NELSON: volume-based CT screening reduces lung cancer deaths with fewer false alarms
In a Dutch-Belgian trial, CT screening using nodule volume rather than diameter cut lung cancer deaths in men by about a quarter at 10 years, with a far lower false-positive rate than NLST.
NELSON randomised 15,789 current or former smokers (13,195 men and 2,594 women) aged 50-74 to CT screening at baseline, 1, 3 and 5.5 years or to no screening. Nodules were managed by volume and volume-doubling time, with an indeterminate category that triggered a repeat scan rather than a biopsy.
The primary analysis in men showed a lung cancer mortality rate ratio of 0.76 at 10 years. Screen-detected cancers were much more often stage I. The positive-screen rate was around 2%, far below NLST's, because the indeterminate category absorbed most small nodules.
NELSON confirmed NLST against a no-screening control and gave European programmes a workable protocol.
- Lung cancer mortality rate ratio in men 0.76 (95% CI 0.61-0.94) at 10 years
- In the smaller female cohort the rate ratio was 0.67 (95% CI 0.38-1.14), not statistically significant but consistent with benefit
- Overall positive-screen rate about 2%, with the majority of screen-detected cancers at stage IA-IB
- Comparator was no screening, removing the concern that chest X-ray in NLST masked part of the effect
Lung screening works when it uses volumetric nodule management, and it works against a no-screening control. The protocol underpins the UK Targeted Lung Health Check programme and European recommendations. Benefit in women remains less precisely estimated.
- Women were under-represented, so the female estimate is imprecise
- Screening intervals lengthened to 2.5 years in the final round; the optimal interval is still debated
- Some overdiagnosis is likely; the authors estimated a modest excess of cancers in the screened arm at 10 years
- Participants were volunteers responding to a population mailing, which may not reflect routine uptake