A point-of-care testing biosensing platform integrated with physiologically based pharmacokinetic model for on-site therapeutic drug monitoring of methotrexate

Methotrexate (MTX) has definite dose-response relationship, in which high level MTX in serum leads to severe adverse events such as hepatotoxicity and myelosuppression. Therapeutic drug monitoring (TDM) of MTX helps to improve clinical outcomes and reduce toxic side effects. Nevertheless, standard methods are restricted by time-consumption, high cost and multiple repetitive sampling. Herein, a point-of-care testing (POCT) sensor was integrated with physiologically based pharmacokinetic (PBPK) model for on-site TDM of MTX to obtain patient-specific concentration-time ( C-T ) profiles, guiding leucovorin rescue and managing toxicities of high dose MTX. The sensor employed graphene oxide (GO)-ionic liquid (IL) modified screen-printed carbon electrodes in conjunction with a portable electrochemical workstation to achieve fast and accurate MTX detection in human serum via cyclic voltammetry, exhibiting two distinct linear dynamic ranges (0.1 nM-100 μM and 100 μM-3 mM) with a detection limit of 1.3 pM. The GO-IL composite enables selective MTX detection through different oxidation potentials, π-π stacking with MTX's pteridine ring, and charge/hydrophobicity-guided exclusion of interferents, achieving discrimination against key interferents including 7-OH MTX. Pearson correlation analysis and Bland-Altman consistency analysis demonstrated that MTX concentrations of 86 clinical samples measured by the sensor were consistent with that of immunoassay. Innovatively, PBPK model of MTX was developed to effectively process the complex MTX concentration data to obtain personalized C-T profiles of each patient. The individual PK parameters like maximum concentration and area under the curve can be easily acquired through C-T profiles, allowing for rapid clinical decision making. The MTX-PBPK model displayed reliable predictive accuracy of PK parameters within 2.0-fold-error. The results demonstrate that the POCT sensor combined with PBPK model is promising to be used for on-site TDM of MTX, promoting the development of individualized medication.