Plasma-based lipidomics reveals potential diagnostic biomarkers for esophageal squamous cell carcinoma: a retrospective study

PeerJ. 2024 Apr 29:12:e17272. doi: 10.7717/peerj.17272. eCollection 2024.

Abstract

Background: Esophageal squamous cell carcinoma (ESCC) is highly prevalent and has a high mortality rate. Traditional diagnostic methods, such as imaging examinations and blood tumor marker tests, are not effective in accurately diagnosing ESCC due to their low sensitivity and specificity. Esophageal endoscopic biopsy, which is considered as the gold standard, is not suitable for screening due to its invasiveness and high cost. Therefore, this study aimed to develop a convenient and low-cost diagnostic method for ESCC using plasma-based lipidomics analysis combined with machine learning (ML) algorithms.

Methods: Plasma samples from a total of 40 ESCC patients and 31 healthy controls were used for lipidomics study. Untargeted lipidomics analysis was conducted through liquid chromatography-mass spectrometry (LC-MS) analysis. Differentially expressed lipid features were filtered based on multivariate and univariate analysis, and lipid annotation was performed using MS-DIAL software.

Results: A total of 99 differential lipids were identified, with 15 up-regulated lipids and 84 down-regulated lipids, suggesting their potential as diagnostic targets for ESCC. In the single-lipid plasma-based diagnostic model, nine specific lipids (FA 15:4, FA 27:1, FA 28:7, FA 28:0, FA 36:0, FA 39:0, FA 42:0, FA 44:0, and DG 37:7) exhibited excellent diagnostic performance, with an area under the curve (AUC) exceeding 0.99. Furthermore, multiple lipid-based ML models also demonstrated comparable diagnostic ability for ESCC. These findings indicate plasma lipids as a promising diagnostic approach for ESCC.

Keywords: Esophageal squamous cell carcinoma; Lipidomics; Machine learning; Plasma-based diagnostic model.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Biomarkers, Tumor* / blood
  • Case-Control Studies
  • Chromatography, Liquid
  • Esophageal Neoplasms* / blood
  • Esophageal Neoplasms* / diagnosis
  • Esophageal Squamous Cell Carcinoma* / blood
  • Esophageal Squamous Cell Carcinoma* / diagnosis
  • Female
  • Humans
  • Lipidomics* / methods
  • Lipids / blood
  • Machine Learning
  • Male
  • Middle Aged
  • Retrospective Studies

Substances

  • Biomarkers, Tumor
  • Lipids

Grants and funding

This research was supported by the National Natural Science Foundation of China (No. 81672315, 81302840); the Medical and the Health Science Project of Zhejiang Province (2022KY622&2020KY487); the Zhejiang Provincial Natural Science Foundation of China (LY23H010002); and the Key R&D Program Projects in Zhejiang Province (2018C04009). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.