A blood test that predicts lung cancer 5 years early? What we know so far

Lung cancer remains the deadliest cancer worldwide, but a major scientific breakthrough could reshape how the disease is detected and potentially prevented.

Researchers have identified a blood-based protein signature capable of predicting lung cancer risk more than five years before diagnosis, opening the possibility of earlier intervention and improved survival outcomes.

The findings, published in the journal Cell on June 4, stem from a large international effort involving more than 80 scientists across four continents.

By combining advanced machine-learning techniques with extensive health data, researchers discovered a group of proteins in the bloodstream that can signal elevated lung cancer risk years before conventional diagnosis.

More from Health

India’s digital sovereignty depends on balancing AI innovation and regulation

Spare the Pope: The AI writing panic is looking at the wrong problem

The study also uncovered evidence suggesting that targeting a specific inflammatory process linked to these proteins may significantly reduce the likelihood of developing lung cancer in certain high-risk individuals.

Experts caution that additional clinical studies are required before the discovery can be translated into routine patient care.

Nevertheless, researchers and oncologists say the work represents an important advance in the long-standing effort to move beyond treatment and focus on prevention.

“Preventing lung cancer has been a missing holy grail for a very, very long time,” Dr. Douglas Arenberg, a professor of medicine at the University of Michigan who was not involved in the study, told the New York Times.

How scientists discovered the protein signature

The research team, led by scientists from the Francis Crick Institute and University College London, analysed approximately 48,000 blood samples collected through the UK Biobank.

Using machine-learning models, they searched for biological patterns associated with future lung cancer development.

Their analysis revealed a set of 14 proteins that consistently appeared in individuals who later developed lung cancer.

When researchers combined information about these proteins with factors such as age, smoking history and pre-existing lung conditions, they were able to predict future lung cancer risk more accurately than many existing assessment methods.

To verify the findings, the team tested the protein signature across eight additional datasets from different regions of the world.

Importantly, the pattern also appeared among people who had never smoked, including participants from Taiwan, suggesting its potential usefulness beyond traditional high-risk groups.

According to the researchers, the proteins do not directly reflect the presence of a tumour. Instead, they appear to indicate biological changes occurring within lung tissue long before cancer develops.

“The signature doesn’t reflect the tumour itself,” explained Roel Vermeulen, a participating researcher and professor at Utrecht University. “It reflects an inflamed lung environment — an alarm signal that dormant mutant cells are being nudged toward malignancy.”

The link between inflammation, pollution and cancer risk

The study sheds new light on the role inflammation may play in the development of lung cancer.

Through experiments involving cell and mouse models, researchers found that the 14 proteins become elevated when a specific inflammatory pathway is activated.

Factors such as cigarette smoke and air pollution can trigger this pathway, creating an environment that may encourage cancer formation.

The findings support a growing body of evidence suggesting that cancer development involves more than genetic mutations alone. Instead, mutations and chronic inflammation may work together to drive disease progression.

“This adds to the evidence that it isn’t just genetic mutations caused by smoking, pollution or other factors that are driving lung cancers,” researchers noted.

Dr. Charles Swanton, an oncologist and clinical director of the Francis Crick Institute and senior author of the study, said the results indicate that “smoke causes mutations and inflammation, which together cause cancer.”

Researchers also observed that the protein signature was elevated in individuals who later developed chronic obstructive pulmonary disease (COPD) and pulmonary fibrosis.

This suggests that several serious lung diseases may emerge from a shared inflammatory environment.

The study also revealed that multiple types of lung cells appear to enter a vulnerable pre-cancerous state before tumours emerge. Air pollution was found to expand this pool of at-risk cells while simultaneously increasing levels of the protein signature.

In laboratory experiments, blocking the inflammatory pathway reduced these cellular changes and slowed the growth of early tumours.

Could an existing drug help prevent lung cancer?

One of the most intriguing aspects of the research involves canakinumab, an anti-inflammatory drug originally developed for other conditions.

To investigate whether suppressing the identified inflammatory pathway could lower cancer risk, researchers revisited data from a previous randomised clinical trial involving 4,650 patients.

Although that study primarily focused on cardiovascular outcomes, it unexpectedly showed a reduction in lung cancer cases among participants receiving the drug.

The new analysis found that among roughly 2,300 individuals with above-average levels of the 14-protein signature, canakinumab reduced lung cancer risk by nearly 50 per cent.

Swanton compared the concept to cholesterol management in heart disease. “This is sort of equivalent to an LDL for cancer,” Swanton said.

However, researchers emphasise that the findings are not yet sufficient to recommend the drug as a preventive therapy. Dedicated clinical trials are still required to determine whether the treatment can safely and effectively prevent lung cancer.

“This is the big ‘if,’” Dr. Roy S. Herbst, chief of medical oncology and hematology at the Yale School of Medicine told the New York Times. “Will it be clinically significant? Will we be able to block this sufficiently at the right stage to prevent cancer?”

Medical experts also point out that canakinumab carries potential risks. According to Dr. Peter Mazzone of the Cleveland Clinic, the drug can increase susceptibility to infections and sepsis, raising questions about whether its benefits would outweigh potential harms in healthy individuals.

Researchers say future studies may identify alternative medicines that target the same inflammatory pathway with fewer side effects.

What this could mean for future lung cancer screening

Beyond prevention, the protein signature may help improve how doctors identify people who would benefit most from screening programmes.

Current screening recommendations in countries such as the United States focus largely on age and smoking history. Yet many eligible individuals never undergo screening, while some people who do not meet traditional criteria still develop lung cancer.

Scientists believe a blood test based on the newly identified proteins could help address both challenges.

It could identify individuals at particularly high risk and encourage earlier imaging through low-dose CT scans. It may also help uncover elevated risk among non-smokers, a growing concern in global lung cancer research.

“In particular, there is ‘a big need’ to better detect lung cancers in people who never smoked,” Mazzone said.

Although more evidence is needed before the approach can be adopted clinically, researchers see the findings as an important step toward more personalised screening strategies.

How AI is changing lung cancer detection

Machine learning was central to identifying the 14-protein signature, demonstrating how AI can uncover disease patterns hidden within massive datasets.

AI-powered tools are increasingly being used alongside blood testing, medical imaging and other diagnostic technologies to improve cancer detection.

Several emerging technologies illustrate this trend. Guardant Health's Shield blood test is being developed as a multi-cancer screening tool and has shown promise for lung cancer applications.

Meanwhile, researchers at Westlake University in China have created a portable device capable of analysing a single drop of blood with high sensitivity for early disease detection.

Artificial intelligence is also being integrated into non-invasive diagnostic approaches. Breath Diagnostics is developing its OneBreath test, which uses machine learning to analyse compounds in exhaled breath and identify patterns associated with lung cancer.

In medical imaging, AI systems developed by companies such as Qure.ai and Median Technologies are helping radiologists detect abnormalities in chest X-rays and low-dose CT scans more efficiently and accurately.

These developments suggest that future lung cancer screening may combine blood tests, imaging technologies and AI-driven analysis to identify disease at much earlier stages.

For many cancer specialists, that possibility represents the ultimate goal.

Herbst reflected on how dramatically lung cancer care has evolved during his career, noting that some patients can now be cured. Yet he stressed that the greatest gains would come from detecting the disease before it takes hold.

“The greatest benefit is still going to be getting it at the earliest stages, or even preventing it,” he said. “This is a step forward in that direction.”

With inputs from agencies