Copyright: ©Author(s) 2026. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution-NonCommercial (CC BY-NC 4.0) license. No commercial re-use. See permissions. Published by Baishideng Publishing Group Inc.
World J Gastroenterol. Dec 14, 2026; 32(46): 121708
Published online Dec 14, 2026. doi: 10.3748/wjg.121708
Published online Dec 14, 2026. doi: 10.3748/wjg.121708
Inflammation-lipid metabolism biomarker-based machine learning models for predicting postoperative pulmonary metastasis in colorectal cancer
Yan-Yuan Du, Yu-Ming Liu, Rui-Ying Fang, Hong-Gang Zheng, Department of Oncology, Guang’anmen Hospital, China Academy of Chinese Medical Sciences, Beijing 100053, China
Co-first authors: Yan-Yuan Du and Yu-Ming Liu.
Author contributions: Du YY and Liu YM were responsible for data curation, drafting the original manuscript, reviewing and editing; Fang RY contributed to data curation; Zheng HG. conceived and designed the study and conducted the investigation. Du YY and Liu YM are designated as co-first authors in recognition of their equal and substantial intellectual contributions to this work. Both authors were primarily responsible for data curation, initial drafting of the manuscript, and critical revision of its content. They jointly contributed to the acquisition, management, and interpretation of study data, and each played a leading role in manuscript preparation and refinement. Their combined efforts ensured the integrity, accuracy, and quality of the final submission, thereby justifying equal contribution as first authors.
AI contribution statement: The authors declare the following regarding the use of artificial intelligence (AI) tools in the preparation of this manuscript: (1) AI tools used: Only Grammarly was used, solely for grammar and spelling correction of author-written text. No generative AI tool was used; (2) Main Text generation: No portion of the Main Text, including the Abstract, Introduction, Materials and Methods, Results, Discussion, and Conclusion, was generated, in whole or in part, by any AI tool. All text was written by the authors; (3) Language polishing, translation, data analysis, and writing assistance: No AI tool was used for substantive language polishing, translation, data analysis, or writing assistance. All statistical analyses and machine-learning modelling were carried out by the authors using standard, non-generative statistical and ML libraries; the complete analysis code is publicly available in the open-source repository cited in this manuscript and is fully reproducible. The use of Grammarly was limited to grammar correction; (4) Study design and interpretation of results: No AI tool participated in the design of the study or in the interpretation of its results. Both were performed entirely by the authors; and (5) Images: No images, figures, or diagrams in this manuscript were generated by AI. All figures were produced by the authors using conventional scientific plotting and illustration software, with the underlying code openly available. The authors take full responsibility for the integrity, accuracy, and originality of all content presented in this manuscript.
Supported by Municipal Natural Science Foundation of Beijing of China, No. 7252262; Beijing-Tianjin-Hebei Academic Inheritance and Promotion Program for Renowned Senior TCM Experts: Piao Bingkui, No. GZY-GCS-2018-071; Central Government Program for Enhancing Clinical Research and Translational Capacity of High-Level TCM Hospitals - Special Project for Inheriting Academic Experience of Famous Senior TCM Experts, No. HLCMHPP2023007; Central Government Program for Enhancing Clinical Research and Translational Capacity of High-Level TCM Hospitals - Special Initiative for Evidence-Based Clinical Research in Traditional Chinese Medicine, No. HLCMHPP2023085; and National Major Science and Technology Project, No. 2023ZD0501700.
Institutional review board statement: This study was conducted in accordance with the Declaration of Helsinki. It has been reviewed and approved by the Ethics Committee of Guang’anmen Hospital, China Academy of Chinese Medical Sciences (approval number: 2025-238-KY), with approval covering the participating external center.
Informed consent statement: The informed consent was waived by the Institutional Review Board.
Conflict-of-interest statement: The authors declare that they have no conflict of interest.
STROBE statement: The authors have read the STROBE Statement—checklist of items, and the manuscript was prepared and revised according to the STROBE Statement—checklist of items.
Data sharing statement: The datasets used and analysed during the current study are available from the corresponding author on reasonable request. The source code used in this study is available at https://github.com/heilegehei/CRCLM.
Corresponding author: Hong-Gang Zheng, MD, PhD, Professor, Department of Oncology, Guang’anmen Hospital, China Academy of Chinese Medical Sciences, No. 5 Beixiange, Xicheng District, Beijing 100053, China. honggangzheng@126.com
Received: March 31, 2026
Revised: April 8, 2026
Accepted: May 26, 2026
Published online: December 14, 2026
Processing time: 203 Days and 19.2 Hours
Revised: April 8, 2026
Accepted: May 26, 2026
Published online: December 14, 2026
Processing time: 203 Days and 19.2 Hours
Core Tip
Core Tip: No prior model predicts colorectal cancer pulmonary metastasis using only preoperative routine blood-derived inflammation and lipid metabolism biomarkers. In this 1246-patient, dual-center study, we systematically compared 14 machine learning algorithms and three ensemble strategies - the most extensive benchmark reported for this task. The optimal Weighted Voting ensemble achieved validated discrimination [internal area under the curve (AUC) = 0.7513; external AUC = 0.7112] using 16 readily obtainable predictors. SHapley Additive exPlanations interpretability analysis revealed the lymphocyte-to-C-reactive protein ratio and body mass index as dominant predictors, offering a practical risk-stratification tool before pathological staging becomes available.