Carcinogenicity Signal

A carcinogenicity signal is an indication that a compound may increase tumour formation. The standard non-clinical package is a pair of two-year bioassays in rat and mouse at several dose levels, supported by a genotoxicity battery that asks the separate question of whether the molecule damages DNA directly. Signals also arise mechanistically, when a drug stimulates a growth-promoting axis or causes sustained hyperplasia of a cell type, and epidemiologically, from cohort studies or disproportionate reporting after launch.

GLP-1 receptor agonists are the case every reader of this field meets. Chronic dosing produced dose- and duration-dependent thyroid C-cell hyperplasia, adenomas and carcinomas in rats and mice. Rodent C-cells express GLP-1 receptors densely; human C-cells express them sparsely, and calcitonin monitoring across the clinical programmes did not reproduce the finding. The European Medicines Agency reviewed accumulated thyroid cancer reports and concluded the evidence did not support a causal association. The growth hormone axis poses a different version: higher circulating IGF-1 is associated with certain cancers in observational cohorts without that establishing causation.

What determines the weight of a signal is whether the mechanism is conserved across species, how large the exposure multiple was at the tumorigenic dose, and whether human follow-up is long enough to matter.

The misuse is symmetric. Rodent tumours at a lifetime maximum tolerated dose are dismissed as irrelevant when species differences have not been demonstrated, and disproportionality scores from reporting databases are quoted as if they were rates. The quieter error concerns unapproved research peptides: no two-year bioassay has ever been run on most of them, so there is no signal to find, and silence is read as a negative result.