Advancing transdermal drug delivery using translational data and mechanistic based modeling

Transdermal drug delivery systems (TDDS) offer patient-centric advantages as drug dosage forms, but candidate selection for transdermal within early discovery remains challenging due to high attrition rates (> 40%) and limited predictive screening methodologies. This work introduces a mechanistic, data-driven in silico model to accelerate candidate selection for transdermal drug delivery, leveraging translational data from four approved products. The model simulates passive diffusive transport across the epidermis, focusing on two critical parameters: the partition coefficient between patch and skin (K_SC/patch) and the diffusion coefficient in the stratum corneum, which together determines the overall transdermal flux. Model reliability was validated using data from four marketed transdermal drugs-fentanyl, rivastigmine, nicotine, and lidocaine-with strong agreement to in vitro and in vivo results. In addition to predictive performance, the platform enables mechanistic interpretation of key transport parameters, establishing in vitro/in vivo correlations via scaling of partition coefficients, and providing a thermodynamic rationale for observed relationships with drug molecular properties. This approach supports rational design and early feasibility assessment of molecules for transdermal delivery systems, supporting efficient identification and optimization of candidate molecules for pre-clinical and clinical translation.