论文部分内容阅读
目的利用31P磁共振波谱分析数据,区别肝细胞肝癌、肝硬化和正常的肝组织。方法从33例包括肝细胞肝癌、肝硬化和正常肝的志愿者中选择66个有效体素样本,利用1.5T超导MR扫描仪采集31P MRS数据,通过BP神经网络和SVM模型的实验来区别肝细胞肝癌、肝硬化和正常肝脏组织。结果有限的样本实现了良好的分类性能,反向传播神经网络(BP)和支持向量机(SVM)模型可以提高31P MRS识别率,识别率可达92.31%。结论基于BP和SVM的31P MRS数据分析,对于活体肝细胞肝癌的诊断提供了一种可选择的有价值的技术。
Objective To analyze the data of 31P magnetic resonance spectroscopy to distinguish hepatocellular carcinoma, cirrhosis and normal liver tissue. Methods Sixty-three valid voxel samples were selected from 33 volunteers including hepatocellular carcinoma, cirrhosis and normal liver. 31P MRS data were acquired using a 1.5T superconducting MR scanner and compared with BP neural network and SVM model experiments Hepatocellular carcinoma, cirrhosis and normal liver tissue. The results of the finite samples achieved good classification performance. The BP neural network (BP) and support vector machine (SVM) models can improve the recognition rate of 31P MRS with the recognition rate of 92.31%. Conclusions 31P MRS data analysis based on BP and SVM provides an alternative valuable technique for the diagnosis of hepatocellular carcinoma in vivo.