Cancer

The number of cancers rises every year in every ageing country, because cancer is a disease of age; the age-standardised picture is more mixed. Lung cancer in men, stomach and cervical cancer have fallen for decades; liver, kidney, uterine and skin cancers have risen. The most-studied rise is under 50: registries in most high-income countries show increasing incidence for several sites since the 1990s, colorectal cancer the most consistent and the largest, following birth cohort — each generation born after about 1960 has higher rates at a given age than the one before — which is the signature of an exposure rather than of detection. Detection is still part of the story: thyroid cancer's rise is almost entirely diagnostic, and lowered screening ages complicate the recent years.

More is found earlier, by design and by accident. Screening for breast, bowel and cervical cancer is national in most of Europe; low-dose CT screening for lung cancer in smokers is now a programme in England and the United States; blood tests that claim to detect many cancers at once are in a large randomised trial in England, with the first results due. The pandemic worked the other way, and registry data now show later-stage diagnosis for some sites in 2020 and 2021 with a measurable mortality cost.

Treatment and prevention have moved survival more than incidence. Immunotherapy has changed the outlook in melanoma and some lung cancers; HPV vaccination has all but eliminated cervical cancer in the first fully vaccinated cohorts; five-year survival has risen for most sites, with the United Kingdom still behind comparable European countries on the international comparisons.

The deaths say that age-standardised cancer mortality has fallen by around one to two per cent a year in high-income countries for decades, driven by lung cancer, and that the fall continues. The exception is the one that matters for the under-50 question: colorectal mortality under 50 has risen in some countries, which detection alone does not produce. The candidate causes of the cohort effect — diet, obesity, antibiotics, the microbiome — remain plausible and unproven.

Status

building

Direction

mixed

Would change our view

Site-specific mortality trends under 50 across several registries moving together; the multi-cancer detection trial reporting on late-stage incidence or mortality; or a cohort study linking a measured exposure to the birth-cohort pattern.

Trackers

Gaps

Registries with reliable stage at diagnosis outside the Nordic countries, the UK, the US and Australia; whether the early-onset pattern holds in Japan and Korea; survival comparisons on a consistent basis beyond the periodic international studies.