Environment International (Jan 2025)

Serum metabolome associated with novel and legacy per- and polyfluoroalkyl substances exposure and thyroid cancer risk: A multi-module integrated analysis based on machine learning

  • Fei Wang,
  • Yuanxin Lin,
  • Lian Qin,
  • Xiangtai Zeng,
  • Hancheng Jiang,
  • Yanlan Liang,
  • Shifeng Wen,
  • Xiangzhi Li,
  • Shiping Huang,
  • Chunxiang Li,
  • Xiaoyu Luo,
  • Xiaobo Yang

Journal volume & issue
Vol. 195
p. 109203

Abstract

Read online

Background: Exposure to per- and polyfluoroalkyl substances (PFAS) may linked to thyroid cancer (TC) risk, but inconsistent findings and a lack of studies on mixed exposures exist, especially regarding novel PFAS compounds. Additionally, little is known about the potential mechanisms underlying the association. Objectives: Explore the effects of PFAS exposure on the serum metabolome and its correlation with TC. Methods: A 1:1 age- and sex-matched case-control study was administered with 746 TC cases and healthy controls. Liquid chromatography-high resolution mass spectrometry determined serum 11 PFAS and untargeted metabolome profile. ENET and LightGBM models were used to explore the exposure patterns and perform variable selection. The mixed exposure effects were assessed using Weighted quantile sum regression and Bayesian kernel machine regression. Metabolome-wide association analyses were performed to assess metabolic dysregulation associated with PFAS, and a structural synthesis analysis was used to detect latent groups of individuals with TC based on PFAS levels and metabolite patterns. Results: Ten of the 11 PFAS were detected in > 80 % of the population. PFHxA and PFDoA exposure associated with increased TC risk, while PFHxS and PFOA associated with decreased TC risk in single compound models (all P < 0.05). Machine learning algorithms identified PFHxA, PFDoA, PFHxS, PFOA, and PFHpA as the key PFAS influencing the development of TC, and mixed exposures have an overall positive effect on TC risk, with PFHxA making the primary contribution. A novel integrative analysis identified a cluster of TC patients characterized by increased PFHxA, PFDoA, PFHpA and decreased PFOA, PFHxS levels, and altered metabolite patterns highlighted by the upregulation of free fatty acids. Conclusions: PFAS exposure is linked to a higher risk of TC, possibly through changes in fatty acid metabolism. Larger, prospective studies are needed to confirm these findings, and the role of short-chain PFAS requires more attention.

Keywords