激光生物学报, 2016, 25 (1): 27, 网络出版: 2016-03-22
体外联合使用指纹区及高波数区拉曼光谱诊断胃癌
Diagnosing Gastric Cancer by Using both Fingerprint and High-wavenumber Raman Spectroscopy
胃癌 拉曼光谱 指纹区 高波数区 偏最小二乘判别分析 gastric cancer Raman spectroscopy high-wavenumber (HW) fingerprint (FP) partial least-squares-discriminant analysis (PLS-D
摘要
目的:研究正常胃粘膜和胃癌组织在拉曼光谱指纹区(800-1 800 cm-1)和高波数区(2 800-3 000 cm-1)的光谱特征,并将其联合使用建立胃癌诊断模型。方法:收集38例正常胃粘膜和37例胃癌组织活检标本,采用785 nm激发光拉曼光谱仪进行拉曼光谱采集。比较正常胃粘膜和胃癌组织在指纹区和高波数区的拉曼光谱异同,使用偏最小二乘判别分析(PLS-DA)结合留一法交叉验证建立诊断模型。结果:1)胃癌组织在853 cm-1,879 cm-1,1 003 cm-1,1 047 cm-1,1 173 cm-1,1 304 cm-1,1 319 cm-1,1 338 cm-1,1 374 cm-1、2 932 cm-1谱峰处与正常胃粘膜的拉曼峰强度差异有统计学意义(P<0.05)2)将拉曼光谱指纹区和高波数区联合,利用PLS-DA建立胃癌诊断模型的敏感性为94.59%(35/37),特异性为86.84(33/38),正确率为90.6%(68/75)。结论:正常胃粘膜和胃癌组织在拉曼光谱指纹区和高波数区均有显著差异,将上述两区联合使用建立模型诊断胃癌能取得良好的诊断效果。
Abstract
Object:The aim of this study was to apply Raman spectroscopy in the fingerprint (FP) region (800-1 800 cm-1) and the High-wavenumber (HW) region (2 800-3 000 cm-1) for ex vivo detection of gastric cancer and compare the difference of the Raman spectra between the normal mucosa and the gastric cancer. Method: Raman spectra were collected to differentiate normal mucosa (n=38) and gastric cancer (n=37) in the FP region and the HW region by Micro-Raman Spectroscopy System at 785-nm excitation. Multivariate diagnostic algorithms was generated based on partial least-squares-discriminant analysis (PLS-DA) together with the leave-one-sample-out, cross-validation. Results: 1)The significant Raman spectral difference between normal and cancer tissue was observed at 853 cm-1,879 cm-1,1 003 cm-1,1 047 cm-1,1 173 cm-1,1 304 cm-1,1 319 cm-1,1 338 cm-1,1 374 cm-1、2 932 cm-1(P<0.05). Multivariate diagnostic algorithms yield the diagnostic sensitivities of 94.59%(35/37),specificities of 86.84%(33/38), diagnostic accuracies of 90.6%(68/75), using the integrated FP/HW Raman spectroscopy for ex vivo diagnosis of gastric cancer. Conclusion: There are distinct differences between the normal mucosa and gastric cancer in the FP region and the HW region. Multivariate diagnostic algorithms using the integrated FP/HW Raman spectroscopy performs well in the diagnoses of gastric cancer.
周学谦, 陈瑶, 于乐泳, 代剑华, 袁月, 彭贵勇. 体外联合使用指纹区及高波数区拉曼光谱诊断胃癌[J]. 激光生物学报, 2016, 25(1): 27. ZHOU Xueqian, CHEN Yao, YU Leyong, DAI Jianhua, YUAN Yue, PENG Guiyong. Diagnosing Gastric Cancer by Using both Fingerprint and High-wavenumber Raman Spectroscopy[J]. Acta Laser Biology Sinica, 2016, 25(1): 27.