Machine Learning Identifies Serological Predictors of COVID-19 Outcomes
• A study of hospitalized and non-hospitalized COVID-19 patients used machine learning to identify key serological markers associated with disease severity.
• Higher levels of binding antibodies, complement activation, and ACE2 inhibition were correlated with reduced risk of intubation or death.
• Random forest models demonstrated that serological data, combined with demographics, improved prediction of COVID-19 outcomes.
• The findings offer insights into immune responses and potential targets for therapeutic interventions in severe COVID-19 cases.
BI 836880 Dose Optimization via Biomarker and PK/PD Modeling in Advanced Solid Tumors
• Researchers utilized a comprehensive biomarker and modeling approach to optimize the dose of BI 836880, a VEGF/Ang-2 inhibitor, in advanced solid tumors.
• Population PK/PD modeling integrated data from Phase I studies to predict BI 836880 concentrations and Ang-2 inhibition, informing dose selection.
• Simulations indicated the probability of achieving target Ang-2 inhibition levels with different BI 836880 doses, supporting individualized treatment strategies.
• The study demonstrates the utility of biomarker-driven modeling in optimizing drug dosing and improving outcomes in cancer therapy.
© Copyright 2025. All Rights Reserved by MedPath