31Phosphorus Magnetic Resonance Spectroscopy Data Analysis of the Hepatocellular Carcinoma Using Artificial Neural Networks

Through the evaluation of the ³¹Phosphorus Magnetic Resonance Spectroscopy (³¹P-MRS), we can distinguish three types of diagnosis: hepatocellular carcinoma, normal and cirrhosis. 71 samples of ³¹P-MRS data are selected including hepatocellular carcinoma, normal and cirrhosis tissue. Back-propagation neural network (BP) and Radial Basis Function Neural Network (RBF) are applied to analyze ³¹P-MRS data, develop neural network models of ³¹P-MRS for the diagnostic classification of hepatocellular carcinoma. The results suggest that BP models have better performance than RBF models. Neural network models based on ³¹P-MRS data offer an alternative and promising technique for diagnostic prediction of hepatocellular carcinoma in vivo.