Generated by Codex with GPT-5

This summary covers The Economist’s June 13th, 2026 Science & technology article listed in the contents as New test for lung cancer and published under the headline Predict and prevent.

Cancer prevention has usually meant reducing exposure to known risks, such as smoking and air pollution. A new study points toward a more targeted approach: use a blood test to identify people whose biology suggests that lung cancer is beginning to develop, then give them an existing anti-inflammatory drug before a tumour forms.

An international team analysed blood samples from nearly 50,000 participants in the UK Biobank. Machine learning identified a pattern involving 14 proteins whose levels rose at least five years before a lung-cancer diagnosis. The same molecular signature predicted risk across eight other data sets, including one from Taiwan where few participants smoked. That matters because it suggests the signal is not merely detecting the effects of tobacco.

From prediction to prevention

The proteins appear to reflect an inflammatory process already linked to lung cancer. Air pollution can trigger the release of interleukin-1 beta, or IL-1 beta, a signalling molecule that activates dormant cancer-causing mutations in lung cells. Those cells can then multiply and form tumours. Earlier experiments found that blocking IL-1 beta prevented tumours in pollution-exposed mice.

Drugs that block this molecule already exist. They are prescribed for some inflammatory conditions, including forms of arthritis. One of them, canakinumab, was tested by Novartis in 2017 as a way to prevent heart attacks. It failed at that task, but trial data unexpectedly showed fewer lung cancers among patients who received it.

The new protein signature helps explain why that overall benefit was modest. Treating everyone indiscriminately would be inefficient: among people with low levels of the 14 proteins, more than 1,500 would need canakinumab treatment to prevent one case of lung cancer. Among those with high levels, however, the drug nearly halved the risk, and treating 55 people prevented one case. That prevention rate is comparable to the use of statins against heart attacks.

A promising foundation, not a finished test

The immediate goal is to turn the 14-protein signature into a commercial blood test. Researchers also want to determine whether other anti-inflammatory medicines can provide the same protection. More UK Biobank data expected over the next two years may reveal comparable early-warning patterns for other cancers.

The research does not yet establish a routine screening programme or prove that canakinumab should be widely prescribed. A useful preventive strategy would still need prospective trials, a practical test and a careful balance between drug risks, costs and benefits. But the study supplies the essential pieces: a measurable biological warning, a plausible mechanism and an existing medicine that appears most effective in the people carrying that warning.

The broader shift is significant. Rather than waiting for scans to reveal a tumour, doctors may eventually be able to detect the molecular conditions that allow cancer to emerge and intervene years earlier. Cancer prevention could become a precise medical treatment, not only a matter of healthier behaviour.