Forget that cancer risk is only about the tumor. Forget that mutations are always the main character. Forget that prevention has to mean giving the same drug to a stadium full of people and hoping the math behaves. This new Cell paper asks a sneakier question: what if your blood can reveal when the lungs are turning into a sketchy neighborhood before cancer has officially moved in?
That is the fascinating, slightly anxiety-inducing idea behind Pandya and colleagues’ study on “plasma signals” of lung tumor promotion. The researchers used machine learning on blood plasma proteins to identify a 14-protein signature linked to future lung cancer risk, including signals detectable more than five years before diagnosis [1]. As someone who has stared too long at heatmaps, I must say: when a blood test starts whispering about future lung biology, we should listen carefully, but not immediately start printing clinic flyers.
The Blood Is Not Snitching on a Tumor. It Is Snitching on the Neighborhood.
Most cancer blood tests try to catch evidence from the tumor itself: DNA crumbs, suspicious proteins, tiny molecular receipts. This study is a little different. The 14-protein signature seemed to reflect an inflammatory lung environment that may help tumors begin, rather than simply detecting an existing hidden tumor [1].
That matters because lung cancer is often diagnosed late, when the villain has already had three sequels and a merchandising deal. Low-dose CT screening helps, but it mostly targets older people with a smoking history. That leaves out many people exposed to air pollution, occupational hazards, or other risks, including some never-smokers.
The team trained models using plasma protein data from more than 48,000 UK Biobank participants, then validated the signature across eight cohorts. That last bit is important. A biomarker that works only in the dataset where it was born is not a biomarker. It is a lab pet.
Air Pollution, IL-1β, and the Lung’s Bad Group Chat
The biology connects to earlier work showing that fine particulate air pollution, PM2.5, can promote lung adenocarcinoma in cells that already carry mutations such as EGFR. The proposed mechanism involves inflammatory immune cells releasing interleukin-1 beta, or IL-1β, which nudges vulnerable lung cells toward a more cancer-friendly state [2].
In this new paper, pollution exposure increased the protein signature and expanded “KAC cells,” an injury-associated cell state that mutant lung cells from different lineages appeared to enter on their way toward cancer [1]. Translation: several different cellular troublemakers may be using the same suspicious back alley. Fascinating. There, I said it again. Journal club has ruined me.
This is not “pollution causes cancer in one simple step.” Biology refuses to be that considerate. The idea is more like: aging lungs can accumulate mutant cells, and inflammatory exposures may provide the promotion signal that lets some of those cells wake up, expand, and cause problems. Mutations load the confetti cannon. Inflammation may press the button.
The Canakinumab Plot Twist
The drug angle comes from CANTOS, a cardiovascular trial of canakinumab, an antibody that blocks IL-1β. In 2017, CANTOS reported fewer lung cancers among people receiving canakinumab, especially at higher doses, but the finding was exploratory and the number needed to treat in an unselected population was too high to be practical [3].
Pandya and colleagues reanalyzed CANTOS data and found that people with a high baseline 14-protein signature appeared to benefit much more clearly from IL-1β blockade. The reported number needed to treat dropped from 1516 in an unselected approach to 55 in the high-signature group [1]. That is the difference between “statistically interesting but clinically awkward” and “wait, maybe this prevention thing has a lane.”
Still, cue the limitation slide. Canakinumab can increase serious infection risk, CANTOS was not originally designed as a lung cancer prevention trial, and a biomarker-guided prevention strategy would need prospective testing. Nobody should interpret this as “ask your doctor for an IL-1β blocker because your lungs seem moody.” Please do not make your clinician age visibly during the appointment.
Why This Is Cool, Carefully
Recent proteomics studies have shown that circulating proteins can improve lung cancer risk prediction near diagnosis and may capture biology beyond smoking history alone [4,5]. Reviews of lung cancer screening and interception increasingly argue that the field needs better ways to identify who is at risk, when to intervene, and what to use [6].
This paper pushes that conversation from “can we detect cancer earlier?” toward “can we detect the conditions that help cancer start?” That is a subtle shift, but a big one. If reproduced prospectively, a plasma signature like this could help select people for prevention trials, personalize screening, or identify inflammatory lung states before a scan shows anything dramatic.
The dream is not a magic blood test that declares your destiny like a very expensive fortune cookie. The dream is risk stratification with biology attached: who is vulnerable, what pathway is active, and which prevention strategy might actually make sense.
For now, the responsible summary is: promising, mechanistically rich, not ready for routine care, and absolutely worth following. My inner grad student wants a replication cohort, a prospective trial, calibration curves, subgroup analyses, and possibly a snack. But yes, this is fascinating.
References
-
Pandya T, Zagorulya M, Leung MM, et al. Plasma signals of lung tumor promotion for molecular cancer prevention. Cell. 2026. DOI: 10.1016/j.cell.2026.05.005. PMID: 42242224
-
Hill W, Lim EL, Weeden CE, et al. Lung adenocarcinoma promotion by air pollutants. Nature. 2023;616:159-167. DOI: 10.1038/s41586-023-05874-3. PMID: 37020004
-
Ridker PM, MacFadyen JG, Thuren T, Everett BM, Libby P, Glynn RJ. Effect of interleukin-1β inhibition with canakinumab on incident lung cancer in patients with atherosclerosis. Lancet. 2017;390:1833-1842. DOI: 10.1016/S0140-6736(17)32247-X. PMID: 28855077
-
The Lung Cancer Cohort Consortium. The blood proteome of imminent lung cancer diagnosis. Nature Communications. 2023;14:3042. DOI: 10.1038/s41467-023-37979-8
-
Kaur G, et al. Plasma protein biomarkers for early prediction of lung cancer. EBioMedicine. 2023;94:104686. DOI: 10.1016/j.ebiom.2023.104686. PMID: 37379654
-
Zhang J, Park MD, Pandya T, et al. Innovative approaches for lung cancer screening and interception. Nature Reviews Clinical Oncology. 2026;23:374-389. DOI: 10.1038/s41571-026-01131-4
Disclaimer: The image accompanying this article is for illustrative purposes only and does not depict actual experimental results, data, or biological mechanisms.