基于显微共聚焦拉曼光谱与多靶标适配体微流控技术鉴定鼻咽癌循环肿瘤细胞的方法  

Identification of circulating tumour cells in nasopharyngeal carcinoma based on microconfocal Raman spectroscopy with multi-target aptamer microfluidics

作　　者：吴紫曼 郭非凡 武薇 豆小文 李健[3] 张秀明[1,2] 熊丹 WU Ziman;GUO Feifan;WU Wei;DOU Xiaowen;LI Jian;ZHANG Xiuming;XIONG Dan(School of Medical Technology,Xinxiang Medical University,Xinxiang 453000,China;Department of Clinical Laboratory Medicine,the Third Affiliated Hospital of Shenzhen University,Shenzhen 518001,China;Department of Otolaryngology,the First Affiliated Hospital,Sun Yat-sen University,Guangzhou 510000,China)

机构地区：[1]新乡医学院医学技术学院,河南新乡453000 [2]深圳大学第三附属医院医学检验科,广东深圳518001 [3]中山大学附属第一医院耳鼻喉科,广东广州510000

出　　处：《分子影像学杂志》2024年第10期1030-1037,共8页Journal of Molecular Imaging

基　　金：深圳市基础研究自由探索项目(JCYJ20230807142806014);深圳市医学重点学科建设经费(SZXK054)。

摘　　要：目的建立一种基于显微共聚焦拉曼光谱技术的鼻咽癌循环肿瘤细胞鉴定方法,实现对鼻咽癌循环肿瘤细胞的无接触、无标记检测。方法体外培养人T淋巴细胞白血病细胞株Jurkat和鼻咽癌细胞株Sune1,利用EpCAM、CD44、EGFR和波形蛋白适配体组装的纳米微流控芯片技术识别和亲和捕获单细胞,并使用显微共聚焦拉曼光谱技术对单细胞进行鉴定,得到单细胞拉曼光谱。以两种机器学习算法:支持向量机和线性判别分析构建分类器,对单细胞拉曼光谱进行建模分析。结果共采集得到474个Jurkat细胞与Sune1细胞的单细胞拉曼光谱,其峰位分析结果显示:与Jurkat细胞相比,Sune1细胞中的腺嘌呤、胸腺嘧啶和鸟嘌呤含量显著下降(P<0.0001);羟脯氨酸、蛋白质和脂类的含量显著上升(P<0.0001),差异有统计学意义。线性判别分析的预测准确率较高,能够有效区分两种细胞,预测准确率高达98.31%。结论本研究基于显微共聚焦拉曼光谱技术和机器学习算法,建立了一种可能适用于鉴定鼻咽癌循环肿瘤细胞的方法,可对鼻咽癌的临床微创诊断产生积极作用。Objective To establish a method for the identification of nasopharyngeal carcinoma circulating tumor cells based on microconfocal Raman spectroscopy,in order to realize the non-contact and marker-free detection of nasopharyngeal carcinoma circulating tumor cells.Methods Human T lymphocytic leukemia cell line Jurkat and nasopharyngeal carcinoma cell line Sune1 were cultured in vitro,and single cells were identified and affinity captured by nanomicronic chip technology assembled by EpCAM,CD44,EGFR and vimentin aptamer.The single cells were identified by microconfocal Raman spectroscopy to obtain single cell Raman spectroscopy.Two machine learning algorithms,support vector machine and linear discriminant analysis,were used to construct a classifier to model and analyze single-cell Raman spectra.Results Single-cell Raman spectra of 474 Jurkat cells and Sune1 cells were collected.Peak location analysis results showed that the contents of adenine,thymine and guanine in Sune1 cells decreased significantly compared with Jurkat cells.The contents of hydroxyproline,protein and lipids increased significantly,and the differences were statistically significant.Linear discriminant analysis had a high prediction accuracy,which could effectively distinguish the two types of cells,and the prediction accuracy was as high as 98.31%.Conclusion Based on microconfocal Raman spectroscopy and machine learning algorithms,this study established a method that may be suitable for the identification of nasopharyngeal carcinoma circulating tumor cells,which has a positive effect on the clinical minimally invasive diagnosis of nasopharyngeal carcinoma.

关 键 词：鼻咽癌 循环肿瘤细胞 显微共聚焦拉曼光谱技术 微流控芯片 机器学习 分离鉴定 

分 类 号：R739.63[医药卫生—肿瘤]