Reduced acquisition time or activity does for ^sup 123^I-FP-CIT SPECT with compressed sensing-based iterative reconstruction

Objectives: The compressed sensing is a signal processing technique that was initially proposed for spare signal recovery. And it is capable of reconstructing high quality images from a substantially smaller number of views than those needed in filtered back projection (FBP). This study aimed to evaluate CS-based iterative reconstruction (CS-IR) for 123I-FP-CIT SPECT to reduce image acquisition time or activity dose. Methods: We used the cylinder/sphere and the striatal digital phantom models. The total projection number was set at 120, 90, 60, and 30. In addition, acquired maximum number of photons in a projection was varied from 25 to 100. SPECT images were reconstructed using FBP, Maximum likelihood expectation maximization (MLEM), and CS-IR. The total-variation transform with local image gradient in L1-norm was adopted in our compressed sensing algorithm. The effects of CS-IR were examined according to the resolution recovery factor, activity concentration linearity, percent coefficient of variation (%CV), and specific binding ratio (SBR). Results For the resolution recovery factor and activity concentration linearity, no differences were observed between each reconstruction when decreasing the projection number and counts. Decreasing the projection number and counts, %CV obtained by CS-IR was superior to that obtained by other reconstruction. %CV obtained by 60 projection CS-IR was equivalent to that obtained by 120 projection FBP and MLEM. The variation of SBR was the highest in FBP and SBR obtained by CS-IR was similar to that of MLEM. Conclusions We demonstrated that CS-IR with decreasing the projection number and acquisition count provided a good image quality compared to commonly SPECT reconstruction. Compressed sensing could help to reduce overall SPECT acquisition time or activity does in 123I-FP-CIT SPECT.