Integrated Lineage Tracing in Hepatoblastoma Finds a Plasticity-Proliferation Axis That Drives Post-Treatment Adaptation

Hepatoblastoma (HB) has one of the lowest mutational burdens among childhood cancers, limiting the role of genetic selection. Nevertheless ~20% of patients relapse, implicating non-genetic mechanisms, such as phenotypic plasticity, in treatment adaptation. The temporal and clonal dynamics of post-treatment plasticity in HB remain poorly defined. Here, we integrate expressed DNA barcoding approaches with single-cell multiomics in preclinical models to simultaneously trace clonal and phenotypic dynamics following cisplatin treatment.We observe that cisplatin selects for progenitor-like states, which persist as reservoirs for phenotypic re-diversification. Barcode lineage tracing reveals a subset of persister clones stochastically resume proliferation and transition from less- to more-differentiated phenotypes, an observation also captured in patient data. Pseudotime and landscape approaches identify a plasticity-proliferation axis underlying recovery post-treatment, driven by coordinated activation of E2F transcription factors and their downstream target BIRC5 (survivin). Downregulation and pharmacological inhibition of BIRC5 disrupt this axis, shifting phenotypic dynamics towards less-differentiated states and killing cisplatin-persister cells, supporting a role for BIRC5 in plasticity-led awakening from persistence. Consistently, elevated BIRC5 expression is associated with poor patient outcome.Together, these findings establish a mechanistic link between persistence, phenotypic plasticity, and stochastic clonal outgrowth in HB, and identify BIRC5 as a regulator of plasticity-driven adaptation and a therapeutically actionable vulnerability to halt treatment adaptation.