PT  - JOURNAL ARTICLE
AU  - Lefteris Livieratos
AU  - Adrien Le Pogam
AU  - Nicolas Boussion
AU  - Catherine Cheze-Le Rest
AU  - Dimitris Visvikis
TI  - Partial volume effect (PVE) correction in SPECT during or post reconstruction: A comparison study
DP  - 2009 May 01
TA  - Journal of Nuclear Medicine
PG  - 1498--1498
VI  - 50
IP  - supplement 2
4099  - http://jnm.snmjournals.org/content/50/supplement_2/1498.short
4100  - http://jnm.snmjournals.org/content/50/supplement_2/1498.full
SO  - J Nucl Med2009 May 01; 50
AB  - 1498 Objectives Reconstruction based resolution recovery has been recently made widely available in clinical systems. In this study we compared the PHILIPS ASTONISHTM with a post-reconstruction deconvolution algorithm for PVE correction in SPECT. Methods The Philips ASTONISHTM reconstruction algorithm is based on 3D-OSEM with built-in corrections for resolution recovery (incorporating a depth dependent resolution model), scatter and attenuation correction. The post-processing algorithms were based on iterative deconvolution using the Lucy-Richardson and Van-Cittert algorithms. Wavelet and Curvelet based denoising were incorporated into the iteration process in order to eliminate the noise observed in deconvolution methods. Reconstruction based and post-processing approaches were compared on both a 99mTc SPECT/CT acquisition of the IEC phantom as well as clinical bone SPECT studies. Corrected images were compared in qualitative and quantitative terms with those reconstructed using a standard OSEM technique with attenuation and scatter correction. Results The post processing deconvolution algorithms with a wavelet/curvelet regularization showed similar or slightly better results compared to the ASTONISHTM reconstruction approach in terms of quantitative accuracy in the different spheres of the phantom. Furthermore, profile plots through the spheres indicate that the Lucy-Richardson algorithm leads to results close to ASTONISHTM. The phantom results were confirmed by the patient images obtained with the two different approaches. Conclusions A comparison of reconstruction based and post-processing deconvolution based approaches for PVE correction in SPECT show similar levels of qualitative and quantitative accuracy improvements.