Computer-Aided Diagnosis and Volumetric Measurement of Hepatic Metastasis

Hepatic metastasis diagnosis is based on visual interpretation that remains subjective and depends on the expert's experience in distinguishing metastasis from other diseases. Among the large number of exams, image analysis becomes tedious and time-consuming. In this context, we propose in this paper a Computer Aided Diagnosis System providing an automatic tool for hepatic metastasis detection and volume measurement. In this work, we use abdominal CT-scan exams obtained from the radiology department of the Tunisian National Oncology Institute "Salah AZAIEZ" and the "3D-IRCADb" open access database. The aim of our work is to detect the liver, identify the hepatic metastasis and give a volumetric measurement of these lesions. Our proposed system includes three parts. A first one dedicated to extract the liver from abdominal slice and to index the region of interest using the Fuzzy C-means algorithm. The second part of the proposed system is dedicated to identify the hepatic metastasis. The third part corresponds to the volumetric measurement of the lesion and presents a 3D visualization of the total volume of the liver and the hepatic metastasis. In results, we obtained over than 99.70% of accuracy for the proposed method when applied on our database. Furthermore, to ensure its effectiveness, we applied it on the "3D-IRCADb" open access database where we achieved 98.98% of accuracy. We have also demonstrated that it has a high performance compared to the level set and the classical region growing algorithm. The results of the proposed method allows quick and reliable hepatic metastasis detection. We consider our proposed system as a very useful tool for post-treatment in detecting and evaluating the hepatic metastasis progress. The volumetric computation is used to measure the functional volume of the liver which is a well used indicator in predicting the efficiency of transplantation and the therapy success.