Many are better than one - next generation multivariate biomarkers for precision oncology

Existing companion diagnostics have helped to match drug treatments to patients. However, they are largely restricted to single-molecule, single-time-point measurements, which cannot capture the full dynamic complexity of cancer biology. The development of multivariate and even dynamic biomarkers for diagnostic assays could allow more patients to benefit from improved drug regimens. Here we describe our work which provides a case study of multivariate biomarkers where we integrated experimental data generated using multivariate profiling technologies with a variety of computational modeling and simulation methods to identify such biomarkers and make clinical predictions on their therapeutic utility. We believe this approach of integrating multivariate profiling technologies and computational models, and iterating between experimental discovery and model predictions, will be required to develop the next generation of multivariate diagnostics and realize the promise of precision medicine.