Early cancer interception—stopping a tumor before it becomes one—has long been hampered by the inability to identify who is actually on that trajectory. A blood test that reliably flags lung cancer risk half a decade before clinical diagnosis could fundamentally reshape who qualifies for prevention trials, sparing low-risk individuals from toxic interventions while concentrating therapeutic effort where it matters most.
Using machine learning applied to plasma proteomics, researchers identified a 14-protein signature capable of predicting lung cancer more than five years before diagnosis. The signature was validated across eight independent cohorts, lending unusual robustness to the finding. Critically, the panel was not merely elevated in smokers—it also tracked exposure to fine particulate matter (PM2.5), implicating environmental air pollution as a biologically active promoter rather than a passive background risk. The signature's origin was traced to lung myeloid and alveolar cell populations, and in EGFR-driven lung adenocarcinoma, multiple epithelial lineages were found to converge on a specific keratin8+/claudin4+ alveolar transitional cell state. Components of the signature could be induced experimentally by particulate matter, oncogenic EGFR activation, or IL-1β stimulation, while IL-1β inhibition suppressed this transitional cell expansion. Reanalysis of the landmark CANTOS trial—which previously showed canakinumab reduced lung cancer incidence but with a prohibitively high number-needed-to-treat—revealed the signature identifies a subgroup deriving substantially greater benefit, compressing the NNT to a potentially practical threshold.
This work is analytically sophisticated and mechanistically grounded, making it considerably more than an incremental biomarker study. The convergence of proteomics, single-cell transcriptomics, environmental exposure data, and a randomized trial validation is rare. Key limitations remain: the cohorts are largely smokers, the transitional cell state findings stem partly from mouse and organoid models, and prospective interventional validation is still needed. For health-conscious adults, the PM2.5 connection is a sobering reminder that air quality is a lung cancer variable, not merely a respiratory one.