Abstract 6811: Plasma proteomic signature for persistent smoking effects and its association with risk of smoking-related cancers among former smokers

Abstract Background: Previous studies have identified circulating protein signatures associated with smoking, which were associated with cancer risk and reduced with years of smoking cessation. A protein signature for persistent smoking effects after quitting smoking is not available. Our study aims to fill this research gap. Methods: Included in the study are 47,166 UK Biobank participants, aged from 39 to 70 years at baseline without any cancer history, with proteomics and complete smoking history. A total of 2,911 plasma proteins with less than 20% missing values were analyzed. Linear regression analyses were conducted to identify smoking-related proteins among current (N=4,036) and never smokers (N=32,664), adjusting for age, sex, race, alcohol consumption, body mass index, and Townsend deprivation index. To identify protein biomarkers capturing persistent smoking effects, we analyzed smoking-related proteins among former (N=10,466) and never smokers (N=32,664) using similar linear regression analyses. Proteins with significant associations with former smoking status at FDR-corrected p-value 0.05 were retained for further analysis. We then applied elastic net regression to the regression residuals of proteins to develop a protein signature for persistent smoking effect (PSPSE), using a 70% training and 30% testing sets, which were randomly split between former and never smokers. The PSPSE was derived by a weighted sum of the selected proteins using their coefficients from the elastic net analysis. We used Cox proportional hazard regressions to examine associations between the PSPSE and risk of 12 smoking-related cancers (SRC), including lung cancer (LC) and upper aerodigestive tract cancer (UATC). Results: Of 2,911 proteins tested, 2081 proteins were associated with current smoking status at FDR p0.05. Among them, 1,039 proteins were associated with former smoking status at FDR p0.05. Of them, 242 proteins were selected to develop the PSPSE among 10,466 former smokers. The PSPSE achieved an area under the curve (AUC) of 0.73 in the training set and 0.70 in the testing set. The PSPSE was significantly associated with the elevated risk of SRC (Hazard ratio (HR)=1.33, 95%CI: 1.17-1.51), LC (HR=1.94, 95%CI: 1.52-2.48), and UATC (HR=1.48, 95%CI: 1.06-2.07) independently of smoking pack-year and quitting duration among former smokers. The PSPSE was not related to an increased risk of non-smoking-related cancers. Conclusions: We developed a proteomics signature capturing the persistent smoking effect among former smokers, which was associated with elevated risks of SRC, particularly LC. If validated, the PSPSE can be used to identify former smokers at high risk of developing cancer for close surveillance and primary prevention. Updated results will be presented at the meeting. Citation Format: Duc Huy Le, Qiuyin Cai, Jirong Long, Sang M. Nguyen, Wei Zheng, Xiao-Ou Shu, . Plasma proteomic signature for persistent smoking effects and its association with risk of smoking-related cancers among former smokers abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6811.