Combined constraints for efficient algebraic regularized methods in fully 3D reconstruction

Abstract

SPECT systems incorporate the use of one or more rotating gamma cameras which can be equipped with cone-beam collimators to improve the trade-off between spatial resolution and sensitivity. The geometry of the cone-beam collimators implies that a specific 3D reconstruction algorithm must be applied. Algebraic methods provide the possibility of including the physical characteristics, such as attenuation, Compton scatter and detector response, in the reconstruction process. However, the reconstruction problem is an ill-posed problem which should be regularized. This paper presents a 3D algebraic method that combines three regularizing constraints. These constraints deal respectively with penalizing negative voxels, local noise smoothing and missing data compensation. The results presented were obtained from imaging simulations, phantom data and from a thyroid clinical study of a normal volunteer.

Clinical Applications of Pinhole Single Photon Emission Tomography

“…Various analytic [5][6][7] and algebraic (iterative) [2,[8][9][10][11][12] reconstruction algorithms were developed for PH-SPET. Analytic algorithms are of filtered backprojection type [5][6][7].…”

“…Although other *Address correspondence to this author at the Université de Liège (ULg), Imagerie médicale expérimentale, Institut de Physique, B5, B-4000 Liège1, Belgium; Tel: 32-4-3663705; Fax: 32-4-3663629; E-mail: aseret@ulg.ac.be alternatives exist, only Feldkamp based algorithms were used for reconstruction of human images. Algebraic reconstruction technique (ART) [8], conjugate gradient (CG) [9], maximum likelihood expectation maximisation (MLEM) [2] and median root prior (MRP) [10] iterative algorithms tailored to PH geometry were developed and validated. The accelerating process called ordered subsets (OS) was successfully used with both MLEM (OSEM) [11] and MRP (OS-MRP) [12].…”

“…All these algorithms were first developed for a full 360° rotation of the γ-camera detector head around the target area and with the aim to image small animals. In view of the application of PH-SPET to humans, most of the algorithms, namely Feldkamp type [6,14], CG [9] and MLEM/OSEM [15,16], were adapted to a limited 180°r otation of the detector, this one being eventually slightly (0-30°) tilted. Tilting the detector allows to avoid the shoulder during the camera rotation and therefore to use a shorter pinhole-to-target area distance which results in an improved final resolution.…”

“…Tilting the detector allows to avoid the shoulder during the camera rotation and therefore to use a shorter pinhole-to-target area distance which results in an improved final resolution. It is worth to notice that a single rotation of the detector, even full 360° rotation, does not allow obtaining a complete set of data for exact 3D tomographic reconstruction [9][10][11]. This is a well-known problem when cone-beam geometry is combined with a single detector circular orbit.…”

“…The distortions in the reconstructed images due to incomplete data collection are much more limited with the iterative methods than with the Feldkamp technique and the use of the iterative methods is advocated (Fig. (2)) [2,9,11]. The reconstruction algorithms take into account the PH zooming effect and its dependency towards the collimator-to-target distance in order to deliver reconstructed images whose pixel size is known.…”