J Nucl Med. 2007; 48 (Supplement 2):42P
Instrumentation & Data Analysis: Image Generation
Motion Effects and Compensation in PET and SPECT |
An anthropomorphic phantom for simulating cardiac and respiratory motion in SPECT/CT and PET/CT
Karthikayan Balakrishnan1,
Rostyslav Boutchko1,
Bryan W Reutter1,
Arkadiusz Sitek2 and
Grant T Gullberg1
1 Lawrence Berkeley National Laboratory, Berkeley, California;
2 Brigham and Women's Hospital and Harvard Medical School, Boston, Massachusetts
140
Objectives: Most physical phantoms do not model respiratory motion which contributes significantly to image degradation in SPECT and PET scans. We investigate the feasibility of constructing a phantom that mimics respiratory and cardiac cycles in a human allowing acquisition of nuclear medicine data that simulates patient cardiac and respiratory motion. The phantom is used to evaluate algorithms that compensate for these physiological motions that result in mismatch between SPECT/PET and CT data in hybrid systems. Methods: Highly flexible liquid silicone rubber was coated over a cast of a life-sized model. When the cast is removed, the coating forms the lungs of the phantom. These are inflated and deflated by a tank of compressed air with the inflation/deflation cycle controlled by computer operated solenoid valves. The heart, adapted from the Jaszczak dynamic phantom, rests between the inflatable rubber lungs. Changes in the lung volume cause non-linear deformation of the lung and, to a lesser extent, the heart, the internal organs and attenuating body. Fluid is pumped in and out of the inner most chamber of the heart by a modified rodent respirator. A limit-switch on the respirator provides an R-wave trigger for acquisition of ECG gated data. Two activity filled sealable rubber sacs independently model the liver and the rest of the abdomen with different uptake levels. All parts are elastic, interconnected and enclosed in a non-rigid container filled with water. Glass beads placed on the surface of the phantom act as fiducial markers. Data collection from the phantom is either the snap-shot mode, where the phantom is in one of many repeatable static stages, or the dynamic mode where the phantom 'breathes' and 'beats' continuously. Results: Images from a preliminary version of the motion phantom show significant conformation changes in all of the simulated organs. Sinograms were successfully obtained for both the snap-shot and the dynamic modes. Conclusions: The snap-shot mode allows one to create a database of SPECT/PET sinograms and attenuation maps for the different cardiac and respiratory phases.
Research Support (if any): NIH R01-HL71253, R01-EB00121 and R01-HL50663; and U.S. DOE Contract DE-AC02-05CH11231
