EQUIPMENTS: QUALITY CONTROL , Abstracts XIX Brazilian ...

Abstracts XIX Brazilian Congress of Nuclear Medicine

Article Nº AJ02-8

3. EQUIPMENTS: QUALITY CONTROL

3.1	Abe R.
	EFFICIENT TESTS OF COLLIMATORS IN SPECT USING A POINT SOURCE
3.2	Aldighieri,F.	flavia.aldighieri@fleury.com.br
	COMPARARIVE STUDY OF SPECT COLLIMATORS
3.3	Amaury de Castro R	jrze@vol.com.br
	COMPARISON OF THE ATTENUATION CORRECTION BY CHANG METHOD IN COMMERCIAL SOFTWARE

3.1 - EFFICIENT TESTS OF COLLIMATORS IN SPECT USING A POINT SOURCE

Abe R., Oliveira M.A. - The Heart Institute of São Paulo,Robilotta,C.C. - Institute of Physics of the University of Sao Paulo,Brasil

In SPECT, it is assumed that the holes and the septa of collimators are precisely orthogonal to the rotation axis, because the lack of orthogonality causes loss of quality in the reconstructed images. Furthermore, if dual-head cameras are used, the alignment between the detectors has to be garanteed in order to ensure the performance of the imaging system. The misalignment of the 2 heads of a Sophy-DST camera was detected from the analysis of a point source sinogram. When the collimators septa angulation test was done using a point source, the images of the collimator 1 presented concentric ellipses with higher counts in the centre that decreased with distance from this point. On the other hand, the images of the collimator 2 were formed by concentric circles, as one would expect. These simple tests showed that one of the collimators is convergent, that is, the holes are not orthogonal to the axis of rotation. In conclusion, the tests with point source are very simple and efficient for Q.C. of collimators.

3.2 - COMPARARIVE STUDY OF SPECT COLLIMATORS

Aldighieri,F.C;Dimenstein,R.;Ikeda,H.H.;Alonso,G.;Martins,L.R.F. & Serviço de Medicina Nuclear & Laboratório Fleury- SP

flavia.aldighieri@fleury.com.br

The purpose of the current study is a comparative evaluation of the performance of collimators commonly used in SPECT systems. A cylindric Jaszczak et al phantom with inserts of rods and spheres was filled with 851MBq of Tc-99m . We used a single head gamma camera GE Starcam 3200 AC/T, to acquire a total of 128 frames in 128 x 128 matrix , 1.000 kc/frame. The images were performed with both low energy collimators, one of high resolution and other of general purpose (LEHR and LEGP). The radius of rotation were 185 mm and 150 mm (with a head holder). The images were processed using Hanning filter with a cut-off frequency of 1.6 cycles/cm. The slice width for spheres was 5 pixels and 25 pixels for rods. The spatial tomographic resolution results obtained for LEHR with radius of rotation of 185 and 150 mm were, 9,5mm e 7,9mm respectively. For LEGP the spatial tomographic resolution results obtained were 11,1mm e 9,5mm with the same radius of rotation mentioned before . In both studies it was possible to see spheres of 12.7 mm diameter. In conclusion, the LEHR collimator has a better performance than the LEGP, specially in smaller radius of rotation.

3.3 - COMPARISON OF THE ATTENUATION CORRECTION BY CHANG METHOD IN COMMERCIAL SOFTWARE

Amaury de Castro R. e Silva Jr(1,2), Cecil Chow Robilotta(2), Rubens Abe(3); 1. Faculdades Integradas de Guarulhos; 2. Instituto de Física da USP; 3. Instituto do Coração do HCFMUSP.

jrze@vol.com.br

The algorithm proposed by L. T. Chang in 1978, for attenuation correction after tomographic reconstruction, is implemented in most of the commercial image processing systems in Nuclear Medicine. This method uses a correction matrix, calculated by considering a constant linear attenuation coefficient within a region of interest and taking into account the pixel size. The objective of this work is to compare the results of the application of the Chang method in some commercial software and the one that we have developed and incorporated to the PIP system (Portable Image Processing ; Leapfrog Ltd.), made available by IAEA within the ARCAL XXIII project. We used the images of a Jaszczak phantom, acquired from five SPECT cameras of different manufacturers. These images were reconstructed by filtered back-projection using Butterworth filter of order 4 and critical frequency 0.4 Wn. The matrix of correction was created in all systems with m=0.12 cm-1 and with their own pixel size. For evaluation, we used the visual analysis, the count profiles in the images before and after correction, the matrices of correction and the tomographic uniformity. The results indicated that the programmes of the manufacturers analysed in this work are not consistent. Although visually the corrected images do not differ much from each other, they presented significant quantitative differences. The influence of these variations in clinical applications needs to be investigated carefully before adopting this method in the daily routine.