Dynamics of Early Cancer Growth Inferred from Tracking Millions of Clonal Tumors in Vivo

Abstract: Although cancers are often recognized as heterogeneous at later stages, recent evidence shows that this variability emerges much earlier: even the smallest lesions can diverge dramatically in growth. This observation suggests that tumor trajectories are shaped not only by later-acquired mutations, but also by stochastic clonal dynamics, variation in initial cellular states, and the surrounding tissue environment. Despite their potential importance in determining which lesions will expand versus remain dormant, we still lack a quantitative understanding of the sources of early heterogeneity. To address this, we leveraged genetic barcoding to track 30 million clonal lung tumors in genetically engineered mice for nearly a year. Tumor sizes follow a broad distribution that is roughly power law over much of their range, with an exponent that decreases over time, constraining possible growth models. We also observe that at later times, the distribution becomes truncated in the upper tail, suggesting that largest tumors stall. Preliminary results suggest that secondary mutations, such as tumor-suppressor inactivation, can sometimes help early clones overcome this limitation. Together, these results provide a high-resolution view of early tumor evolution and a framework for connecting empirical data to theoretical models.