TY - JOUR AU - Sachs, Michael C. PY - 2015 TI - Statistical principles for omics-based clinical trials JF - Chinese Clinical Oncology; Vol 4, No 3 (September 25, 2015): Chinese Clinical Oncology (Clinical Trials in the Genome Era—Study Designs and Endpoints, Practical Considerations - Guest Editors: Daniel Sargent, Sumithra Mandrekar and Axel Grothey) Y2 - 2015 KW - N2 - High-throughput technologies enable the measurement of a large number of molecular characteristics from a small tissue specimen. High-dimensional molecular information (referred to as omics data) offers the possibility of predicting the future outcome of a patient (prognosis) and predicting the likely response to a specific treatment (prediction). Embedded in the vast amount of data is the hope that there exists some signal that will enable practitioners to deliver therapy personalized to the molecular profile of a tumor, thereby improving health outcomes. The challenges are to determine that the omics assays are valid and reproducible in a clinical setting, to develop a valid and optimal omics-based test that algorithmically determines the optimal treatment regime, to evaluate that test in a powerful and unbiased manner, and finally to demonstrate clinical utility: that the test under study improves clinical outcome as compared to not using the test. We review the statistical considerations involved in each of these stages, specifically dealing with the challenges of high-dimensional, omics data. UR - https://cco.amegroups.org/article/view/5940