1304Pan-cancer analysis of pre-diagnostic blood metabolite concentrations in the EPIC study. (2nd September 2021)

• Record Type: Journal Article
• Authors: Breeur, Marie; Ferrari, Pietro; Schmidt, Julie; Travis, Ruth; Key, Tim; Johansson, Mattias; Dossus, Laure; Jenab, Mazda; Rinaldi, Sabina; Gunter, Marc; Viallon, Vivian
• Is Part Of: International journal of epidemiology. Volume 50(2021)Supplement 1
• Publication Date: 2021-09-02
• Subjects: Epidemiology -- Periodicals; 614.4
• DOI: 10.1093/ije/dyab168.685
• Languages: English
• ISSNs: 0300-5771

Abstract

Abstract: Background: Metabolomics studies in cancer epidemiology have mostly focused on single metabolite-cancer site associations. Pan-cancer analyses may have larger statistical power when identifying metabolites showing consistent associations across cancer sites, while allowing the identification of site-specific associations. Methods: Data from seven cancer-specific case-control studies nested within the European Prospective Investigation into Cancer and Nutrition Cohort (EPIC) were pooled, resulting in a total sample of 7, 957 case-control pairs from eight cancer types (breast, colorectal, endometrial, gallbladder, kidney, localized prostate and advanced prostate cancer, and hepatocellular carcinoma). A total of 117 pre-diagnostic blood metabolites were measured. After clustering the most highly correlated ones together, we studied the association between 50 features (metabolites or clusters of metabolites) and cancer risk in multivariate penalized conditional logistic regression models controlled for body mass index using the data shared lasso. Results: We identified: (i) 8 features with consistent associations across cancer sites: e.g., glutamine and C4-acylcarnitine, one cluster of lysophosphatidylcholines and one of phosphatidylcholines were inversely associated with cancer, while C10-acylcarnitine, valine and proline showed positive associations; (ii) 11 features with heterogeneous associations across cancer sites: e.g., arginine was positively associated with colorectal cancer only, while one cluster of sphingomyelins was associated inversely with hepatocellular carcinoma and positively with endometrial cancer. Conclusions: Our pan-cancer analysis notably identified metabolites showing consistent associations with cancer risk across different cancer-types. Key messages: Our results could lead to the identification of common pathways shared across different cancer types.