Enrol every new systemic therapy into a registry linkage that will still report in thirty years

The honest position on the late effects of modern treatment is that they are unmeasurable, not that they are absent. Follow-up of commercial CAR-T cohorts is short relative to the latency of second cancers; the FDA's 2024 labelling action on secondary T-cell malignancy states no incidence; whether the immunotherapies and antibody-drug conjugates now in first-line use carry such a risk is not yet measurable. The same applies in the reassuring direction: whether smaller modern radiotherapy fields lowered the forty-year second cancer rate is not known because nobody has been followed that long.

Waiting will not fix this, because the follow-up mechanism does not exist. Trial follow-up ends when the registration endpoint is met; extension studies are voluntary and lose people; pharmacovigilance describes reporting, not incidence. The paediatric cohorts show what works: abstract exposures at entry, link to registries, and the answer arrives decades later without anyone having to remember.

The proposal is to make that routine. At the point of approval, every new systemic cancer therapy enters a national registry linkage keyed to the treatment record, with second cancers, cardiovascular events and mortality reported from routine data rather than from sponsor follow-up. The marginal cost is in the linkage and the consent architecture, and it is the same infrastructure a late-effects registry needs, which is why the two proposals belong together.