Plasma proteomics markers for risk prediction of incident lung cancer
Principal Investigator
Name
Charles Swanton
Degrees
MBCHB Ph.D
Institution
Francis Crick Institute
Position Title
Professor
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
2025-0069
Initial CDAS Request Approval
Feb 19, 2026
Title
Plasma proteomics markers for risk prediction of incident lung cancer
Summary
Lung cancer screening programs in the USA, UK, and Europe are currently limited to individuals above a certain age with significant smoking histories. However, the low positive predictive value of these inclusion criteria limits their utility in selecting participants for lung cancer prevention trials, which typically require large cohorts and long follow-up periods to achieve statistical power.
A detailed understanding of the mechanisms underlying lung cancer initiation is essential for accurately identifying at-risk individuals and developing molecularly targeted prevention strategies. We hypothesize that stratifying individuals who may benefit from anti-IL-1β therapy as a preventive intervention will be possible by understanding how inhaled environmental particulate matter, whether from pollution or tobacco smoke, drives pulmonary inflammation and influences the fate of early oncogene-mutant progenitor cells.
To support this approach, we identified and validated a 14 protein plasma-based signature that predicts incident lung cancer 2–5 years before diagnosis. After training on UK Biobank data, we validated the signature in eight independent cohorts (2,198 cases, 53,641 controls), as well as in a mouse model , which links the signature to the underlying inflammatory biology of tumour promotion, where the proteins are linked to epithelial dysregulation in the alveolar niche, in particular AT2 cells.
However, because anti-IL-1β has demonstrated efficacy only in preventing lung cancer and not treating established disease, it is critical to define protein expression thresholds for identifying high-risk individuals who do not yet have lung cancer.
To achieve this, we aim to use samples from the PLCO cohort alongside TRACERx, a clinical study of patients with lung cancer, to analyse the temporal dynamics of the 14-protein panel. Using absolute quantification, we will determine threshold levels that enable effective stratification for future prevention trials.
Aims
We will use absolute quantification of the 14-protein panel in both the PLCO and TRACERx cohorts to investigate associations with lung cancer risk and to define a minimal set of proteins suitable for risk stratification in prevention trials.
We aim to:
(1) identify a parsimonious subset of proteins using penalized regression (e.g., LASSO), optimized for predictive performance and practicality; and
(2) determine absolute protein thresholds that distinguish high-risk individuals in the pre-diagnostic setting. This combined analysis will enable robust calibration of protein-based risk scores and support their application in future precision-prevention trials, particularly those targeting IL-1β pathways.
This combined analysis will enable robust calibration of protein-based risk scores and support their application in future precision-prevention trials, particularly those targeting IL-1β pathways.
Collaborators
Charles Swanton (Francis Crick Institute)
Tej Pandya (Francis Crick Institute)
Robert Samstein (Icahn School of Medicine at Mount Sinai)
Miriam Merad (Icahn School of Medicine at Mount Sinai)