PO.CL01.03 · 临床研究

泰国尿液农药生物标志物与肝病风险：一个基于机器学习的风险预测模型

Urinary pesticide biomarkers and liver disease risk in Thailand: A machine-learning-based risk-prediction model

编号 2443 展板 13 时间 4/20 09:00–12:00 区域 Section 40 主讲 Daxeshkumar Patel, PhD

分会场 Biomarkers Predictive of Therapeutic Benefit 3

查看 PDF 下载 PDF 🔒 查看 / 下载完整 PDF 需登录并开通下载套餐 · 查看套餐 / 开通 AACR 官方页面

作者与单位 Authors & Affiliations

Daxeshkumar P. Patel¹, Christopher Loffredo², Majda Haznadar³, Mohammed Khan⁴, Amelia Parker⁴, Benjarath Pupacdi⁵, Siritida Rabibhadana⁶, Panida Navasumrit⁷, Nirush Lertprasertsuke⁸, Anon Chotirosniramit⁹, Chawalit Pairojkul,¹⁰, Vor Luvira,¹¹, Ake Pugkhem¹², Wattana Sukeepaisarnjaroen¹², Teerapat Ungtrakul¹³, Thaniya Sricharunrat¹⁴, Kannika Phornphutkul¹⁵, Frank J. Gonzalez¹⁶, Anuradha Budhu¹⁷, Chulabhorn Mahidol⁶, Xin Wei Wang¹⁸, Mathuros Ruchirawat⁷, Curtis C. Harris¹⁸, TIGER-LC Consortium

¹LHC, CCR, NIH-NCI, Bethesda, MD,²Georgetown University, Washington, DC,³FDA, Silver Spring, MD,⁴Laboratory of Human Carcinogenesis, Center for Cancer Research, National Cancer Institute, Bethesda, Maryland, USA, Bethesda, MD,⁵Translational Research Unit, Chulabhorn Research Institute, Bangkok, Thailand,⁶Laboratory of Chemical Carcinogenesis, Chulabhorn Research Institute, Bangkok, Thailand,⁷Laboratory of Environmental Toxicology, Chulabhorn Research Institute, Bangkok, Thailand,⁸Department of Pathology, Faculty of Medicine, Chiang Mai University, Chiang Mai, Thailand,⁹Department of Surgery, Faculty of Medicine, Chiang Mai University,, Chiang Mai, Thailand,¹⁰Faculty of Medicine,, Khon Kaen University, Khon Kaen, Khon Kaen,, Thailand,¹¹Khon Kaen University, Khon Kaen, Thailand;, Khon Kaen, Thailand,¹²Faculty of Medicine,, Khon Kaen University, Khon Kaen, Thailand;, Khon Kaen, Thailand,¹³Princess Srisavangavadhana Faculty of Medicine,, Chulabhorn Royal Academy, Bangkok, Thailand;, Bangkok, Thailand,¹⁴Princess Srisavangavadhana Faculty of Medicine, Chulabhorn Royal Academy, Bangkok, Thailand,¹⁵Rajavej Hospital, Chiang Mai, Thailand,¹⁶Cancer Innovation Laboratory, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD,¹⁷Laboratory of Human Carcinogenesis, Center for Cancer Research, National Cancer Institute, Bethesda, MD,¹⁸National Cancer Institute, Bethesda, MD

摘要 Abstract

中文摘要

背景基于将尿液草甘膦（glyphosate）与慢性肝病（CLD）和肝细胞癌（HCC）联系起来的证据，我们开发了将尿液农药分析与机器学习风险预测（MLRP）相整合的方法，以对高暴露人群进行风险分层。方法我们在“泰国肝癌基因组与表达研究计划”（TIGER-LC；2011-2016；n=593）内开展了一项病例对照研究：228例CLD、116例HCC和249例对照。通过LC-MS/MS定量了8种尿液农药（二甲戊灵、恶草酮、甲磺隆、丁草胺、2,4-二氯苯氧乙酸[2,4-D]、氯氰菊酯、氟鼠灵、溴敌隆）。构建了含与不含草甘膦的复合农药负荷评分（PLS）以估计暴露负担。开发了两个预测模型：基于逻辑回归的农药知情肝癌风险评分（PILCRS）和纳入年龄、性别、饮酒、职业和PLS的极端梯度提升（XGBoost）分类器。内部效度采用1,000次bootstrap重抽样，并进行乐观度校正的校准。结果预测的CLD概率从PLS最低四分位数组的30%上升至最高组的70%以上，HCC则从10%上升至40%（p<0.0001）。相对估计值一致；最高四分位数组相比最低组，CLD的比值比为2.84（95% CI 1.66-4.91），HCC为4.76（2.30-10.29）。氯氰菊酯仍保持独立相关。经乐观度校正后，两个模型均表现出良好的区分度和校准度。解读该框架建立了一个可扩展的、基于暴露信息的肝病预测工具。研究结果强调农药负荷是一种可改变的风险因素，并与可持续发展目标3.9及世卫组织-粮农组织在中低收入国家（LMICs）的优先事项相一致。外部验证至关重要。

查看英文原文 English abstract

Background Building on evidence linking urinary glyphosate to chronic liver disease (CLD) and hepatocellular carcinoma (HCC), we developed urinary pesticide profiling integrated with machine learning risk prediction (MLRP) to stratify risk in high-exposure populations.Methods We conducted a case-control study within the Thailand Initiative in Genomics and Expression Research for Liver Cancer (TIGER-LC; 2011-2016; n=593): 228 CLD, 116 HCC, and 249 controls. Eight urinary pesticides were quantified by LC-MS/MS (pendimethalin, oxadiazon, metsulfuron-methyl, butachlor, 2,4-dichlorophenoxyacetic acid [2,4-D], cypermethrin, flocoumafen, bromadiolone). A composite Pesticide Load Score (PLS), with and without glyphosate, estimated burden. Two predictive models were developed: a logistic-regression Pesticide-Informed Liver Cancer Risk Score (PILCRS) and an Extreme Gradient Boosting (XGBoost) classifier that incorporated age, sex, alcohol use, occupation, and PLS. Internal validity used 1,000 bootstrap resamples with optimism-corrected calibration.Findings Predicted CLD probability increased from 30% in the lowest PLS quartile to over 70% in the highest, and HCC from 10% to 40% (p<0ꞏ0001). Relative estimates were consistent; the highest versus lowest quartile yielded odds ratios of 2ꞏ84 (95% CI 1ꞏ66-4ꞏ91) for CLD and 4ꞏ76 (2ꞏ30- 10ꞏ29) for HCC. Cypermethrin remained independently associated. After optimism correction, both models demonstrated strong discrimination and calibration.Interpretation This framework establishes a scalable, exposure-informed tool for liver disease prediction. Findings underscore pesticide burden as a modifiable risk factor and align with Sustainable Development Goal 3ꞏ9 and WHO-FAO priorities in low- and middle-income countries (LMICs). External validation is essential.

利益披露 Disclosure

D. P. Patel, None.. C. Pairojkul,, None.. V. Luvira,, None.. A. Pugkhem, None.. W. Sukeepaisarnjaroen, None.. T. Ungtrakul, None.

← 返回 AACR 2026 检索