Metabolome profiling by untargeted metabolomics and biomarker panel selection using machine-learning for patients in different stages of peripheral neuropathy induced by oxaliplatin

Q: What was the study design and patient cohort size?

The study used a prospective cohort of 129 colorectal cancer patients receiving oxaliplatin-based chemotherapy, stratified into four OIPN severity grades (Level 0-3).

Q: How were metabolites measured and how many were detected?

Serum samples were analyzed using untargeted LC-MS/MS metabolomics, detecting 521 metabolites for downstream modeling.

Q: Which machine-learning approach was used to select biomarkers, and what key discrimination result was reported?

SHAP-guided random forest with multivariate statistical analysis identified six biomarkers. Models validated across six classifiers showed near-perfect discrimination for early-stage OIPN (AUC nearly 1).

Aug 5, 2026

Upload by Clementine

Oxaliplatin-induced peripheral neuropathy (OIPN) threatens colorectal cancer treatment continuity and can lead to dose interruptions and eventual therapy failure. This longitudinal untargeted metabolomics study profiles serum metabolites in a prospective cohort of 129 patients across OIPN severity levels and applies multivariate statistics with SHAP-guided random forest models to prioritize biomarkers. Cumulative oxaliplatin dose, tumor markers, immune inflammation signals, and metabolic/liver dysfunction markers are associated with progression. A six-biomarker panel discriminates early-stage OIPN with near-perfect accuracy and supports prediction and management.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Introduction

Background and clinical challenge

Methods

Prospective cohort and OIPN grading

Untargeted LC-MS/MS metabolomics

Machine-learning biomarker prioritization and validation

Results

Associations with dose, tumor markers, inflammation, and metabolic/liver function

Six-biomarker panel discrimination

Performance for early vs intermediate OIPN grades

Pathway enrichment findings

Conclusions

TYPE Original Research PUBLISHED 19 September 2025 DOI 10.3389/fonc.2025.1617207 OPEN ACCESS EDITED BY Rocco Ricciardi, Massachusetts General Hospital and Harvard Medical School, United States REVIEWED BY Dong-Joo (Ellen) Cheon, Albany Medical College, United States Jinping Gu, Zhejiang University of Technology, China *CORRESPONDENCE Jing-hua Chen [chenjinghua@jiangnan.edu.cn](chenjinghua@jiangnan.edu.cn)[ ](chenjinghua@jiangnan.edu.cn)Yong-juan Ding [dingyongjuan2021@163.com](dingyongjuan2021@163.com)[ ](dingyongjuan2021@163.com)Yan-yan Chen [chenyanyan59@163.com](chenyanyan59@163.com) †These authors have contributed equally to this work and share ﬁrst authorship RECEIVED 24 April 2025 ACCEPTED 01 September 2025 PUBLISHED 19 September 2025 CITATION Hua Y-j, Zhang Y, Wu R-R, Lv J, Zhang Y, Chen Y-y, Ding Y-j and Chen J-h (2025) Metabolome proﬁling by untargeted metabolomics and biomarker panelselection using machine-learning for patients in different stages of peripheral neuropathy induced by oxaliplatin. Front. Oncol. 15:1617207 . doi: 10.3389/fonc.2025.1617207 COPYRIGHT © 2025 Hua, Zhang, Wu, Lv, Zhang, Chen, Ding and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Metabolome proﬁling by untargeted metabolomicsand biomarker panel selection using machine-learning for patients in different stages of peripheral neuropathy induced by oxaliplatin Yu-jiao Hua 1,2,3†, Ying Zhang 4†, Rui-Rong Wu 4†, Juan Lv 1, Yan Zhang 1, Yan-yan Chen 5*, Yong-juan Ding 3* and Jing-hua Chen 1,2* 1School of Life Sciences and Health Engineering, Jiangnan University, Wuxi, Jiangsu, China, 2School of Chemical & Material Engineering, Jiangnan University, Wuxi, Jiangsu, China, 3 Department of Clinical Pharmacy, Afﬁliated Hospital of Jiangnan University, Wuxi, Jiangsu, China, 4 Department of Medical Oncology, Afﬁliated Hospital of Jiangnan University, Wuxi, Jiangsu, China, 5Cancer Institute, Institute of Integrated Chinese and Western Medicine, Afﬁliated Hospital of Jiangnan University, Wuxi, Jiangsu, China Background: Oxaliplatin-induced peripheral neuropathy (OIPN) poses asigniﬁcant challenge for patients with colorectal tumor, often resulting in treatment interruption or discontinuation and subsequent treatment failure. Herein, a longitudinal untargeted metabolomic study to reveal the metabolomic proﬁles and biomarkers associated with the progression of OIPN. Methods: A prospective cohort of 129 colorectal cancer patients receiving oxaliplatin-based chemotherapy was stratiﬁed into four OIPN severity grades (Level 0-3) . Serum samples underwent untargeted LC-MS/MS metabolomic analysis, detecting 521 metabolites. Multivariate statistical models and SHAPguided random forest algorithms were employed to prioritize biomarkers. Machine learning validation included six classiﬁers assessed via ROC-AUC. Results: The cumulative dose of Oxaliplatin chemotherapy plays an important role in OIPN. At the same time, our ﬁndings implied that the occurrence of OIPN may be associated with the progression of the disease and the patients ’ tumor markers (CEA, CA19-9, CA72-4), as well as immune response and inﬂammation (ANC, PLT), and metabolic and liver function abnormalities (GGT and UA)(P<0.05).Multivariate statistical analysis combined with SHAP-guided machine learning identiﬁed six biomarkers, including thiabendazole, 1-methylxanthine, imidazol-5-yl-pyruvate, 5-hydroxypentanoic acid, spermidine, and 4 ’ -oxolividamine that consistently distinguished OIPN patients (Level 1-3) from non-OIPN controls (Level 0) . Ma

FAQs

What was the study design and patient cohort size?

The study used a prospective cohort of 129 colorectal cancer patients receiving oxaliplatin-based chemotherapy, stratified into four OIPN severity grades (Level 0-3).

How were metabolites measured and how many were detected?

Serum samples were analyzed using untargeted LC-MS/MS metabolomics, detecting 521 metabolites for downstream modeling.

Which machine-learning approach was used to select biomarkers, and what key discrimination result was reported?

SHAP-guided random forest with multivariate statistical analysis identified six biomarkers. Models validated across six classifiers showed near-perfect discrimination for early-stage OIPN (AUC nearly 1).

Aug 5, 2026

Upload by Clementine