Tkachenko, S.; Hinczewski, M.; McFarland, C. D.

Cancer progression is increasingly understood as an evolutionary process shaped not only by competition but also by cooperative interactions including those mediated through diffusible ``public goods'' (PGs). Classical evolutionary game theory predicts that PG-producing (altruistic) subclones cannot invade well-mixed populations of non-producers, creating a paradox given their observed emergence in tumors. Here, we resolve this contradiction by combining stochastic spatial simulations with an analytically tractable Moran model to study the invasion dynamics of PG-producing cells in structured populations. Starting from a single producer cell, we explicitly model stochastic PG secretion, diffusion, binding/unbinding, and cell proliferation across biologically relevant parameter ranges. We demonstrate that spatial structure fundamentally alters invasion dynamics, enabling PG producers to invade and establish even when production incurs a fitness cost. Both numerical and analytical approaches converge on a key unifying parameter, a characteristic length scale {delta}, that captures the combined effects of diffusivity, binding kinetics, and degradation. This length scale determines the spatial extent of PG availability and thus the selective advantage of producers. We identify distinct regimes: when PGs are localized (small {delta}), producers preferentially benefit and invasion is likely; when PGs are widely dispersed (large {delta}), benefits are shared and invasion approaches neutrality or is suppressed by costs. Our results highlight that invasion of cooperative traits is governed by spatially mediated resource localization rather than intrinsic fitness alone. This framework provides a mechanistic basis for understanding the emergence of cooperative subclones in tumors and suggests that modulating biophysical transport properties of signaling molecules could influence tumor evolution, metastasis, and therapeutic resistance.