Modeling Magnetic Hyperthermia of ZnO:Fe₃O₄:Chitosan Nanocomposites in Perfused Tumor Tissue

Magnetic hyperthermia models often use arbitrarily chosen or post hoc adjusted inputs, which obscures what determines tumor heating and thermal dose. In the present MATLAB model, every input is traced to a published measurement and the heat source is computed, not prescribed. A temperature-dependent two-level model of the non-linear Stoner–Wohlfarth hysteresis of immobilized magnetite cores is coupled to the Pennes bioheat equation and to the CEM43 thermal dose, for spheres with the volumes of pediatric CNS neuroblastoma (FOXR2-activated) under the glioblastoma clinical protocol. Monte Carlo and Sobol analyses over literature bounds identify the controlling inputs. At the clinical dose, the ZnO:Fe₃O₄:chitosan mass ratio does not affect the heat source, because only the magnetic phase generates heat. Linear response theory overestimates the loss by 63% at the Atkinson–Brezovich (AB) field. A median of 43 °C in a 115 mL tumor requires about 4.9 kA m⁻¹, whereas the center of a 10.5 mL lesion stays below 43 °C at the AB field. The reported range of the Néel attempt time can reduce the loss of monodisperse cores a hundredfold, whereas the reported size spread triples it; over these bounds, the median temperature at the AB field spans 37–59 °C (5th–95th percentiles).