Combined respiratory and rigid body motion compensation in cardiac perfusion SPECT using a visual tracking system

P. Hendrik Pretorius, University of Massachusetts Medical School
Michael A. King, University of Massachusetts Medical School
Karen L. Johnson, University of Massachusetts Medical School
Joyeeta M. Mukherjee, University of Massachusetts Medical School
Joyoni Dey, University of Massachusetts Medical School
Arda Konik, University of Massachusetts Medical School

Abstract

We report on the validation of our combined respiratory and rigid body motion compensation strategy through acquisitions of the Data Spectrum anthropomorphic phantom, and investigation of clinical efficacy and robustness in 25 cardiac perfusion patient studies, employing a visual tracking system (VTS). The heart and liver was filled with a 2:1 concentration of Tc-99m and two sets of SPECT data were acquired. Each set of SPECT data consisted of a rest-perfusion baseline frame-mode emission acquisition, a Beacon (Philips, Cleveland, OH) transmission acquisition, and a list-mode emission acquisition. Respiratory motion was simulated during the list-mode acquisitions using the Quasar (Modus Medical Devices Inc. ON, Canada). Rigid-body motion was introduced in one of the two list-mode acquisitions by rotating the phantom around the x-axis and translating the phantom in x, y, and z. Patient volunteers with written consent were similarly acquired and asked to execute some predefined body motion during the list-mode acquisition. Motion tracking was performed using 6 near infrared Vicon cameras in combination with 7 retro-reflective markers, 5 placed on the chests of both patient volunteers and phantom, 2 placed on the abdomen of patient volunteers, and 2 placed on the vertical motion stage of the Quasar to simulate abdominal phantom motion. Processing steps included, down sampling VTS positional data to 10 Hz (100 ms) and synchronized with 100 ms SPECT frames, separating rigid body and respiratory motion and estimating 6 DOF rigid body motion, amplitude bin 100 ms frames into respiratory projection sets, reconstruct with rigid body motion compensation respiratory projection sets, estimated respiratory motion employing intensity based estimation, combine rigid body and respiratory motion, and reconstruct with combined compensation. We showed for both phantom and patient acquisitions that combined respiratory and rigid body motion compensation improve the visual appearance of slices. © 2011 IEEE.