Why Childhood Cancer Arises Without a Precancerous Sequence: Evidence, Gaps, and a Measurement Agenda

Childhood cancer is the most successfully treated group of malignancies in oncology: five-year overall survival for ages 0–19 has reached about 90%, and more than 90% of acute lymphoblastic leukaemia is cured. Three clinical facts limit that success. Relapse remains the leading cause of death and is concentrated among children classified as standard risk; further intensification has not improved survival while adding grade 3–4 toxicity; and the cost is deferred, with cumulative chronic conditions reaching 73.4% at 30 years and 96.0% for grade 3–5 by age 50. Childhood cancer aetiology adds three anomalies that adult templates do not accommodate: the window for mutation accumulation is measured in months rather than decades; no recognisable multi-step precancerous sequence exists in the general paediatric population; and tumours arise along developmental programs, with the same gene inactivation tumorigenic only in a particular cell at a particular stage. This review synthesises the human evidence bearing on these anomalies and asks whether they share one measurable entry point. We organise the evidence around three gates in series—cell-intrinsic, tissue and systemic—and examine the possibility that childhood cancer occupies the failure path in which the cell-intrinsic gate was never built. The built state has a candidate physical readout: the resting membrane potential, which completes a hyperpolarising fall during differentiation and stalls between the proliferative and differentiated positions when construction fails. We specify three fingerprints that can be read in existing registry, pathology and electrophysiological data, and seven tests that would distinguish this account from the mutation-centred one. The contribution is a measurement agenda: it names the state to be read, the material in which it can be read, and the comparison that would count. The readouts are specified so that existing cohorts can be re-analysed without new technology.