Researchers from University College London (UCL) and the Francis Crick Institute identified 14 proteins in the blood that allow predict the risk of developing lung cancer more than five years prior to diagnosis. The discovery, published in the magazine Cellopens the door to preventive treatments in people at high risk before the disease takes hold.
The study analyzed blood samples from more than 48,000 people from the UK Biobank using machine learning. The team combined variables such as age, smoking habit and history of lung diseases with the levels of these 14 proteins to build a biological signature capable of anticipating a future diagnosis, as detailed by the UCL in a statement.
The results were validated in eight data sets from different parts of the world, with the collaboration of more than 80 researchers on five continents.
A central fact of the work is that this protein signature does not come from the tumor itself. According to the analysis, it reflects a inflammatory environment in the lung that precedes cancer. Factors such as air pollution or cigarette smoke would be responsible for this previous inflammatory state.
Inflammation as an early warning sign
The team determined that the environmental pollution It not only raises the levels of the protein signature but also increases the number of abnormal lung cells associated with the onset of cancer. This link reinforces previous research from the same group, which had already shown for the first time how air pollution can cause lung cancer in people who never smoked.
The study – funded by Cancer Research UK and supported by the National Institute for Health and Care Research – suggests that drugs that block the action of an inflammatory molecule called IL-1β could prevent lung cancer in those who present the high-risk signature. It is an immune system molecule whose sustained activity contributes to tissue damage and tumor development.
The work is part of the TRACERx programme, a reference study on the evolution of lung cancer led by Professor Charles Swanton, head of Personalized Oncology Medicine at the UCL Cancer Institute and director of Clinical Research at the Crick.
Swanton drew an analogy to cardiology to illustrate the magnitude of the challenge: «Drugs like statins have transformed cardiovascular disease prevention, used to treat individuals with high levels of LDL cholesterol. But we still don’t have a similar risk marker or a statin for lung cancer,» he explained.
The researcher also highlighted the clinical urgency behind the finding. «In the clinic we see firsthand the impact of diagnosing late-stage lung cancer, so being able to identify people at higher risk and intervene before the disease develops is critical,» Swanton said.
Tej Pandya, clinical PhD student at the UCL Institute of Health Informatics and visiting scientist at the Crick, described the scope of the validation process. «We used machine learning on plasma data from more than 48,000 people to identify the signature of 14 proteins, and it was incredible to validate it in eight data sets with more than 80 collaborators on five continents,» he said.
Pandya also spelled out the broader significance of the result. «We have shown that the signature reflects an altered inflammatory environment in the lung before cancer develops. It is a proof of concept that, one day, we could use this signature to offer preventive treatment to people at risk of lung cancer,» he said.
Swanton also raised a hypothesis that goes beyond lung cancer in particular. «This work supports a relatively new idea in the field: that some common age-related diseases, which generate a high disease burden in the community, share a presymptomatic state of inflammation. We believe the signature could help in the future to predict and prevent lung cancer and other lung diseases,» he said.
