基于代谢组学标记识别的高发区人群食管癌筛检模型研究

The technology “Iodine staining under endoscope + indicative biopsy” is widely used for esophageal cancer (EC) screening in high-incidence area of China. However, the reported screening positive rate is no more than 20-30%, while the positive predictive value (PPV) is under 2%, which make it inefficiency. It is known that EC is caused by the interaction of multiple genes and environmental factors, manifesting as the gradual accumulation of environmental risk factors and continuous damage to the homeostasis of various metabolic pathways. Our previous metabolomics studies confirmed this hypothesis, and discover some differential expressed metabolites between screening positive and negative groups, such as amino acids, organic acids and phospholipids. Therefore, based on the above promising result, this project further design a large-scale population based metabolomics study for EC screening, based on EC screening cohort "National EC early detection and treatment bases (Feicheng city)", LC-QTOF/MS platform, statistical simulation and pattern recognition. The aim of this project is to identify the metabolic biomarkers for early EC screening, and established the optimal early EC screening model for high-risk population in China. The project will realize EC screening through the simple, rapid determination of serum metabolic markers, so as to identify the endoscope target population, which provide important bases for early EC screening in high-risk population.

我国食管癌筛查采用“内镜下碘染色+指示性活检”技术，在高发区内镜筛查碘染色阳性率在20%-30%，经病理证实筛检出的可干预阳性病变不足2%，存在一定的局限性。研究表明食管癌由多基因和环境交互所致，表现为环境因素逐渐积累并不断损伤机体各代谢通路稳态。本课题组前期研究利用内镜筛查碘染色技术,发现了阳性和阴性对照组间存在大量氨基酸、有机酸和磷脂类物质代谢差异;为深入研究,本项目拟依托“国家食管癌早诊早治示范基地-肥城市”，借助LC-QTOF/MS检测平台、统计模拟和模式识别方法，开展大样本的人群食管癌筛查代谢组学研究。预期目标是筛选可用于食管癌筛查和高危人群早期识别的代谢组学标记，构建适合高发区人群推广的食管癌筛查模型，并在外部人群中应用验证。该模型的建立将实现通过简单、快速测定血清中代谢组标记确定内镜筛查的目标人群，对于食管癌高发区人群筛查的实施和推广有着重要的理论价值和实际意义。

目前我国食管癌早诊早治工作主要采取对高发区内高危人群（即40-69岁人群）直接应用“内镜+碘染色指示性活检”技术进行筛查，该方法为有创检查，操作复杂、成本较高，且筛查出的阳性可干预病变不足2%，在一定程度上制约了我国食管癌防治工作继续深入。课题组依托“国家食管癌早诊早治示范基地（肥城市）”建立食管癌筛查与随访人群队列和生物标本库，借助高通量UPLC-QTOF/MS代谢组学检测技术，以及高维统计分析与模式识别方法，旨在开发一种新的、简便的初筛方法，以便精准锁定高危人群，有效缩小内镜检查范围，提高筛查效率。课题共入组调查对象5449例（其中包括正常2115例，食管癌前病变570例，食管癌56例），并采集血清标本和流行病学等资料。经研究发现通过代谢组学轮廓分析可以清楚地区分食管癌患者和健康对照，食管癌患者的血清中有16种代谢物受到干扰，涉及到脂肪酸生物合成，甘油磷脂代谢，癌症中的胆碱代谢和亚油酸代谢4种代谢通路。诊断模型在外部验证集中的AUC值为0.895，灵敏度为0.850，特异度为0.905，能够较好的识别食管癌患者。食管原位癌（0期）、早期食管癌（I/II期）和晚期食管癌（III期）患者的AUC分别为0.881、0.881和0.929，显示其具有良好的应用价值。另外，研究发现6种代谢物可用于食管癌的初步分期，其中3种生物标记物，十二烷酸，LysoPA(18:1)和LysoPC(14:0)表现出明显的随疾病程度进展的趋势。本研究成果可为食管癌筛查和高危人群早期识别提供重要信息，对于食管癌人群筛查的实施和推广有着重要的理论价值和实际意义。