Gradient-driven microfluidic framework for the screening of ratiometric drug combinations towards precision oncology in paediatric leukaemia.

Improving outcomes for high-risk acute lymphoblastic leukaemia (ALL) patients requires increasingly personalized treatment strategies that leverage molecular and omics-based assays as well as functional screens. However, to date, only a few of these approaches are clinically actionable and easily adopted in clinical settings. To that end, we report a combination drug screening assay (GradientScreen) that correlates with clinical outcomes using standard paediatric ALL treatment protocols. This assay uses a microfluidic device to generate overlapping drug gradients and implements an analysis framework that characterizes drug synergy and efficacy in a ratiometric manner. Using partial least squares regression, we show that our combination response metrics correlate with both minimal residual disease (MRD) following initiation of treatment and later clinical end-points. Here, we further demonstrate that this system could be used to evaluate response to experimental therapeutic regimens in patients who did not respond to standard treatment. This assay can serve as a research-enabling tool for the field, with potential to integrate into clinical pathology workflows to identify difficult-to-treat ALL patients and screen for alternative treatment strategies at diagnosis.