RT Journal Article SR Electronic T1 Demonstration of real-time imaging system for the OpenPET toward PET-guided tumor-tracking radiation therapy JF Journal of Nuclear Medicine JO J Nucl Med FD Society of Nuclear Medicine SP 1947 OP 1947 VO 52 IS supplement 1 A1 Hideaki Tashima A1 Eiji Yoshida A1 Shoko Kinouchi A1 Fumihiko Nishikido A1 Naoko Inadama A1 Hideo Murayama A1 Kyosan Yoshikawa A1 Mikio Suga A1 Hideaki Haneishi A1 Taiga Yamaya YR 2011 UL http://jnm.snmjournals.org/content/52/supplement_1/1947.abstract AB 1947 Objectives We have proposed the OpenPET, an open type PET, which has a physically opened field-of-view between two detector rings. Especially, the OpenPET has great potential for the real-time tumor-tracking radiation therapy, if real-time imaging is realized. The real-time tumor-tracking is desired especially for the treatment of tumors in mobile organs such as lung moving with respiration. However, the real-time imaging of the PET is challenging task because of its computational burden for the image reconstruction. To the best of our knowledge, no one has ever actually demonstrated the real-time imaging of the PET. Therefore, the aim of this study is to demonstrate the real-time OpenPET imaging by proposing a new system architecture and implementing it on a small prototype system. Methods For real-time image reconstruction, we used recently developed fast PET reconstruction algorithm, 3D one-pass list-mode dynamic row-action maximum likelihood algorithm, and implemented on the GPU. In addition, we proposed a system architecture to stabilize the reconstruction speed by incorporating a data transfer control system which controls data flow and limits maximum data usage in the reconstruction step in every time frame. The proposed architecture was implemented on a single computer connected to the data acquisition system of the small OpenPET prototype. Results Demonstrational experiments of the developed system showed real-time imaging; tracking of a moving point source (1 MBq) was possible in 4D reconstructed images with a frame rate of 2 frames per second. Conclusions The demonstration showed possibility of realizing the PET-guided tumor-tracking radiation therapy. We expect the real-time system also enables PET-guided biopsy. Our next work will include experiments using more complicated and realistic phantoms