Robust genomic copy number predictor of pan cancer metastasis.

Copy number alterations(CNAs) are the most common genetic changes observed in many cancers, reflecting the innate chromosomal instability of this disorder. Yet, how these alterations affect gene function to promote metastases across different tumor types has not been established. In this study, we developed a pan-cancer metastasis potential score (panMPS) based on observed CNAs. panMPS predicts metastasis and metastasis-free survival in cohorts of patients with prostate cancer, triple negative breast cancer and lung adenocarcinoma, and overall survival in the Metabric breast cancer cohort and three cohorts from The Cancer Genome Atlas (TCGA), including prostate, breast and lung adenocarcinoma. These CNAs are present in cell lines of metastatic tumors from eight different origins, reflected by an elevated panMPS for all cell lines. Many copy number alterations involve large chromosomal segments that encompass multiple genes ("clumps"). We show that harnessing this structural information to select only one gene per clump captures the contributions of other genes within the clump, resulting in a robust predictor of metastasis outcome. These sets of selected genes are distinct from cancer drivers that undergo mutation, and in fact, metastasis-related functions have been published for over half of them.

Utilization of the 21-Gene Recurrence Score in a Diverse Breast Cancer Patient Population: Development of a Clinicopathologic Model to Predict High-Risk Scores and Response to Neoadjuvant Chemotherapy.

INTRODUCTION: The 21-gene expression profile [Oncotype DX Recurrence Score (RS)] stratifies benefit from adjuvant chemotherapy in hormone receptor (HR)-positive, HER2/neu-negative, node-negative breast cancer. It is not routinely applied to predict neoadjuvant chemotherapy (NACT) response; data in diverse patient populations also are limited. We developed a statistical model based on standard clinicopathologic features to identify high-risk cases (RS > 30) and then evaluated ability of predicted high RS to predict for NACT downstaging. METHODS: Primary surgery patients with Oncotype DX RS testing 2012-2016 were identified from a prospectively-maintained database. A RS predictive model was created and applied to a dataset of comparable NACT patients. Response was defined as tumor size decrease ≥ 1 cm. RESULTS: Of 394 primary surgery patients-60.4% white American; 31.0% African American-RS distribution was similar for both groups. No single feature reliably identified high RS patients; however, a model accounting for age, HR expression, proliferative index (MIB1/Ki67), histology, and tumor size was generated, with receiver operator area under the curve 0.909. Fifty-six NACT patients were identified (25 African American). Of 21 cases with all relevant clinicopathology, 14 responded to NACT and the model generated high-risk RS in 14 (100%); conversely, of 16 cases generating high-risk RS, only 2 did not respond. CONCLUSIONS: Predictive modelling can identify high RS patients; this model also can identify patients likely to experience primary tumor downstaging with NACT. Until this model is validated in other datasets, we recommend that Oncotype-eligible patients undergo primary surgery with decisions regarding chemotherapy made in the adjuvant setting.