Several blood were associated with future lung cancer risk among people with smoking history.
Evidence
This analysis of four prospective cohorts with prediagnostic blood samples included 789 lung cancer cases and 691 controls and compared age-adjusted DNAm clocks, smoking measures, and PLCO 2012 risk discrimination.
Caveat
The study is observational and restricted to participants with smoking history, and smoking explained only part of the methylation-clock associations.
Simplified
BACKGROUND: Biological age, estimated by DNA methylation-based (DNAm) clocks, has been reported to be associated with lung cancer risk. However, the extent to which tobacco smoking behaviours can explain this association and the extent to which DNAm clocks and their components can inform risk assessment for lung cancer remains to be elucidated. This study aimed to evaluate the relationship between DNAm clocks, smoking, and lung cancer risk.
METHODS: We analyzed four prospective cohorts (MCCS, Australia, 324 cases/324 controls; NSHDS, Sweden, 190 cases/190 controls; EPIC, Italy, 160 cases/107 controls; and NOWAC, Norway, 115 case/70 controls) with blood samples collected before lung cancer diagnosis. Study participants were restricted to those with a history of smoking. Incidence sampling was used to match one control to each of the lung cancer cases by cohort, sex, date of blood collection, age, and smoking status in MCCS and NSHDS. The risk discriminative performance of age-adjusted DNAm clocks and their components was compared with that of the Prostate, Lung, Colorectal, and Ovarian model 2012 (PLCO) lung cancer risk model. m2012
RESULTS: We found several DNAm clocks positively associated with lung cancer risk (Hannum: = 1.13, 95% CI = 1.02-1.26; PhenoAge: OR = 1.25, 95% CI = 1.12-1.40; DunedinPACE: OR = 1.44, 95% CI = 1.29-1.62; PCGrimAge (a principal component-denoised GrimAge): OR = 1.79, 95% CI = 1.56-2.06), after adjustment for age and tobacco smoking. Tobacco smoking explained a modest proportion of variance in most age-adjusted DNAm clocks (R < 11%), except for PCGrimAge, where it accounted for ~ 30% of variance in both lung cancer cases and controls. Detailed smoking adjustments attenuated the PCGrimAge association with lung cancer risk by 13%. In a secondary analysis adjusting for PCGrimAge components and the PLCOscore, DNA methylation-predicted packyears emerged as an independent predictor of lung cancer risk (OR = 2.23, 95% CI = 1.58-3.14). The area under the receiver operating characteristic curve (AUC) for the PLCOmodel was 0.66 (95% CI = 0.61-0.71) compared with 0.72 (95% CI = 0.67-0.77) for the PCGrimAge model (P = 0.03). Combining PCGrimAge with PLCOprovided similar risk discrimination as PCGrimAge alone (AUC = 0.72, 95% CI = 0.67-0.77). 2m2012m2012difference m2012
CONCLUSIONS: Methylation-based biological clocks capture epigenetic marks left by exposure to tobacco smoke, and some clocks may inform lung cancer risk assessment by complementing or replacing traditional prediction models.
Key numbers
1.79
PCGrimAge Association with Lung Cancer Risk
for lung cancer risk per standard deviation increase in PCGrimAge
30%
Variance Explained by Tobacco Smoking
Proportion of variance in PCGrimAge explained by tobacco smoking
0.72
AUC for PCGrimAge Model
Area under the curve for PCGrimAge model in lung cancer risk prediction
Full Text
We can’t show the full text here under this license.
