COMPARATIVE STUDY BETWEEN COLLIMATORS COMMONLY USED IN SPECT

Article Nº AJ03-6

 

SUMMARY

The current study is a comparative evaluation of the performance of collimators commonly used in SPECT. The results show the difference in tomographic spatial resolution using a parallel hole high resolution collimator and the parallel hole general purpose collimator and how the distance between the collimator surface and the phantom influences the quality of the image.

Key words: SPECT systems, Jaszczack phantom, resolution, uniformity.

INTRODUCTION

The purpose of the current study is a comparative evaluation of tomographic spatial resolution of two collimators commonly used in SPECTS systems . This study was performed using a Jaszczak phantom , which is a cylindrical acrylic phantom with inserts of rods and spheres representing "cold" areas, and a region of uniformity⁽¹⁾. The phantom study is able to show the performance of the single photon emission computed tomography system (SPECT) regarding spacial resolution and uniformity. The tomographic resolution is determined through the observation of the cold rods section . The section without inserts, filled with water and 851 MBq of Tc99m (TcO₄^_) is used as a reference only for the visual uniformity analysis. The test with the phantom must be part of a quality control routine program in SPECT systems, which enables the user to evaluate the performance of the whole system every six months.

MATERIAL AND METHODS

For this study an acrylic cylindric Jaszczak ⁽¹⁾ phantom, with insert of spheres measuring 9.5, 12.7, 15.9, 19.1, 25.4 and 31.0 mm diameter and insert of rods measuring 4.8, 6.4, 7.9, 9.5, 11.1 and 12.7 mm diameter was used. The phantom was filled with water and 851 MBq activity of Tc99m(TcO₄^_ ) were uniformly distributed. The images were acquired with a single-head GE Starcam 3200 AC/T gamma camera having 3.1% integral uniformity, 2.3% differential uniformity in the CFOV, 11.07% energy resolution and center of rotation deviation of 0.4 mm. The photopeak was set at 140 keV with a 20% centered energy window .

Image Acquisition:

One hundred and twenty eight frames in 128 x 128 matrix were acquired with 1000 kc per frame in word mode and clockwise circular 360^o orbit. The collimators used were the parallel holes, Low Energy High Resolution (LEHR) and Low Energy General Purpose (LEGP). The radius of rotation were 185 and 150 mm (a head holder was used in order to place the detector as close as possible to the phantom). A bubble level was used to ensure that the phantom was levelled. It was also adjusted so that the phantom would lie parallel the axis of rotation of the gamma camera. The height of the scan table was adjusted so that the distance from the collimator to the middle of the phantom was the same, with the detector at 0^o and at 180^o .

Image Processing:

The images were processed using Hanning filter with cut-off frequency of 1.6 cycles.cm^-1, slice thickness of 5 pixels in the spheres section , 25 pixels at the rods section and 2 pixels at the uniform section. A 100 million uniformity correction map and an attenuation coefficient of 0,11 ^-1 cm correction ⁽²⁾ were applied.

 

RESULTS:

  1. Rods section

2. Spheres section

The smallest sphere observed was the 12.7 mm in diameter in both collimators

 

3. Uniform section

Absence of ring artifacts and uniform radioactivity distribution.

Figure1

( click = zoom)

Transaxial slices of rods and spheres sections. The images were obtained with LEGP collimator with 150 and 185 mm radius of rotation.

Figure 2

( click = zoom)

– Transaxial slices of rods and spheres sections. The images were obtained with LEHR collimator with 150 and 185 mm radius of rotation.

Figure 3

( click = zoom)

Transaxial slices of the uniform section.

 

DISCUSSION

It is well known that the hight resolution collimators increase the ability to visualize small lesions and structure in a lesion provided that count density is satisfactory, the best will be the detection ability of the spECT system. The results show that the LEHR collimator offers the best resolution when compared with the LEGP collimator, with a difference of 1.6 mm between them. The best resolution was obtained with the LEHR collimator at the smallest radius of rotation used in this study: 150 mm.

Ring artifacts were not seen at the reconstructed images due to the application of an uniformity correction map with a large total count (100 million kc).

 

CONCLUSION

The "cost-effectiveness" of choosing collimators should be evaluated before the decision for better resolution or better sensitivity for each particular study.

This study showed that the LEHR collimator can contribute effectively to a better resolution than the LEGP collimator. Although this conclusion is already available in the medical literature, mathematical data were obtained for the equipment GE Starcam 3200 AC/T. These data allow other users to have a better evaluation in comparison to other equipment commonly used in Nuclear Medicine.

 

REFERENCES

1. Data spectrum’s SPECT user’s manual; Data Spectrum Corporation; Aug 1996.

2. O’Connor, M.K.: The Mayo Clinic Manual of Nuclear Medicine; 1996.

3. Muellehener et al Physics in Medicine and Biology 30:163, 1985

4. Madsen et al. Physics in Medicine and Biology 37: 1625, 1992