Tumorigenesis is the consequence of game interactions involving tumor cells, tumor-microenvironments, and adjacent normal cells. Charting the roadmap of such tripartite interactions can reveal how tumors grow and metastasize. This perspective introduces a generalized statistical mechanics platform – GameTalker – for tracing cell-cell crosstalk and communication that occurs not only pairwise but also at high orders to mediate tumor growth. GameTalker unifies evolutionary game theory, ecological niche theory, modularity theory, and graph theory into a computational framework, under which all intra- and extra-tumoral agents are assembled into sparsely-linked but well-coordinated networks. By decomposing the overall “payoff” of individual agents into independent components (due to their own intrinsic capacity) and dependent components (resulting from regulation of other agents), GameTalker can characterize in which way each agent contributes to tumor growth, directly or indirectly, actively or passively, and singly or collectively. Such information is fundamental for downstream researchers to design precision cancer therapies for obstructing tumor progression through altering cell-cell communication.
Wang et al. (2026) studied this question.