Usage of the XCAT Phantom to Investigate Respiratory Motion and Estimation of Attenuation Maps in Pediatric SPECT DMSA Studies

Forum Details

Date: June 17, 2022 12-1 PM Eastern time

Speakers: Justin Pijanowski and Clifford Lindsay Ph.D., UMass Chan Medical School

The aim of this project is to minimize radiation dose administered to pediatric DMSA SPECT images assessed for renal disease while retaining image quality. For this talk, we discuss our efforts in addressing image degradation in two areas, specifically patient motion, and attenuation estimation with the use of CT. The effect of respiration on the kidneys was evaluated and the feasibility of data driven motion compensation in DMSA SPECT kidney imaging was studied. XCAT phantoms were used to introduce respiratory motion into the kidneys during simulation. Additionally, DMSA SPECT studies acquired without CT for attenuation compensation to reduce dose, requires manual segmentation of the emission data to generate a uniform attenuation map. To improve attenuation compensation, we developed a Deep learning (DL) method to predict variable linear attenuation coefficients from reconstructed photopeak and scatter window data. Our DL method utilizes a large population of XCAT phantoms and SIMIND Monte Carlo simulations to create an initial training dataset.