基于稳健统计的新鲜乳腺组织拉曼光谱分析

采用便携式拉曼光谱仪对新鲜乳腺正常组织、良性组织和恶性组织进行检测,通过稳健统计方法对拉曼光谱数据进行分析处理,建立乳腺组织拉曼光谱标准图谱,根据标准图谱特征峰归纳3类组织的主要区别和特征.在3类乳腺组织中,正常组织有明显的脂类特征峰(1078,1297,1437,1653,1746 cm-1),而在良性和恶性组织中则出现了较明显的蛋白特征峰(1259,1530,1650 cm-1),正常、良性和恶性组织的主要区别集中在1340和1534 cm-1处,应归属为蛋白和类胡萝卜素,这一结果并不能由经典统计方法得出.基于稳健统计建立的新鲜乳腺组织拉曼光谱标准图谱为构建数学模型来鉴别乳腺病灶的性质奠定了基础.

Study on Fresh Breast Tissues by Raman Spectroscopy Based on Robust Statistics

A portable Raman spectrometer was used for distinguishing the characteristics of normal, malignant and benign fresh breast tissues. Robust statistics method was used to analyze the Raman spectrum data, and a standard Raman spectral atlas of fresh breast tissues was established. Based on the spectral profiles of the Raman spectral standard atlas, the presence of lipids(1078, 1297, 1437, 1653, 1746 cm-1) is indicated in normal tissue. And proteins(1259, 1530, 1650 cm-1) are found in benign and malignant tissues. Among normal, benign and malignant tissues, proteins(1340 cm-1) and β-carotene(1534 cm-1) are found to be good discrimination parameters, which are hardly achieved by other typical statistical methods. The established fresh breast tissue Raman spectral standard atlas based on robust statistics laid the foundation of benign and malignant breast lesions identify model.