Development of a Spatial Drug Penetration Algorithm for Optimizing Therapeutic Molecules in Breast Tumor Microenvironments

Limited penetration of therapeutic molecules into solid breast tumors is a major cause of heterogeneous exposure and incomplete response. We developed SpaDPA, a spatial drug penetration algorithm that couples heterogeneous tumor fields, reaction-diffusion transport, cellular uptake, extracellular-matrix binding, molecular degradation, and vascular source terms with multi-objective molecular optimization. The algorithm constructs two-dimensional digital breast-tumor phantoms from normalized maps of collagen density, cellularity, hypoxia, and vascular accessibility. Twelve candidate molecular profiles were evaluated using an explicit finite-difference solver and ranked by intratumoral penetration index, volumetric coverage, penetration depth, and predicted systemic exposure. On held-out digital phantoms, SpaDPA achieved an R² of 0.02, RMSE of 0.024, and MAE of 0.020, outperforming homogeneous-diffusion and conventional porous-media models. The leading molecular profile, SPD-07, achieved a penetration index of 0.332 and maintained measurable concentration across 0 µm of tumor depth. Sensitivity analysis identified effective diffusivity, molecular weight, lipophilicity, and matrix binding as dominant determinants of spatial delivery. These findings establish a reproducible computational framework for designing molecules that are optimized not only for target affinity but also for access to spatially protected tumor regions.