On-Chip Prediction of Combination Drug Efficacy Using a Magnetofluidic Platform with Integrated Capacitive Sensors

Automated and rapid combination drug screening using lab-on-a-chip platforms could benefit significantly from real-time trials for effective dose identification and optimization without human involvement, while decreasing the tedious pipetting procedures and measuring errors. Most reported platforms focus on individual drug testing using a significantly high number of volume-based compartments, complicating dose detection and imposing the handling of large samples. Meanwhile, probing the efficacy of combination drugs remains largely at its inception. To address such challenges efficiently, we propose a compact magnetically regulated microfluidic (MRM) platform with integrated interdigitated electrode-based capacitive sensors. The device design comprises 3 inlets, each paired with a staircase-configured patterned magnet that directs flow into 5 dispersion microchannels per inlet─yielding a total of 15 microchannels independently routed to 5 outlet wells. Such an inlet-to-outlet architecture mimics a multiplex-demultiplex configuration, with isolated dispersion microchannels that prevent cross-contamination prior to collection at the outlet wells. It is experimentally demonstrated that clinically relevant concentrations (10–485 μg mL –1 ) of different magnetically coated drug combinations can be sorted rapidly within the five wells of the designed device, and the delivered doses are quantified using the integrated sensor interface. Such unique concentration variation is attributed to the localized magnetic field-induced deflection of the magnetically tagged drugs in the predefined microchannel regions of the device, controlled by tuning the flow rate in the range of 0.9–2 μL min –1 . Human breast cancer cells (MCF-7) are cultured in these specifically designed wells to probe the efficacy of exposure to several drug combinations, including Curcumin, Paclitaxel, and Quercetin. Using the developed platform, we identified the most efficacious drug combinations, including the appropriate concentration and ratio that cytotoxically affected these cells. Also, the collected data sets are used in an OLS (Ordinary Least Squares) regression model to predict the drug efficacy in test samples, which allows for the reduction of iterative on-chip experiments while reducing cost and conserving precious reagents.