, the scaffolding adapter protein, GRB2-associated binding protein 2 (GAB2), functions in ErbB signaling and was recently shown to enhance mammary epithelial cell proliferation, and metastasis of ERBB2 (HER2/neu)-driven murine breast cancer. However, the amplification status and function of GAB2 in the context of amplification remain undefined. In this study, by genomic profiling of 172 breast tumors, and fluorescence in situ hybridization validation in an independent set of 210 scorable cases, we observed focal amplification spanning GAB2 (11q14.1) independent of CCND1 (11q13.2) amplification, consistent with a driver role. Further, small interfering RNA (siRNA)-mediated knockdown of GAB2 in breast cancer lines (SUM52, SUM44PE and MDA468) with GAB2 amplification revealed a dependency on GAB2 for cell proliferation, cell-cycle progression, survival and invasion, likely mediated through altered phosphatidylinositol 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK) signaling. GAB2 knockdown also reduced proliferation and survival in a cell line (BT474) with ERBB2 amplification, consistent with the possibility that GAB2 can function downstream of ERBB2. Our studies implicate focal amplification of GAB2 in breast carcinogenesis, and underscore an oncogenic role of scaffolding adapter proteins, and a potential new point of therapeutic intervention.

Somatic mutation databases as tools for molecular epidemiology and molecular pathology of cancer: proposed guidelines for improving data collection, distribution, and integration.

There are currently less than 40 locus-specific databases (LSDBs) and one large general database that curate data on somatic mutations in human cancer genes. These databases have different scope and use different annotation standards and database systems, resulting in duplicated efforts in data curation, and making it difficult for users to find clear and consistent information. As data related to somatic mutations are generated at an increasing pace it is urgent to create a framework for improving the collecting of this information and making it more accessible to clinicians, scientists, and epidemiologists to facilitate research on biomarkers. Here we propose a data flow for improving the connectivity between existing databases and we provide practical guidelines for data reporting, database contents, and annotation standards. These proposals are based on common standards recommended by the Human Genome Variation Society (HGVS) with additions related to specific requirements of somatic mutations in cancer. Indeed, somatic mutations may be used in molecular pathology and clinical studies to characterize tumor types, help treatment choice, predict response to treatment and patient outcome, or in epidemiological studies as markers for tumor etiology or exposure assessment. Thus, specific annotations are required to cover these diverse research topics. This initiative is meant to promote collaboration and discussion on these issues and the development of adequate resources that would avoid the loss of extremely valuable information generated by years of basic and clinical research.

High frequency of skewed X inactivation in young breast cancer patients.

INTRODUCTION: Patients with invasive ovarian cancer were recently shown to have a higher frequency of skewed X chromosome inactivation in peripheral blood cells compared to patients with borderline cancer and controls. In this study, we analysed the X inactivation pattern in peripheral blood from 216 breast cancer patients. METHODS: X inactivation analysis was performed using HpaII predigestion of DNA followed by PCR of the highly polymorphic CAG repeat of the androgen receptor gene (AR), which amplifies the undigested inactive X chromosome only. The X inactivation pattern was classified as skewed when 90% or more of the cells preferentially used one X chromosome. RESULTS: Young breast cancer patients (27-45 years) had a higher frequency of skewed X inactivation than young controls (13 and 1%, respectively) (p=0.009), whereas no difference was found for middle aged and older patients compared to controls of a similar age. CONCLUSIONS: A germline mutation in an X linked tumour suppressor gene may give a proliferative advantage to cells with this mutation on the active X chromosome, thus causing skewed X inactivation and an increased risk for developing cancer. Another possible explanation could be that females with a constitutionally skewed X inactivation pattern are more susceptible to develop breast cancer because of an X linked low penetrance susceptibility allele that is affected by the inactivation pattern.

BRCA1 screening in patients with a family history of breast or ovarian cancer.

Several studies have shown that the frequency of BRCA1 mutations in high-risk families differs widely between populations. In a recently published study from the Stockholm region, we found BRCA1 mutations in about 35% of the breast/ovarian families, but only in 1% of the families with site-specific breast cancer. To determine the frequency of BRCA1 mutations in families with a less increased risk for breast or ovarian cancer, a second study was performed. A total of 94 families with two and six families with only one affected member were included. Six mutations were found, all localized in exon 11, and five of them were previously known Swedish founder mutations. The mutation frequency was 6%, similar to the finding in families fulfilling the criteria for hereditary breast cancer (7%) that was disclosed in our first study. All families with a mutation had at least one individual with ovarian cancer. Thus, our study further implies that for a woman with breast cancer, a family history of ovarian cancer is far more important than a family history of breast cancer for predicting a BRCA1 mutation. Our results can be used to increase the specificity in selection of families for genetic testing.