Model using blood tests better predicts lung cancer risk
In 3,195 Danish people who had smoked, a score measuring predictive performance was 0.783 for the model versus 0.720 for an established model.
6 years
An observational study in Danish people who had ever smoked developed and validated a model of six-year lung cancer risk. Researchers used clinical information and blood tests from 689 lung cancer cases and 1,924 controls to train the Copenhagen lung cancer risk model. They tested it in a separate cohort of 3,195 people aged 50 or older, including 68 who developed lung cancer within six years.
In that validation cohort, the model better distinguished people who developed lung cancer from those who did not than an established risk model. The researchers concluded that adding blood test information made prediction more sensitive in the Danish population: it identified a larger share of future lung cancers, particularly among women and younger people.
Why it matters
For older adults who have smoked, risk assessment informs decisions about lung cancer screening. The study addressed whether blood test information could help select people for imaging-based screening.
Caveats
This observational study tested risk prediction, not whether using the model for screening reduced deaths. The validation cohort consisted of Danish people aged 50 or older who had ever smoked.
The paper
Herlev and Gentofte Hospital
Journal of Clinical Oncology, 12 Oct 2026
