Reliable integration of CyTOF and IMC data in biomarker discovery clinical studies

Speaker: Brice Gaudilliere, MD, PhD, Associate Professor of Anesthesiology, Perioperative and Pain Medicine

Multi-omic predictive modeling in biomarker discovery studies uses transcriptomics, proteomics, metabolomics, cytomics and spatialomics to generate a set of features that can accurately predict a disease state and progression. Gaudilliere discusses a clinical use case to predict the onset of labor in term and preterm pregnancies, for which there are no good diagnostic tests.
To search for common immune pathways of labor onset in these pregnancies, a longitudinal study outlined what changes occur within the maternal immune system as labor approaches. A new machine learning framework, STABL, was then applied to bridge predictive modeling of high-dimensional omics data and selectivity and stability requirements of effective biomarker discovery to reliably predict preterm labor.