The evaluation and calibration of fan-beam collimators

The aims of this study were (a) to determine the true focal length of a fan-beam collimator and (b) to calibrate image size (mm/pixel) for each collimator to permit inter-comparison of image data acquired on different gamma camera systems. A total of six fan-beam collimators on three dual-head gamma camera systems were evaluated using a set of four cobalt-57 point source markers. The markers were arranged in a line in the transverse plane with a known separation between them. Tomographic images were obtained at three radii of rotation. From reconstructed transaxial images the distance between markers was measured in pixels and used to determine pixel size in mm/pixel. The system value for the focal length of the collimator was modified by up to ±100 mm and transaxial images were again reconstructed. To standardize pixel size between systems, the apparent radius of rotation during a single-photon emission tomography (SPET) acquisition was modified by changes to the effective collimator thickness. SPET images of a 3D brain phantom were acquired on each system and reconstructed using both the original and the modified values of collimator focal length and thickness. Co-registration and subtraction of the reconstructed transaxial images was used to evaluate the effects of changes in collimator parameters. Pixel size in the reconstructed image was found to be a function of both the radius of rotation and the focal length. At the correct focal length, pixel size was essentially independent of the radius of rotation. For all six collimators, true focal length differed from the original focal length by up to 26 mm. These differences in focal length resulted in up to 6% variation in pixel size between systems. Pixel size between the three systems was standardized by altering the value for collimator thickness. Subtraction of the co-registered SPET images of the 3D brain phantom was significantly improved after optimization of collimator parameters, with a 35%–50% reduction in the standard deviation of residual counts in the subtraction images. In conclusion, we have described a simple method for measurement of the focal length of a fan-beam collimator. This is an important parameter on multidetector systems for optimum image quality and where accurate co-registration of SPET to SPET and SPET to MRI studies is required. Received 17 October and in revised form 12 December 1998     

High-resolution cardiac SPET study using fanbeam collimators in infants.

High-resolution three-headed single photon emission computed tomography (SPET) equipped with fan-beam collimators was applied to myocardial perfusion imaging in infants aged from 1 to 11 months (n = 5). A tabletop designed specifically for infants was fixed on the SPET couch to reduce the radius of camera rotation to 13.2 cm. Significant improvement in resolution was achieved with the fan-beam collimators compared to parallel-hole high-resolution collimators. With the administration of approximately 37 MBq (26-44 MBq) 201Tl, 5 min acquisition time was possible for SPET imaging, which provided good image quality in all patients. Thus, a smaller administration dose is possible within a practical short acquisition time. High-resolution fan-beam SPET imaging can be a routine diagnostic method for heart disease in newborn babies and infants.     

Acquisition of linograms in SPET: implementation and benefits

Compared with other tomographic modalities, single-photon emission tomography (SPET), the most widely used tomographic modality in nuclear medicine, suffers from poor quality image since the collimator stops 99.99% of the emitted gamma rays reaching the detector. This paper describes a new SPET acquisition modality using a very short focal length (12.5 cm) fan-beam collimator and a very short transverse field of view detector (25 cm). The detector moves along at least two linear orthogonal orbits in such a way that the focal line travels through the source target. This linear orbit acquisition (LOrA) generates linograms forming a complete set of tomographic data, i.e. sufficient to exactly reconstruct the activity map using a modified filtered back-projection algorithm. In contrast to the classical fan-beam tomography, truncation is not a problem, even when the source transverse size is much larger than the detector transverse size. When the collimator hole length/diameter ratio is adapted to obtain a spatial resolution similar to that of classical SPET, LOrA SPET offers an improvement in sensitivity by a factor of about 2.5 for a 20-cm source size. This improvement is achieved with a detector that is half as large, and thus half as expensive. As with classical fan-beam SPET, the sensitivity increases further if the target size decreases. When fitting the collimator to obtain a similar sensitivity to that of classical SPET, a significant improvement in spatial resolution is obtained.     

Collimator design for single photon emission tomography.

We discuss recent trends in collimator design and technology, with emphasis on theoretical and practical issues of importance for single photon emission tomography (SPET). The well-known imaging performance parameters of parallel-hole collimators are compared with those of fan-beam collimators, which have enjoyed considerable success in recent years, particularly for brain SPET. We review a simplistic approach to the collimator optimization problem, as well as more sophisticated "task-dependent" treatments and important considerations for SPET collimator design. Practical guidance is offered for understanding trade-offs that must be considered for clinical imaging. Finally, selective comparisons among different SPET systems and collimators are presented for illustrative purposes.     

Simulating technetium-99m cerebral perfusion studies with a three-dimensional Hoffman brain phantom: Collimator and filter selection in SPECT neuroimaging

The choice of a collimator and the selection of a filter can affect the quality of clinical SPECT images of the brain. The compromises that 4 different collimators make between spatial resolution and sensitivity were studied by imaging a three-dimensional Hoffman brain phantom. The planar data were acquired with each collimator on a three-headed SPECT system and were reconstructed with both a standard Butterworth filter and a Wiener pre-filter. The reconstructed images were then evaluated by specialists in nuclear medicine and were also quantitatively analyzed with specific regions of interest (ROI) in the brain. All observers preferred the Wiener filter reconstructed images regardless of the collimator used to acquire the planar images. With this filter, the ultrahigh-resolution fan-beam collimator was the most subjectively preferable and quantitatively produced the highest contrast ratios. The findings support suggestions that higher resolution collimators are preferable to higher sensitivity collimators, and indicate that fan-beam collimators are preferable to parallel-hole collimators for clinical SPECT studies of cerebral perfusion. The results also suggest that the Wiener filter enhances the quality of SPECT brain images regardless of which collimator is used to acquire the data.     

Evaluation of SPET quantification of simultaneous emission and transmission imaging of the brain using a multidetector SPET system with the TEW scatter compensation method and fan-beam collimation

A gamma camera system which is able to acquire simultaneous single-photon emission tomographic (SPET) data and gamma ray transmission computed tomography (TCT) data for brain study using external rod sources and fan-beam collimators was developed and evaluated. Since the three external rod sources were located at the focal points of fan-beam collimators, which also happened to be the apexes of the equilateral triangle defined by the three detectors, simultaneous SPET and TCT scan could be performed using a 120° shared scan. Therefore, the proposed system required less than one third of the scanning time of a single-head system. Since the combination of rod sources and fan-beam collimators decreased the scatter component in transmission data without a slit collimator for each rod source, the radioactivity of the rod source was less than one-tenth of the previous investigations. For evaluation, we used two isotopes, thallium-201 for TCT and technetium-99m for SPET. The cross-contamination of transmission and emission was well compensated using the triple energy window (TEW) method. In a separate TCT scan, the measured attenuation coefficient of²⁰¹Tl for water was 0.19±0.01 cm^–1, while in a simultaneous scan, it was 0.20±0.01 cm^–1. The measured attenuation coefficient for water agreed well with the narrow-beam (theoretical) value of 0.187 cm^–1. In SPET images, scatter compensation was also performed using the TEW method and attenuation compensation was done using the measured attenuation map. The results showed the feasibility of simultaneous SPET and TCT scanning using the TEW method to obtain quantitative SPET images.     

Aperture correction with an asymmetrically trimmed Gaussian weight in SPECT with a fan-beam collimator

Objectives

The aim of the study is to improve the spatial resolution of SPECT images acquired with a fan-beam collimator.

Methods

The aperture angle of a hole in the fan-beam collimator depends on the position of the collimator. To correct the aperture effect in an iterative image reconstruction, an asymmetrically trimmed Gaussian weight was used for a model. To confirm the validity of our method, point source phantoms and brain phantom were used in the simulation, and we applied the method to the clinical data.

Results

The results of the simulation showed that the spatial resolution of point sources improved from about 6 to 2 pixels full width at half maximum, and the corrected point sources were isotropic. The results of the simulation with the brain phantom showed that our proposed method could improve the spatial resolution of the phantom, and our method was effective for different fan-beam collimators with different focal lengths. The results of clinical data showed that the quality of the reconstructed image was improved with our proposed method.

Conclusions

Our proposed aperture correction method with the asymmetrically trimmed Gaussian weighting function was effective in improving the spatial resolution of SPECT images acquired with the fan-beam collimator.     

Collimator design for single photon emission tomography

We discuss recent trends in collimator design and technology, with emphasis on theoretical and practical issues of importance for single photon emission tomography (SPET). The well-known imaging performance parameters of parallel-hole collimators are compared with those of fan-beam collimators, which have enjoyed considerable success in recent years, particularly for brain SPET. We review a simplistic approach to the collimator optimization problem, as well as more sophisticated task-dependent treatments and important considerations for SPET collimator design. Practical guidance is offered for understanding trade-offs that must be considered for clinical imaging. Finally, selective comparisons among different SPET systems and collimators are presented for illustrative purposes. Offprint requests to: S.C. Moore     

Physical assessment of the GE/CGR Neurocam and comparison with a single rotating gamma-camera

The GE/CGR Neurocam is a triple-headed single photon emission tomography (SPET) system dedicated to multi-slice brain tomography. We have assessed its physical performance in terms of sensitivity and resolution, and its clinical efficacy in comparison with a modern, single rotating gamma-camera (GE 400XCT). Using a water-filled cylinder containing technetium-99m, the tomographic volume sensitivity of the Neurocam was 30.0 and 50.7 kcps/MBq · ml · cm for the high-resolution (HR) and general-purpose (GP) collimators, respectively; the corresponding values for the single rotating camera were 7.6 and 12.8 kcps/(MBq/ml)/cm. Tomographic resolution was measured in air and in water. In air, the Neurocam resolution at the cente of the field-of-view (FOV) is 9.0 and 10.7 mm full width at half-maximum (FWHM) with the HR and GP collimators, respectively, and is isotropic in the three orthogonal planes; the resolution of the GE 400XCT with 13 cm radius of rotation is 10.3 and 11.7 mm, respectively. For the Neurocam with the HR collimator, the transaxial FWHM values in water were 9.7 mm at the centre and 9.5 mm radial (6.6 mm tangential) at 8 cm from the centre. The physical characteristics of the Neurocam enable the routine acquisition of brain perfusion data with technetium-99m hexamethyl-propylene amine oxime (^99mTc-HMPAO) in about 14 min, yielding better image quality than with a single rotating camera in 40 min. Offprint requests to: K. Kouris     

Characterisation of fan-beam collimators

Fan-beam collimators offer a good balance between resolution and noise. The collimator response may be included in iterative reconstruction algorithms in order to improve single-photon emission tomography (SPET) resolution. To this end, accurate determination of the focal region and characterisation of the collimator response as a function of the source co-ordinates must be performed. In this paper, a method to characterise fanbeam collimators is evaluated. First, we calculated the real focal region and the accuracy of the collimator convergence. Then, we confirmed the hypothesis that Gaussian distributions adequately fit the collimator responses, although no individualised treatment was performed for the tails of detector response which are associated with scattering and septal penetration. Finally, analytical functions were used to model the resolution and sensitivity. The parameter values in these functions were obtained from experimental measures by non-linear regression fitting. Our findings show differences of 1.43% between nominal and real focal length and standard deviations of 2.5 mm in the x-direction and 7.1 mm in the y-direction for the focal convergence. The correlation coefficients between experimental and predicted values were 0.994 for resolution and 0.991 for sensitivity. As a consequence, the proposed method can be used to characterise the collimator response.     

SPECT using a specially designed cone beam collimator

A specially designed high resolution converging collimator having a focal length of 50 cm has been evaluated for cone beam single photon emission computed tomography (SPECT). The focal region was investigated by imaging a point source placed at the expected focal point and along the central ray of the collimator in front of and behind the focal point. Technetium-99m point source sensitivities measured in air at 5, 10, 15, and 20 cm from the collimator surface are 4.2, 5.5, 7.3, and 10.5 cts.sec-1.microCi-1 when used with a single camera SPECT system. A commercially available parallel hole collimator, with similar resolution characteristics has a measured sensitivity of 3.3 cts.sec-1.microCi-1. Volume sensitivities of 9,780 and 4,945 (cts.sec-1)/(microCi.ml-1) were measured for the cone beam and parallel hole collimators, respectively, using a 17-cm-diameter spherical source. Reconstructed spatial resolution (FWHM) on the axis-of-rotation ranged between 10 and 11 mm for both collimators when the radius of rotation was equal to 15 cm. Using equal acquisition times SPECT images of phantoms scanned with the cone beam collimator were visually improved compared with images acquired using the parallel hole collimator. These results demonstrate that a factor of 2 improvement in volume sensitivity can be demonstrated with a cone beam collimator compared with a commercially available parallel hole collimator. Further improvements are possible using shorter focal lengths, astigmatic focusing, and larger field of view cameras.     

Physical assessment of the GE/CGR Neurocam and comparison with a single rotating gamma-camera.

The GE/CGR Neurocam is a triple-headed single photon emission tomography (SPET) system dedicated to multi-slice brain tomography. We have assessed its physical performance in terms of sensitivity and resolution, and its clinical efficacy in comparison with a modern, single rotating gamma-camera (GE 400XCT). Using a water-filled cylinder containing technetium-99m, the tomographic volume sensitivity of the Neurocam was 30.0 and 50.7 kcps/MBq.ml.cm for the high-resolution (HR) and general-purpose (GP) collimators, respectively; the corresponding values for the single rotating camera were 7.6 and 12.8 kcps/(MBq/ml)/cm. Tomographic resolution was measured in air and in water. In air, the Neurocam resolution at the cente of the field-of-view (FOV) is 9.0 and 10.7 mm full width at half-maximum (FWHM) with the HR and GP collimators, respectively, and is isotropic in the three orthogonal planes; the resolution of the GE 400XCT with 13 cm radius of rotation is 10.3 and 11.7 mm, respectively. For the Neurocam with the HR collimator, the transaxial FWHM values in water were 9.7 mm at the centre and 9.5 mm radial (6.6 mm tangential) at 8 cm from the centre. The physical characteristics of the Neurocam enable the routine acquisition of brain perfusion data with technetium-99m hexamethyl-propylene amine oxime (99mTc-HMPAO) in about 14 min, yielding better image quality than with a single rotating camera in 40 min.     

Evaluation of optimally designed planar-concave collimators in single-photon emission tomography

The imaging properties of optimally designed planar-concave (PC) collimators were evaluated by means of Monte Carlo simulations. The evaluation was done with respect to total system spatial resolution and the overall image noise distribution in single-photon emission tomography. The results showed that the non-isotropy with PC collimators, assessed by the ratio of the full-width at half-maximum in the radial and tangential directions, was reduced by about 60% as compared with a conventional parallel-hole collimator for sources located 200 mm away from the centre of rotation. Furthermore, the image noise distribution along the object radius became more uniform when the curved collimator was used. The maximum increase in noise due to use of the curved collimator was about 45% close to the edge of the phantom, where the hole length was about 3 times longer. We also showed with Monte Carlo simulations that the spatial resolution of the lateral cortex when using the curved collimator was significantly improved due to improved radial resolution. Received 1 May and in revised form 4 August 1997     

Choice of collimator for cardiac SPET when resolution compensation is included in iterative reconstruction

In clinical cardiac single-photon emission tomography (SPET) studies, collimators of different spatial resolution and geometric efficiency are available for imaging. In selecting the appropriate collimator for clinical use, there is a trade-off between spatial resolution, which can limit the contrast of the reconstructed image, and detection efficiency, which determines the noise in the image. Our objective was to assess which collimator is best suited for cardiac SPET when reconstruction is performed with and without compensation for distance-dependent resolution (CDR). The dynamic MCAT thorax phantom was used to simulate 180 degree technetium-99m cardiac data, acquired using either a general-purpose (GP) or high-resolution (HR) collimator. For GP and HR, the resolution at 15 cm was 11.5 mm and 9.5 mm respectively, and the corresponding relative efficiency was 1.0 and 0.52 respectively. Distance-dependent resolution, attenuation and noise were included in the projection data; scatter was not included. Ordered subsets expectation maximisation reconstruction (subset size 4) was performed with and without CDR. Results were evaluated by comparing the myocardial recovery coefficient and contrast between myocardium and ventricle relative to the original phantom, each plotted for different noise levels corresponding to increasing iteration number. The study demonstrated that, without CDR, HR gave the best results. However, for any given noise level with CDR, GP gave superior recovery and contrast. These findings were confirmed in a physical phantom study. Results suggest that improved reconstruction can be achieved using a GP collimator in combination with resolution compensation.     

Choice of collimator for cardiac SPET when resolution compensation is included in iterative reconstruction

In clinical cardiac single-photon emission tomography (SPET) studies, collimators of different spatial resolution and geometric efficiency are available for imaging. In selecting the appropriate collimator for clinical use, there is a trade-off between spatial resolution, which can limit the contrast of the reconstructed image, and detection efficiency, which determines the noise in the image. Our objective was to assess which collimator is best suited for cardiac SPET when reconstruction is performed with and without compensation for distance-dependent resolution (CDR). The dynamic MCAT thorax phantom was used to simulate 180° technetium-99m cardiac data, acquired using either a general-purpose (GP) or high-resolution (HR) collimator. For GP and HR, the resolution at 15 cm was 11.5 mm and 9.5 mm respectively, and the corresponding relative efficiency was 1.0 and 0.52 respectively. Distance-dependent resolution, attenuation and noise were included in the projection data; scatter was not included. Ordered subsets expectation maximisation reconstruction (subset size 4) was performed with and without CDR. Results were evaluated by comparing the myocardial recovery coefficient and contrast between myocardium and ventricle relative to the original phantom, each plotted for different noise levels corresponding to increasing iteration number. The study demonstrated that, without CDR, HR gave the best results. However, for any given noise level with CDR, GP gave superior recovery and contrast. These findings were confirmed in a physical phantom study. Results suggest that improved reconstruction can be achieved using a GP collimator in combination with resolution compensation.     

Effects of crystal pixel size and collimator geometry on the performance of a pixelated crystal gamma-camera using Monte Carlo simulation

Dedicated γ-cameras based on pixelated scintillators have long been used for breast tumor imaging. Intercrystal scattering (ICS) increases the background counting rate and degrades the image quality when small crystal pixels are used. Because of the small size of applied collimators, scattered radiation and septal penetration are high, and therefore collimator characteristics must be carefully considered. In our study, we investigated the influence of ICS events on position-detection accuracy (PDA) for pixelated crystals and the effects of different geometries of hexagonal-hole collimators on the performance of these cameras, using Monte Carlo simulation to optimize camera design. The arrays of thallium-doped cesium iodide detectors with different pixel dimensions that had been exposed to 140-keV photons of isotropic point source, 50 mm from the collimator surface, were simulated. Hexagonal-hole collimators were 10.5, 15, and 21 mm long. The septal thickness varied from 0.1 to 0.5 mm, with 3 different hole diameters. The results confirmed that by increasing the crystal pixel size, ICS was decreased and change of detection efficiency was negligible, but PDA, contrast-to-noise ratio, and spatial resolution (full width at half maximum) were increased. Our experiences confirmed that 2 × 2 mm was an optimum crystal pixel size, especially for a lower ICS fraction and an appropriate full width at half maximum. Because collimators are the limiting factor for spatial resolution and sensitivity, careful collimator design is of great importance.     

Simultaneous acquisition of iodine-123 emission and technetium-99m transmission data for quantitative brain single-photon emission tomographic imaging

The aim of this study was to obtain quantitative iodine-123 brain single-photon emission tomographic (SPET) images with scatter and attenuation correction. We used a triple-headed SPET gamma camera system equipped with fan-beam collimators with a technetium-99m line transmission source placed at one of the focal lines of the fan-beam collimators. Four energy windows were employed for data acquisition: (a) 126–132 keV, (b) 132–143 keV, (c) 143–175 keV and (d) 175–186 keV. A simultaneous transmission-emission computed tomography scan (TCT-ECT) was carried out for a brain phantom containing ¹²³I solution. The triple energy window scatter correction was applied to the ¹²³I ECT data measured by means of the windows (b), (c) and (d) acquired by two detectors. Attenuation maps were reconstructed from ^99mTc TCT data measured by means of the windows (a), (b) and (c) acquired by one detector. Chang’s iterative attenuation correction method using the attenuation maps was applied to the ¹²³I ECT images. In the phantom study cross-calibrated SPET values obtained with the simultaneous mode were almost equal to those obtained with the sequential mode, and they were close to the true value, within an error range of 5.5%. In the human study corrected images showed a higher grey-to-white matter count ratio and relatively higher uptake in the cerebellum, basal ganglia and thalamus than uncorrected images. We conclude that this correction method provides improved quantification and quality of SPET images and that the method is clinically practical because it requires only a single scan with a ^99mTc external source. Received 6 June and in revised form 27 July 1998     

The influence of collimators on SPECT center of rotation measurements: artifact generation and acceptance testing

Misalignment between the electronic and mechanical axes of rotation will result in artifact generation and image degradation during single photon emission computed tomography (SPECT) reconstruction. Acceptance and quality control testing procedures have not emphasized the variability in center of rotation (COR) measurements caused by collimators and the need to verify uniformity across the full collimator field of view (FOV). Variation from the mean COR across the FOV was tested in four different collimators using multiple point source acquisitions. The mean COR was different for each collimator and two of the four had a greater than 0.5 pixel difference from the mean COR on some area of the FOV. This variation makes these collimators unacceptable for SPECT acquisition. Thus, initial acceptance testing of SPECT collimators should verify a uniform COR across the full FOV and collimators with a variability from the mean COR greater than 0.5 pixels should be rejected.     

Development of a converging collimator for thyroid scintigraphy

Thyroid scintigraphy is generally performed using a pin-hole or a parallel-hole collimator. The former yields a high resolution image, however, the acquisition time prolongs because the collimation system magnifies a thyroid image. The latter has somewhat low resolution, because the collimation system cannot magnify the thyroid image and it is difficult to close the collimator surface to a patient thyroid. The defects of the above two collimators are distortion or blurs at the thyroid edge of the acquired image. To improve the image quality of the thyroid scintigram, we made computer simulations concerning contrast resolution in comparison of a parallel-hole, pin-hole and converging collimator. The simulations results showed that the contrast resolution by the converging collimator is superior to the other collimators. According to the results, we developed a specially-designed converging collimator (thickness: 100 mm, focal length: 195 mm). We cut an edge of the collimator to decrease the distance between the neck and the collimator surface. Under almost the same conditions for spatial resolution and sensitivity, we compared the effectiveness of the converging collimator with that of the other two collimators; the pin-hole type [medium resolution] and the parallel-hole type [high resolution]. As a result of experiments using a phantom, the converging collimator showed less distortion compared with the other collimators at the deep area of the thyroid phantom.(ABSTRACT TRUNCATED AT 250 WORDS)     

Attenuation correction using asymmetric fanbeam transmission CT on two-head SPECT system

For transmission computed tomography (TCT) systems using a centered transmission source with a fan-beam collimator, the transmission projection data are truncated. To achieve sufficiently large imaging field of view (FOV), we have designed the combination of an asymmetric fan-beam (AsF) collimator and a small uncollimated sheet-source for TCT, and implemented AsF sampling on a two-head SPECT system. The purpose of this study is to evaluate the feasibility of our TCT method for quantitative emission computed tomography (ECT) in clinical application. Sequential Tc-99m transmission and Tl-201 emission data acquisition were performed in a cardiac phantom (30 cm in width) with a myocardial chamber and a patient study. Tc-99m of 185 MBq was used as the transmission source. Both the ECT and TCT images were reconstructed with the filtered back-projection method after scatter correction with the triple energy window (TEW) method. The attenuation corrected transaxial images were iteratively reconstructed with the Chang algorithm utilizing the attenuation coefficient map computed from the TCT data. In this AsF sampling geometry, an imaging FOV of 50 cm was yielded. The attenuated regions appeared normal on the scatter and attenuation corrected (SAC) images in the phantom and patient study. The good quantitative accuracy on the SAC images was also confirmed by the measurement of the Tl-201 radioactivity in the myocardial chamber in the phantom study. The AsF collimation geometry that we have proposed in this study makes it easy to realize TCT data acquisition on the two-head SPECT system and to perform quantification on Tl-201 myocardial SPECT.