A Multiscale Mechanistic Framework Linking Infection Dynamics, Oxidative Chemiexcitation, and Ultraweak Photon Emission

Ultraweak photon emission (UPE) provides a noninvasive optical signature of oxidative chemistry, but the relationship between infection-associated reactive oxygen species (ROS) and photon emission remains mechanistically uncertain. Here, we develop a multiscale framework linking pathogen dynamics, immune activation, signaling ROS, emission-relevant oxidative chemistry, chemiexcitation-capable intermediates, electronically excited molecular states, and wavelength-resolved UPE. Chemiexcitation is represented as incoherent Lindblad pumping, allowing stochastic oxidative reactions to populate molecular excited states without assuming coherent ROS-driven optical excitation. The model was evaluated using literature-constrained oxidative inputs for healthy, severe COVID-19, and sepsis conditions, together with Latin Hypercube uncertainty propagation, Sobol sensitivity analysis, fixed-ROS counterfactual testing, spectral robustness analysis, and a pathway-level PMA/DPI intervention consistency test. Two admissible oxidative-to-photon mappings produced sharply different quantitative predictions from the same clinical ROS inputs. Under a high-gain structure, 9.01-fold and 13.86-fold oxidative increases produced 42.21-fold and 89.02-fold increases in peak UPE, whereas a saturating structure compressed the same inputs to 1.445-fold and 1.450-fold. Local elasticity remained near 1.7 under the high-gain mapping but declined to 0.011 and 0.005 at the COVID-19 and sepsis inputs under the saturating mapping. At fixed ROS, downstream parameter uncertainty produced more than a 200-fold spread in predicted UPE under the high-gain structure. Sobol analysis identified the saturation scale, emission-relevant oxidative lifetime, and signaling-to-emission conversion as the dominant contributors to output variance. In a pathway-level consistency test, a subset of high-gain realizations reproduced the reported DPI/PMA residual-UPE interval under physically admissible residual oxidative drive, whereas none of the sampled saturating realizations did. The model also predicted a progressive redistribution of spectral intensity toward longer wavelengths with increasing oxidative burden, although absolute spectral centroids remained dependent on the assumed spectral representation and emitter weighting. These results show that current ROS measurements constrain the direction of the UPE response more strongly than its quantitative magnitude and do not identify a unique universal ROS-to-UPE transfer function.