HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Staelens, S. Articles by Lemahieu, I. PubMed Articles by Staelens, S. Articles by Lemahieu, I.

GATE as an on-the-fly forward projector in iterative SPECT image reconstruction

¹ ELIS, UGent, Ghent, OVL, Belgium

1683

Objectives: SPECT images are mainly degraded by the improper detection of scattered photons and photons that penetrate the collimator septa. We aim to accelerate the GATE Monte Carlo simulator to become an on-the-fly projector for geometric photons as well as patient scatter without storing the system matrix. It is our goal to incorporate this accelerated GATE version in a unified reconstruction package in order to improve contrast-to-noise, lesion detectability and quantification while maintaining clinical realistic reconstruction times.

Methods: The GATE cluster framework was used to simulate a symmetrical thorax phantom and no variance reduction techniques were applied. Next to that, we recently developed convolution-based forced detection (CFD) with multiple projection sampling for GATE. This fast GATE version was used as a forward projector to generate projection data on-the-fly. Accordingly, the sinogram is stored in memory and passed to the update step of the MLEM algorithm. Image quality is evaluated by measurements of an anthropomorphic torso phantom with a cardiac insert including two solid defects, acquired on a refurbished PRISM 3000XP.

Results: From the simulated data we generated a gold standard subset with only geometric photons and an evaluation subset with only geometric and scattered photons. We compared our model based reconstruction versus standard MLEM on both the evaluation set, the gold standard and also the experimental data. We could conclude an image quality improvement and ongoing work is performed to derive the gain in contrast-to-noise, lesion detectability and quantification.

Conclusions: A high speed version of the GATE Monte Carlo method is presented to act as a forward projector in iterative SPECT image reconstruction. Improved image quality results while the computational load remains clinically realistic.

This Article Services Email this article to a friend Similar articles in this journal Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Staelens, S. Articles by Lemahieu, I. PubMed Articles by Staelens, S. Articles by Lemahieu, I.