Background and objective Theranostic-guided transepithelial corneal cross-linking uses patterned iontophoresis for minimally invasive, spatially controlled riboflavin delivery in progressive keratoconus. This study developed a computational model to investigate the effects of applied current and reservoir-cornea conformity on riboflavin transport. Methods A 2D axisymmetric finite element model of the layered cornea and a clinically available patterned iontophoresis device was developed. The electric field was coupled to Nernst-Planck equations incorporating diffusion, electromigration, electroosmosis, tissue porosity, and tortuosity. The patterned mask was explicitly modeled, while a conformity parameter accounted for reservoir-cornea contact. Model parameters were constrained using two independent in vivo measurements: the effective electrical resistance of the iontophoresis–cornea system and the volume-averaged stromal riboflavin concentration. Results Electromigration promoted localized transport through mask openings, whereas electroosmosis contributed to deeper stromal penetration; passive diffusion was limited by the intact epithelium. At 420 s, the predicted volume-averaged stromal concentrations across the total treatment region were 97 and 110 μg/cm3 at 0.5 and 1.0 mA, respectively, consistent with corresponding in vivo measurements of 90 ± 20 and 111 ± 21 μg/cm3 in 28 keratoconic eyes. The central-to-total concentration ratio increased from 1.95 at 0.5 mA to 2.96 at 2.0 mA, indicating greater spatial selectivity with increasing applied current. Conclusions The model describes transepithelial riboflavin delivery by corneal iontophoresis, highlighting the influence of mask geometry and applied current on spatial riboflavin distribution. Limitations include the 2D axisymmetric approximation, simplified boundary conditions, and lack of direct in vivo validation of riboflavin concentrations at currents above 1.0 mA.
