Evaluation of brain dopamine transporter SPECT images using iterative reconstruction and attenuation correction

Objectives The aim of this study was to examine the optimal reconstruction parameters for brain dopamine transporter SPECT images using iterative reconstruction with attenuation correction.

Methods The striatal phantom was filled with ¹²³I solution at a striatum-to-background radioactivity ratio of 8.08, 6.03, 4.03 and 3.01. Data acquisition was performed with using two SPECT/CT devices: a Symbia T6 and an Infinia Hawkeye4 (device A and B) equipped with low-medium energy general purpose and low-energy high-resolution collimators (C1, C2). The SPECT images were reconstructed using iterative reconstruction with both CT and Chang’s attenuation correction (AC). Scatter correction was also performed. The images were evaluated using the specific binding ratio (SBR) of the striatum and the visual assessment based on a five-step scale for image quality. Regions-of-interest for the background were placed on both the frontal and occipital regions.

Results An approximate update number of 100 for device A-C1 and 50 for the other devices were considered to be necessary. The optimal numbers of iteration and subset were determined to be (10, 10) for A-C1, (10, 5) for A-C2 and (7, 6) for both B-C1 and B-C2. The recovery of the SBR was 69.6%, 47.1%, 48.4% and 52.3%, respectively. The scores for the visual assessment were 4, 3, 3 and 4, respectively. The SBR and image quality were superior to that of the images reconstructed using filtered back projection (FBP) in our previous study. The SBR calculated with the background activity in the frontal region was lower than that in the occipital region for the images with Chang’s AC, and higher for those calculated with CTAC.

Conclusions Using iterative reconstruction, the image quality and SBR were improved compared with that obtained with the FBP method. The CT method therefore was considered to be useful for attenuation correction.