A respiratory gating system for PET imaging of lung cancer

Abstract

The accurate characterization of lesions in FDG-PET imaging of lung cancer in terms of their size, shape, location and quantitation, can be significantly affected by motion of the lungs during respiration. The objective of this research was to develop a system to acquire and process the patient's breathing signal and use it to "gate" the acquisition of FDG-PET images in order to reduce the effects of respiratory motion. Methods: A high-resolution CCD laser system that senses the displacement of the chest or abdomen during the patient's breathing cycle was employed. The displacement information was appropriately processed to reconstruct the respiratory signal, which in turn was used to trigger the gated acquisition of FDG-PET images in a synchronized fashion from inspiration to expiration. The system was attached to a custom designed mechanical assembly with robotic movements and was tested on a Philips C-PET and a Philips Gemini PET/CT scanner. A motorized oscillating phantom and a ventilator supplied lung simulator were developed for testing and validating the system with positron emitting sources simulating lung lesions at variable speeds of breathing and motion displacement. Results: There was a significant reduction in the effects of respiratory motion with the proposed respiratory gating system. Improved accuracy was noted in determining the spatial location, obtaining the size, and recovering the actual counts of simulated lung lesions with gated vs. non-gated phantom experiments. The reliable operation of the respiratory gating system was also validated in 3 patient studies. Conclusions: A system was developed that enables respiratory-gated acquisition in FDG-PET imaging in order to reduce the effects of respiratory motion, which can affect lung lesions. Phantom experiments and limited clinical applications demonstrated the reliable performance of the system and improved accuracy in the characterization of lung lesions. © 2009 Springer Berlin Heidelberg.