Objective:

The article examines CEACAM5 and CEACAM6 as potential lung cancer risk biomarkers in people who never smoked and discusses the need for risk-based screening models for this population.

Approach:

• Biomarker discovery: Zahed and colleagues analyzed 2,821 proteins in prediagnostic blood samples from 180 matched case-control pairs of never-smokers, collected up to 5 years before diagnosis.
• Replication: Associations were assessed in 49,291 UK Biobank participants, regardless of smoking status.
• Screening context: The article reviews never-smoker screening programs, risk-model development, and other candidate biomarkers.

Key Findings:

• Each SD increase in CEACAM5 and CEACAM6 concentration was associated with 65% and 72% higher lung cancer risk within 5 years, respectively, in the discovery analysis.
• In UK Biobank, corresponding risk increases were 86% for CEACAM5 and 92% for CEACAM6; the proteins were weakly correlated and provided largely independent risk information.
• Only 4 of 36 previously identified markers of imminent smoking-related lung cancer were associated with lung cancer in PWNS, indicating that smoking-derived risk models may not transfer to this population.
• TALENT detected lung cancer in 2.6% of participants, and a subsequent Taiwanese national program reported a 1.7% detection rate; both selected participants with additional risk factors, and 85% of tumors were early-stage.
• No comparable validated risk instrument currently exists for PWNS, and the other biomarker approaches described have not been validated in this population.

Interpretation:

The article presents CEACAM5 and CEACAM6 as promising risk-associated biomarkers, while emphasizing that screening PWNS requires risk-based selection and a validated model incorporating biomarkers and relevant exposures.

Limitations:

• The discovery analysis included 180 matched case-control pairs.
• No comparable risk instrument currently exists for people who never smoked.
• The Taiwanese screening programs selected participants with additional risk factors, so their results do not represent unselected never-smoker populations.
• The source describes other biomarker categories under study but does not establish their validation for risk estimation in people who never smoked.

Conclusion:

Risk-model development for people who never smoked will require incorporating biomarkers and relevant exposures and calibrating estimated risk against observed incidence.

Sources:

• Article context provided