Translator Disclaimer
30 March 2007 Dynamic-thresholding level set: a novel computer-aided volumetry method for liver tumors in hepatic CT images
Author Affiliations +
Measurement of the volume of focal liver tumors, called liver tumor volumetry, is indispensable for assessing the growth of tumors and for monitoring the response of tumors to oncology treatments. Traditional edge models, such as the maximum gradient and zero-crossing methods, often fail to detect the accurate boundary of a fuzzy object such as a liver tumor. As a result, the computerized volumetry based on these edge models tends to differ from manual segmentation results performed by physicians. In this study, we developed a novel computerized volumetry method for fuzzy objects, called dynamic-thresholding level set (DT level set). An optimal threshold value computed from a histogram tends to shift, relative to the theoretical threshold value obtained from a normal distribution model, toward a smaller region in the histogram. We thus designed a mobile shell structure, called a propagating shell, which is a thick region encompassing the level set front. The optimal threshold calculated from the histogram of the shell drives the level set front toward the boundary of a liver tumor. When the volume ratio between the object and the background in the shell approaches one, the optimal threshold value best fits the theoretical threshold value and the shell stops propagating. Application of the DT level set to 26 hepatic CT cases with 63 biopsy-confirmed hepatocellular carcinomas (HCCs) and metastases showed that the computer measured volumes were highly correlated with those of tumors measured manually by physicians. Our preliminary results showed that DT level set was effective and accurate in estimating the volumes of liver tumors detected in hepatic CT images.
© (2007) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Wenli Cai, Hiroyuki Yoshida, and Gordon J. Harris "Dynamic-thresholding level set: a novel computer-aided volumetry method for liver tumors in hepatic CT images", Proc. SPIE 6514, Medical Imaging 2007: Computer-Aided Diagnosis, 65142W (30 March 2007);

Back to Top