#39 · Systems biology

Reduced dynamical models convert treatment failure into dosing and relapse predictions

The common thread is the use of reduced or hybrid dynamical models to turn treatment failure into testable predictions about dosing and relapse. The CLL framework couples patient-specific ibrutinib pharmacokinetics with stochastic Boolean gene regulatory dynamics, using single-cell data to model relapse and screen BTK-associated targets. The CML study uses a minimal two-equation model with the immune compartment reduced by quasi-steady-state, reporting that leukemic fitness predicts treatment-free remission and can guide adaptive therapy. The persister model is phenomenological, describing plasticity-driven drug persistence across a continuum and predicting an optimal drug concentration and drug holiday schedule. What is new is the explicit pharmacokinetic-to-regulatory coupling in the CLL work and the fitness-based prediction of treatment-free remission in CML. All results are model-derived predictions; the summaries report no experimental validation beyond these stated predictions.