Margaret E. Daube-Witherspoon

Benefit of Time-of-Flight in PET: Experimental and Clinical Results

Significant improvements have made it possible to add the technology of time-of-flight (TOF) to improve PET, particularly for oncology applications. The goals of this work were to investigate the benefits of TOF in experimental phantoms and to determine how these benefits translate into improved performance for patient imaging. Methods: In this study we used a fully 3-dimensional scanner with the scintillator lutetium-yttrium oxyorthosilicate and a system timing resolution of ∼600 ps. The data are acquired in list-mode and reconstructed with a maximum-likelihood expectation maximization algorithm; the system model includes the TOF kernel and corrections for attenuation, detector normalization, randoms, and scatter. The scatter correction is an extension of the model-based single-scatter simulation to include the time domain. Phantom measurements to study the benefit of TOF include 27-cm- and 35-cm-diameter distributions with spheres ranging in size from 10 to 37 mm. To assess the benefit of TOF PET for clinical imaging, patient studies are quantitatively analyzed. Results: The lesion phantom studies demonstrate the improved contrast of the smallest spheres with TOF compared with non-TOF and also confirm the faster convergence of contrast with TOF. These gains are evident from visual inspection of the images as well as a quantitative evaluation of contrast recovery of the spheres and noise in the background. The gains with TOF are higher for larger objects. These results correlate with patient studies in which lesions are seen more clearly and with higher uptake at comparable noise for TOF than with non-TOF. Conclusion: TOF leads to a better contrast-versus-noise trade-off than non-TOF but one that is difficult to quantify in terms of a simple sensitivity gain improvement: A single gain factor for TOF improvement does not include the increased rate of convergence with TOF nor does it consider that TOF may converge to a different contrast than non-TOF. The experimental phantom results agree with those of prior simulations and help explain the improved image quality with TOF for patient oncology studies.

An Iterative Image Space Reconstruction Algorthm Suitable for Volume ECT

The trend in the design of scanners for positron emission computed tomography has traditionally been to improve the transverse spatial resolution to several millimeters while maintaining relatively coarse axial resolution (1-2 cm). Several scanners are being built with fine sampling in the axial as well as transverse directions, leading to the possibility of the true volume imaging. The number of possible coincidence pairs in these scanners is quite large. The usual methods of image reconstruction cannot handle these data without making approximations. It is computationally most efficient to reduce the size of this large, sparsely populated array by back-projecting the coincidence data prior to reconstruction. While analytic reconstruction techniques exist for back-projected data, an iterative algorithm may be necessary for those cases where the point spread function is spatially variant. A modification of the maximum likelihood algorithm is proposed to reconstruct these back-projected data. The method, the iterative image space reconstruction algorithm (ISRA), is able to reconstruct data from a scanner with a spatially variant point spread function in less time than other proposed algorithms. Results are presented for single-slice data, simulated and actual, from the PENN-PET scanner.

Intensity-dependent regional cerebral blood flow during 1-Hz repetitive transcranial magnetic stimulation (rTMS) in healthy volunteers studied with H215O positron emission tomography: i. effects of primary motor cortex rTMS

BACKGROUND Repetitive transcranial magnetic stimulation (rTMS) affects the excitability of the motor cortex and is thought to influence activity in other brain areas as well. We combined the administration of varying intensities of 1-Hz rTMS of the motor cortex with simultaneous positron emission tomography (PET) to delineate local and distant effects on brain activity. METHODS Ten healthy subjects received 1-Hz rTMS to the optimal position over motor cortex (M1) for producing a twitch in the right hand at 80, 90, 100, 110, and 120% of the twitch threshold, while regional cerebral blood flow (rCBF) was measured using H(2)(15)O and PET. Repetitive transcranial magnetic stimulation (rTMS) was delivered in 75-pulse trains at each intensity every 10 min through a figure-eight coil. The regional relationship of stimulation intensity to normalized rCBF was assessed statistically. RESULTS Intensity-dependent rCBF increases were produced under the M1 stimulation site in ipsilateral primary auditory cortex, contralateral cerebellum, and bilateral putamen, insula, and red nucleus. Intensity-dependent reductions in rCBF occurred in contralateral frontal and parietal cortices and bilateral anterior cingulate gyrus and occipital cortex. CONCLUSIONS This study demonstrates that 1-Hz rTMS delivered to the primary motor cortex (M1) produces intensity-dependent increases in brain activity locally and has associated effects in distant sites with known connections to M1.

A system for the 3D reconstruction of retracted-septa PET data using the EM algorithm

We have implemented the EM reconstruction algorithm for volume acquisition from current generation retracted-septa PET scanners. Although the software was designed for a GE Advance scanner, it is easily adaptable to other 3D scanners. The reconstruction software was written for an Intel iPSC/860 parallel computer with 128 compute nodes. Running on 32 processors, the algorithm requires approximately 55 minutes per iteration to reconstruct a 128/spl times/128/spl times/35 image. No projection data compression schemes or other approximations were used in the implementation. Extensive use of EM system matrix (C/sub ij/) symmetries (including the 8-fold in-plane symmetries, 2-fold axial symmetries, and axial parallel line redundancies) reduces the storage cost by a factor of 188. The parallel algorithm operates on distributed projection data which are decomposed by base-symmetry angles. Symmetry operators copy and index the C/sub ij/ chord to the form required for the particular symmetry. The use of asynchronous reads, lookup tables, and optimized image indexing improves computational performance. >

An approximation formula for the variance of PET region-of-interest values

An approximation formula for the variance of positron emission tomography (PET) region-of-interest (ROI) values has been developed, implemented, and evaluated. This formula does not require access to the original projection data and is therefore convenient for routine use. The formula was derived by applying successive approximations to the filtered-backprojection reconstruction algorithm. ROI variance is estimated from the product of mean pixel variance within the region and a term accounting for the intercorrelation of all pixel pairs inside the region. The formula accounts for radioactivity distribution, attenuation, randoms, scatter, deadtime, detector normalization, scan length, decay, and reconstruction filter. The algorithm was tested by comparison to the exact ROI variance as calculated with Huesmans algorithm. Tests with scan data from phantoms, animals, and humans obtained on the Scanditronix PC2048-15B tomograph showed the approximation formula to be accurate to within +/-10%

The imaging performance of a LaBr3-based PET scanner

A prototype time-of-flight (TOF) PET scanner based on cerium-doped lanthanum bromide [LaBr(3) (5% Ce)] has been developed. LaBr(3) has a high light output, excellent energy resolution and fast timing properties that have been predicted to lead to good image quality. Intrinsic performance measurements of spatial resolution, sensitivity and scatter fraction demonstrate good conventional PET performance; the results agree with previous simulation studies. Phantom measurements show the excellent image quality achievable with the prototype system. Phantom measurements and corresponding simulations show a faster and more uniform convergence rate, as well as more uniform quantification, for TOF reconstruction of the data, which have 375 ps intrinsic timing resolution, compared to non-TOF images. Measurements and simulations of a hot and cold sphere phantom show that the 7% energy resolution helps to mitigate residual errors in the scatter estimate because a high energy threshold (>480 keV) can be used to restrict the amount of scatter accepted without a loss of true events. Preliminary results with incorporation of a model of detector blurring in the iterative reconstruction algorithm not only show improved contrast recovery but also point out the importance of an accurate resolution model of the tails of LaBr(3)s point spread function. The LaBr(3) TOF-PET scanner demonstrated the impact of superior timing and energy resolutions on image quality.

Evaluation of 3D reconstruction algorithms for a small animal PET camera

The use of paired, opposing position-sensitive phototube scintillation cameras (SCs) operating in coincidence for small animal imaging with positron emitters is currently under study. Because of the low sensitivity of the system even in 3D mode and the need to produce images with high resolution, it was postulated that a 3D expectation maximization (EM) reconstruction algorithm might be well suited for this application. We investigated four reconstruction algorithms for the 3D SC PET camera: 2D filtered back-projection (FBP), 2D ordered subset EM (OSEM), 3D reprojection (3DRP), and 3D OSEM. Noise was assessed for all slices by the coefficient of variation in a simulated uniform cylinder. Resolution was assessed from a simulation of 15 point sources in the warm background of the uniform cylinder. At comparable noise levels, the resolution achieved with OSEM (0.9-mm to 1.2-mm) is significantly better than that obtained with FBP or 3DRP (1.5-mm to 2.0-mm.) Images of a rat skull labeled with /sup 18/F-fluoride suggest that 3D OSEM can improve image quality of a small animal PET camera.

Imaging performance of a-PET: a small animal PET camera

The evolution of positron emission tomography (PET) imaging for small animals has led to the development of dedicated PET scanner designs with high resolution and sensitivity. The animal PET scanner achieves these goals for imaging small animals such as mice and rats. The scanner uses a pixelated Anger-logic detector for discriminating 2 /spl times/ 2 /spl times/ 10 mm/sup 3/ crystals with 19-mm-diameter photomultiplier tubes. With a 19.7-cm ring diameter, the scanner has an axial length of 11.9 cm and operates exclusively in three-dimensional imaging mode, leading to very high sensitivity. Measurements show that the scanner design achieves a spatial resolution of 1.9 mm at the center of the field-of-view. Initially designed with gadolinium orthosilicate but changed to lutetium-yttrium orthosilicate, the scanner now achieves a sensitivity of 3.6% for a point source at the center of the field-of-view with an energy window of 250-665 keV. Iterative image reconstruction, together with accurate data corrections for scatter, random, and attenuation, are incorporated to achieve high-quality images and quantitative data. These results are demonstrated through our contrast recovery measurements as well as sample animal studies.

Noise characteristics of 3-D and 2-D PET images

The authors analyzed the noise characteristics of two-dimensional (2-D) and three-dimensional (3-D) images obtained from the GE Advance positron emission tomography (PET) scanner. Three phantoms were used: a uniform 20-cm phantom, a 3-D Hoffman brain phantom, and a chest phantom with heart and lung inserts. Using gated acquisition, the authors acquired 20 statistically equivalent scans of each phantom in 2-D and 3-D modes at several activity levels. From these data, they calculated pixel normalized standard deviations (NSDs), scaled to phantom mean, across the replicate scans, which allowed them to characterize the radial and axial distributions of pixel noise. The authors also performed sequential measurements of the phantoms in 2-D and 3-D modes to measure noise (from interpixel standard deviations) as a function of activity. To compensate for the difference in axial slice width between 2-D and 3-D images (due to the septa and reconstruction effects), they developed a smoothing kernel to apply to the 2-D data. After matching the resolution, the ratio of image-derived NSD values (NSD/sub 2D//NSD/sub 3D/)/sup 2/ averaged throughout the uniform phantom was in good agreement with the noise equivalent count (NEC) ratio (NEC/sub 3D//NEC/sub 2D/). By comparing different phantoms, the authors showed that the attenuation and emission distributions influence the spatial noise distribution. The estimates of pixel noise for 2-D and 3-D images produced here can be applied in the weighting of PET kinetic data and may be useful in the design of optimal dose and scanning requirements for PET studies. The accuracy of these phantom-based noise formulas should be validated for any given imaging situation, particularly in 3-D, if there is significant activity outside the scanner field of view.

Post-injection transmission attenuation measurements for PET

The acquisition of positron-emission-tomography (PET) transmission information after tracer injection using ring and rotating pin transmission sources is discussed. A combined transmission/emission scan was acquired, followed by an emission scan used to substract the emission counts from the transmission/emission data. The ratio of emission count rate for brain scans to transmission count rate is 50-100% for a 5-mCi ring source and less than 5% for a 5-mCi pin source. Windowing of the sinogram, which rejects most random and scattered coincidences, also eliminates most emission counts. The magnitude and effects of residual random and scattered coincidences as well as increases in variability from transmission/emission scans were studied. The results of combined transmission/emission scans for a high-contrast emission source distribution using ring and pin sources are also described. >