Session PB7.2

Delineation of Tumor Volume in Cardiac Gated PET Images

J Gubbi*, M Palaniswami, K Tomas, D Binns

The University of Melbourne
Melbourne, Australia

Positron Emission Tomography is a state of the art functional imaging technique used in accurate detection of cancer. It plays a major role in detecting cancer in thoracic regions, which is one of the most lethal of cancers worldwide. Tumors in the thoracic region, particularly closer to heart, face a unique problem of movement due to the involuntary motion of heart and lungs. They keep changing their position, shape and size during each cardiac cycle. Radiation therapy is one of the most widely used therapies. It aims to deliver the necessary therapeutic dose of ionizing radiation to the tumor tissues whilst minimizing irradiation to the surrounding healthy tissues. Movement causes changes in the position of the Gross Tumor Volume (GTV) that can be as much as 2.5 cm depending on the tumor and its location. To avoid irradiation of healthy tissues, accurate tracking of tumor during different stages of cardiac cycle is recommended. Cardiac gating is a process of dividing a single cardiac cycle into several phases and sorting the data into each of these phases according to the location of the acquired image. This helps in localizing the tumor accurately at specific stage of the cardiac cycle. The aim of this study is to accurately track the movement of the tumor in cardiac gated PET image. The proposed method is a two stage approach where the initial stage is based on the standard uptake value. In the second stage, deformable models with a modified level set method are proposed for tracking the tumor accurately. The proposed method is tested on five patient cardiac gated dataset (5 patients, 8 frames per patient with 36 slices per frame) and a very good accuracy of 92% in terms of percentage volume overlap is obtained using the deformable models. As a result, the complete information about tumor movement during cardiac cycle is available in relatively short time. Thus, the modified level set accurately delineated the tumor volume from all eight frames, thereby providing a scope of using PET images towards planning an accurate and effective radio therapy treatment for thoracic cancer.

(Abstract Control Number: 200)