Hsu BL, Ljungberg M, Case JA, Bateman TM and Cullom SJ

Objectives : Spectral reconstruction previously has been demonstrated to improve the energy resolution (ER) of gamma camera and then utilized for scatter correction (SC) in cardiac SPECT (Hsu et al, ASNC 2004). We validated the IESR algorithm on energy data of NCAT phantom generated from Monte Carlo simulation (MCS) using the SIMIND code.

Methods: IESR assumed the energy spectrum (ES) of gamma camera is a linear combination of energy components (EC) presented by photon counts contributed from specific energies. It employed the energy impulse responses (EIR) of system (50-200 keV) to model the response of camera to the energies of EC for the probability matrix. EC in the spectrum were then iteratively reconstructed with MLEM method. EIRs with 1 keV sampling interval were provided from MCS of Cardio60 SPECT system (Philips, Milpitas CA) with ER=10.5% at 140 keV. ES of “in-air” point sources of Tc99m and Tl- 201 were reconstructed and compared with MCS results. IESR was then applied to MCS cardiac NCAT phantom projections. Integrated reconstructed spectrum (IRS) defined by summing ES of 64*64 pixels was compared to the summed ES simulated for hypothetically ultra fine ER=2%. SC was then performed for each pixel of projection by only including the EC near photopeaks.

Results : Using 500 iterations (i), IESR effectively separated 70 and 81 keV peaks of Tl- 201 agreed with the ES simulated for ER=2% with minimal differences (Δ) of FWHM (Δ70=1.1 keV, Δ80=1.75 keV) and FWTM (Δ70=2.2 keV, Δ80=2.7 keV) for the two peaks. For mono energetic photopeaks of 135 and 167 keV of Tl-201 and 140 keV of Tc99m, the Δ to MCS were overall less than 2.5 keV. For NCAT cardiac phantom, IESR with i=5000 produced IRS matched with MCS spectrum for ER=2% with (3.2±1.2)% Δ of counts per keV for Tc99m and (8.3±4.2)% for Tl-201. Using IESR for SC, the contrast of wall to cavity was improved 25.4% for Tc99m phantom and 51.2% for Tl-201 with only a slight increase of image noise.

Conclusions : IESR is a physical approach to improve the energy resolution of gamma camera with the capability of reconstructing energy spectrum matched with 2% energy resolution. Scatter correction with IESR enhanced the image contrast of myocardial perfusion images of NCAT phantom generated with Monte Carlo simulation.