Associations of prediagnostic metabolomic profiles with colorectal cancer risk-a collaborative reanalysis within the COnsortium of METabolomic Studies (COMETS).

Abstract Background: An improved understanding of the etiology of colorectal cancer is needed to allow more accurate disease prediction. This study will use a state-of-the-art metabolomics strategy leveraging existing data from 8 prospective cohort studies to discover and validate novel blood-based biomarkers that are associated with colorectal cancer risk. This project is in its early phase and the abstract is focusing on describing the study design. Methods: We will perform a meta-analysis using data from 8 large-scale prospective cohort studies that are part of the Consortium of Metabolomics Studies (COMETS): European Prospective Investigation into Cancer and Nutrition, Estonian Biobank, Women Health Study – Observational Study, Prostate, Lung, Colorectal and Ovarian Cancer Observational Study, Nurses’ Health Study, Health Professional Follow-up Study, Shanghai Women’s and Men’s Health Studies. All studies used a nested case-control study design (1:1 matching at a minimum by age and sex). Prediagnostic plasma and serum samples were analyzed using state-of-the art metabolomics platforms: Metabolon, Inc., Broad Institute and Biocrates. We expect n~150 overlapping identified metabolites across platforms. Data will be split into discovery and validation set. We will investigate associations of metabolites with colorectal cancer risk using multivariable conditional logistic regression models and random effects meta-analysis. Analyses will be stratified by age (<45, 45-55, ≥55 years) and site (colon versus rectal). Heterogeneity between studies will be assessed using Q-value and I2 statistics. Metabolites at the 0.05 false discovery rate adjusted significance levels will be carried forward to the validation set. Results: The present study includes n=2,093 case-control pairs in the discovery dataset and n=1,082 case-control pairs in the validation dataset. Demographic data such as age at blood draw and age at diagnosis, race, body mass index, smoking, and dietary intake at baseline (red meat, vegetables, fruits, fat, and fiber) are collected. Data on tumor site are >95% complete. Tumor stage is available for >85% of patients and is expected to increase substantially since medical data abstraction is under way in some cohorts. Conclusions: This is the largest study, to our knowledge, investigating the associations of prediagnosis metabolomic profiles with colorectal cancer risk. The proposed analyses have the potential to identify biomarkers that may enhance population-wide and tailored screening of colorectal cancer risk. Citation Format: Jennifer Ose, Cornelia M. Ulrich, Mary Playdon, Ken Boucher, A. Heather Elisassen, Xia-Ou Shu, Wei Zheng, Andres Metspalu, Tonu Esku, Jessica Lasky-Su, Steve Moore, Marc Gunter. Associations of prediagnostic metabolomic profiles with colorectal cancer risk—a collaborative reanalysis within the COnsortium of METabolomic Studies (COMETS) [abstract]. In: Proceedings of the AACR Special Conference on Modernizing Population Sciences in the Digital Age; 2019 Feb 19-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(9 Suppl):Abstract nr A06.