Abstract 3432: Blood metabolomic profile of glioma risk: A pooled, multi-cohort analysis in COMETS

Abstract

Abstract Although gliomas are relatively rare, they are among the most lethal malignancies. Yet, with the exception of ionizing radiation and family history, their etiologic factors remain largely unknown. We conducted an untargeted metabolomics investigation pooling nested case-control studies from 9 prospective cohorts (median age 63 years, 59% women) in the Consortium of Metabolomics Studies (COMETS). One cancer-free control was individually matched to each of the 803 glioma cases based on age, sex, self-identified race/ethnicity, and blood collection date. Using an ultrahigh-performance LC-MS/MS platform (Metabolon, Inc.), we identified 874 known metabolites in pre-diagnostic serum or plasma, and logistic regression models conditioned on the matching factors) estimated odds ratios (ORs) and 95% confidence intervals for the associations between 1-SD increases in circulating metabolites and glioma risk (Bonferroni corrected significance threshold P=5.7 × 10−5). A total of 58 metabolites were associated with glioma (P<0.05), with the top three being 3-methylglutarylcarnitine (OR=1.19; P=0.001), cholate (OR=0.85; P=0.001) and a sphingomyelin (SM) (d18:2/21:0, d16:2/23:0, OR=1.19; P=0.002). Analysis restricted to cases diagnosed >5 years after blood collection showed 51 metabolites associated with risk (P<0.05), including SM d18:2/23:1 (OR=1.32; P=5.2 × 10−5), and two other SMs (d18:2/21:0, d16:2/23:0 and d17:1/24:1(15Z)) (ORs=1.27 and 1.26, respectively; P<5.6 × 10−4). Our findings were not materially altered with further adjustment for smoking, BMI, height and diabetes. Analysis restricted to high-grade glioma cases (n=595) showed similar associations, whereas low-grade glioma (n=207) showed fewer metabolites related to risk (top signals of sphingosine, sphinganine and L-glutamic acid, each inversely associated). We did, however, observe that the 2-hydroxyglutarate/alpha-ketoglutarate ratio was significantly associated with low-grade disease (OR=2.9; P=0.0008), which was most pronounced among cases diagnosed >10 years following blood collection (OR=9.99; P=0.003). Glioblastoma (n=485) showed similar positive risk estimates including for 3-methylglutarylcarnitine which achieved Bonferroni significance (OR=1.35; P=3 × 10−5) and several SMs. In contrast, the top metabolite for astrocytoma risk (n=88) was tyrosine (OR=0.55; P=0.002). Gene-set and principal components analyses confirmed that the sub-pathways of SM, dihydro-SM, fatty acid, and dihydroceramide metabolism were significantly associated with overall glioma risk, and dihydro-SM, dihydroceramide, SM, lactosylceramide, glutathione, and purine metabolism were associated with high-grade glioma (all P values <0.05). These novel findings regarding altered metabolic pathways/dysregulation occurring years before glioma diagnoses may be related to etiologic factors or subclinical disease. Citation Format: Jiaqi Huang, Bin Zhao, Stephanie J. Weinstein, Marian L. Neuhouser, Steve C. Moore, Heather Eliassen, David J. Cote, Loic Le Marchand, Lorelei Mucci, Ying Yang, Victoria Stevens, Wei Zheng, Xiao-Ou Shu, Demetrius Albanes. Blood metabolomic profile of glioma risk: A pooled, multi-cohort analysis in COMETS [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3432.