论文部分内容阅读
以腹膜接种人胃癌细胞SGC-7901的裸鼠为胃癌腹膜播散的动物模型,进行模拟外科手术,在体探测不同种植期的裸鼠腹膜癌结节及正常腹膜组织的激光拉曼光谱,对比光谱差异,采用支持向量机(SVM)算法对光谱进行分类和分期判决。结果表明,癌结节和正常组织拉曼光谱差异显著,用支持向量机算法进行分类的灵敏度、特异度和诊断准确度分别为95.73%、70.73%和90.73%;不同生长期的癌结节组织拉曼光谱也存在明显差异,用支持向量机算法进行分期的结果分别为98.82%、98.73%和98.78%。从分类结果可以看出,此方法对指导外科手术中癌变组织的识别有重要的意义。
The peritoneal inoculation of human gastric cancer cell line SGC-7901 in nude mice is an animal model of peritoneal dissemination of gastric cancer. Simulated surgical procedures were performed to detect laser Raman spectra of peritoneal nodules and normal peritoneal tissue in nude mice at different planting stages. Spectral differences, the support vector machine (SVM) algorithm to classify the spectrum and staging decisions. The results showed that there were significant differences in Raman spectra between normal nodules and normal nodules. The sensitivity, specificity and diagnostic accuracy of SVM classification were 95.73%, 70.73% and 90.73%, respectively. Cancer nodules There are also significant differences in Raman spectra. The results of staging using support vector machine are 98.82%, 98.73% and 98.78% respectively. It can be seen from the classification results that this method is of great significance to the identification of cancerous tissues in surgical operation.