Computer-aided hepatocellular carcinoma analysis

Abstract

A computer-aided hepatocellular carcinoma system is introduced in this paper. Tumor size and tumor intensity changes are heavily referenced in clinical practice while diagnosing hepatocellular carcinoma. In the enhancement scanning, the tumor might have different intensity appearance in three consecutive phases: arterial phase, portal vein phase and delay phase. By comparing the intensity of tumor area to normal liver tissue, the hepatocellular carcinoma diagnosis is promising. The proposed system first segments the original CT images from three different phases and uses the tumor size and intensity information to estimate the possibility of hepatocellular carcinoma. The experiment results of 30 dataset using our developed system prove the feasibility of our system. Further improvement of this system has been discussed to enable the proposed system to handle more complicated clinical cases.