Search

In this article, two latest silicon photomultiplier (SiPM) arrays from Hamamatsu and Broadcom were evaluated for high-resolution dual-ended readout positron emission tomography (PET) detectors. The Hamamatsu SiPM array and the Broadcom SiPM array are both $8\times8$ arrays of $\sim 4\times4$ mm2 SiPMs and were fabricated using 50 and 30- $\mu \text{m}$ microcells, respectively. To evaluate the SiPM arrays for PET applications, two dual-ended readout PET detectors were fabricated, while each detector contains two identical SiPM arrays coupled to both ends of the same $20\times20$ lutetium yttrium oxyorthosilicate array with a pitch size of 1.5 mm and a length of 20 mm. The results show that the two detectors provide excellent performance for high-resolution PET applications. At a temperature of $23~^\circ \text{C}$ , the flood histogram quality, energy resolution, depth-of-interaction resolution, and timing resolution obtained using the Hamamatsu SiPM arrays at a bias voltage of 41.5 V are 4.83 ± 0.64, 10.5 ± 3.5%, 1.88 ± 0.17 mm, and 0.86 ± 0.13 ns, respectively, and those obtained using the Broadcom SiPM arrays at a bias voltage of 32.5 V are 4.01 ± 0.53, 15.7 ± 3.6%, 2.01 ± 0.19 mm, and 0.88 ± 0.11 ns, respectively.

Background Reducing the radiation tracer dose and scanning time during positron emission tomography (PET) imaging can reduce the cost of the tracer, reduce motion artifacts, and increase the efficiency of the scanner. However, the reconstructed images to be noisy. It is very important to reconstruct high-quality images with low-count (LC) data. Therefore, we propose a deep learning method called LCPR-Net, which is used for directly reconstructing full-count (FC) PET images from corresponding LC sinogram data. Methods Based on the framework of a generative adversarial network (GAN), we enforce a cyclic consistency constraint on the least-squares loss to establish a nonlinear end-to-end mapping process from LC sinograms to FC images. In this process, we merge a convolutional neural network (CNN) and a residual network for feature extraction and image reconstruction. In addition, the domain transform (DT) operation sends a priori information to the cycle-consistent GAN (CycleGAN) network, avoiding the need for a large amount of computational resources to learn this transformation. Results The main advantages of this method are as follows. First, the network can use LC sinogram data as input to directly reconstruct an FC PET image. The reconstruction speed is faster than that provided by model-based iterative reconstruction. Second, reconstruction based on the CycleGAN framework improves the quality of the reconstructed image. Conclusions Compared with other state-of-the-art methods, the quantitative and qualitative evaluation results show that the proposed method is accurate and effective for FC PET image reconstruction.

Positron emission tomography (PET) is an advanced medical imaging technique widely used in various clinical applications, such as tumor detection and neurologic disorders. Reducing the radiotracer dose is desirable in PET imaging because it decreases the patient’s radiation exposure. However, reducing the dose can also increase noise, affecting the image quality. Therefore, an advanced image reconstruction algorithm based on low-dose PET data is needed to improve the quality of the reconstructed image. In this article, we propose the use of a direct PET image reconstruction network (DPIR-Net) using an improved Wasserstein generative adversarial network (WGAN) framework to enhance image quality. This article provides two main findings. First, our network uses sinogram data as input and outputs high-quality PET images direct, resulting in shorter reconstruction times compared with traditional model-based reconstruction networks. Second, we combine perceptual loss, mean square error, and the Wasserstein distance as the loss function, which effectively solves the problem of excessive smoothness and loss of detailed information in traditional network image reconstruction. We performed a comparative study using maximum-likelihood expectation maximization (MLEM) with a post-Gaussian filter, a total variation (TV)-norm regularization, a nonlocal means (NLMs) denoising method, a neural network denoising method, a traditional deep learning PET reconstruction network, and our proposed DPIR-Net method and evaluated the proposed method using both human and mouse data. The mouse data were obtained from a small animal PET prototype system developed by our laboratory. The quantitative and qualitative results show that our proposed method outperformed the conventional methods.

BACKGROUND: Radiation risk from computed tomography (CT) is always an issue for patients, especially those in clinical conditions in which repeated CT scanning is required. For patients undergoing repeated CT scanning, a low-dose protocol, such as sparse scanning, is often used, and consequently, an advanced reconstruction algorithm is also needed. OBJECTIVE: To develop a novel algorithm used for sparse-view CT reconstruction associated with the prior image. METHODS: A low-dose CT reconstruction method based on prior information of normal-dose image (PI-NDI) involving a transformed model for attenuation coefficients of the object to be reconstructed and prior information application in the forward-projection process was used to reconstruct CT images from sparse-view projection data. A digital extended cardiac-torso (XCAT) ventral phantom and a diagnostic head phantom were employed to evaluate the performance of the proposed PI-NDI method. The root-mean-square error (RMSE), peak signal-to-noise ratio (PSNR) and mean percent absolute error (MPAE) of the reconstructed images were measured for quantitative evaluation of the proposed PI-NDI method. RESULTS: The reconstructed images with sparse-view projection data via the proposed PI-NDI method have higher quality by visual inspection than that via the compared methods. In terms of quantitative evaluations, the RMSE measured on the images reconstructed by the PI-NDI method with sparse projection data is comparable to that by MLEM-TV, PWLS-TV and PWLS-PICCS with fully sampled projection data. When the projection data are very sparse, images reconstructed by the PI-NDI method have higher PSNR values and lower MPAE values than those from the compared algorithms. CONCLUSIONS: This study presents a new low-dose CT reconstruction method based on prior information of normal-dose image (PI-NDI) for sparse-view CT image reconstruction. The experimental results validate that the new method has superior performance over other state-of-art methods.

Breast cancer is the most frequently diagnosed cancer in women worldwide. Digital breast tomosynthesis (DBT), which is based on limited-angle tomography, was developed to solve tissue overlapping problems associated with traditional breast mammography. However, due to the problems associated with tube movement during the process of data acquisition, stationary DBT (s-DBT) was developed to allow the X-ray source array to stay stationary during the DBT scanning process. In this work, we evaluate four widely used and investigated DBT image reconstruction algorithms, including the commercial Feldkamp-Davis-Kress algorithm (FBP), the simultaneous iterative reconstruction technique (SIRT), the simultaneous algebraic reconstruction technique (SART) and the total variation regularized SART (SART-TV) for an s-DBT imaging system that we set up in our own laboratory for studies using a semi-elliptical digital phantom and a rubber breast phantom to determine the most superior algorithm for s-DBT image reconstruction among the four algorithms. Several quantitative indexes for image quality assessment, including the peak signal-noise ratio (PSNR), the root mean square error (RMSE) and the structural similarity (SSIM), are used to determine the best algorithm for the imaging system that we set up. Image resolutions are measured via the calculation of the contrast-to-noise ratio (CNR) and artefact spread function (ASF). The experimental results show that the SART-TV algorithm gives reconstructed images with the highest PSNR and SSIM values and the lowest RMSE values in terms of image accuracy and similarity, along with the highest CNR values calculated for the selected features and the best ASF curves in terms of image resolution in the horizontal and vertical directions. Thus, the SART-TV algorithm is proven to be the best algorithm for use in s-DBT image reconstruction for the specific imaging task in our study.

Purpose To reduce the cost of positron emission tomography (PET) scanning systems, image reconstruction algorithms for low-sampled data have been extensively studied. However, the current method based on total variation (TV) minimization regularization nested in the maximum likelihood-expectation maximization (MLEM) algorithm cannot distinguish true structures from noise resulting losing some fine features in the images. Thus, this work aims to recover fine features lost in the MLEM-TV algorithm from low-sampled data. Method A feature refinement (FR) approach previously developed for statistical interior computed tomography (CT) reconstruction is applied to PET imaging to recover fine features in this study. The proposed method starts with a constant initial image and the FR step is performed after each MLEM-TV iteration to extract the desired structural information lost during TV minimization. A feature descriptor is specifically designed to distinguish structure from noise and artifacts. A modified steepest descent method is adopted to minimize the objective function. After evaluating the impacts of different patch sizes on the outcome of the presented method, an optimal patch size of 7×7 is selected in this study to balance structure-detection ability and computational efficiency. Results Applying MLEM-TV-FR algorithm to the simulated brain PET imaging using an emission activity phantom, a standard Shepp-Logan phantom, and mouse results in the increased peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) as comparing to using the conventional MLEM-TV algorithm, as well as the substantial reduction of the used sampling numbers, which improves the computational efficiency. Conclusions The presented algorithm can achieve image quality superior to that of the MLEM and MLEM-TV approaches in terms of the preservation of fine structure and the suppression of undesired artifacts and noise, indicating its useful potential for low-sampled data in PET imaging.

X-ray radiation is harmful to human health. Thus, obtaining a better reconstructed image with few projection view constraints is a major challenge in the computed tomography (CT) field to reduce radiation dose. In this study, we proposed and tested a new algorithm that combines penalized weighted least-squares using total generalized variation (PWLS-TGV) and dictionary learning (DL), named PWLS-TGV-DL to address this challenge. We first presented and tested this new algorithm and evaluated it through both data simulation and physical experiments. We then analyzed experimental data in terms of image qualitative and quantitative measures, such as the structural similarity index (SSIM) and the root mean square error (RMSE). The experiments and data analysis indicated that applying the new algorithm to CT data recovered images more efficiently and yielded better results than the traditional CT image reconstruction approaches.

Recently, low-dose computed tomography (CT) has become highly desirable due to the increasing attention paid to the potential risks of excessive radiation of the regular dose CT. However, ensuring image quality while reducing the radiation dose in the low-dose CT imaging is a major challenge. Compared to classical filtered back-projection (FBP) algorithms, statistical iterative reconstruction (SIR) methods for modeling measurement statistics and imaging geometry can significantly reduce the radiation dose, while maintaining the image quality in a variety of CT applications. To facilitate low-dose CT imaging, we in this study proposed an improved statistical iterative reconstruction scheme based on the penalized weighted least squares (PWLS) standard combined with total variation (TV) minimization and sparse dictionary learning (DL), which is named as a method of PWLS-TV-DL. To evaluate this PWLS-TV-DL method, we performed experiments on digital phantoms and physical phantoms, and analyzed the results in terms of image quality and calculation. The results show that the proposed method is better than the comparison methods, which indicates the potential of applying this PWLS-TV-DL method to reconstruct low-dose CT images.

To reduce overall patient radiation exposure in some clinical scenarios (since cancer patients need frequent follow-ups), noncontrast CT is not used in some institutions. However, although less desirable, noncontrast CT could provide additional important information. In this article, we propose a deep subtraction residual network based on adjacency content transfer to reconstruct noncontrast CT from contrast CT and maintain image quality comparable to that of a CT scan originally acquired without contrast. To address the slight structural dissimilarity of the paired CT images (noncontrast CT and contrast CT) due to involuntary physiological motion, we introduce a contrastive loss network derived from the adjacency content-transfer strategy. We evaluate the results of various similarity metrics (MSE, SSIM, NRMSE, PSNR, MAE) and the fitting curve (HU distribution) of the output mapping to estimate the reconstruction performance of the algorithm. To build the model, we randomly select a total of 15,405 CT paired images (noncontrast CT and contrast-enhanced CT) for training and 10,270 CT paired images for testing. The proposed algorithm preserves the robust structures from the contrast-enhanced CT scans and learns the noncontrast attenuation pattern from the noncontrast CT scans. During the evaluation, the deep subtraction residual network achieves higher MSE, MAE, NRMSE, and PSNR scores (by 30%) than those of the baseline models (BEGAN, CycleGAN, Pixel2Pixel) and better simulates the HU curve of noncontrast CT attenuation. After validation based on an analysis of the experimental results, we can report that the noncontrast CT images reconstructed by our proposed algorithm not only preserve the high-quality structures from the contrast-enhanced CT images, but also mimic the CT attenuation of the originally acquired noncontrast CT images.

Both monolithic and semi-monolithic scintillator positron emission tomography (PET) detectors can measure the depth of interaction with single-ended readout. Usually scintillators with a thickness of 10 mm or less are used since the position resolutions of the detectors degrade as the scintillator thickness increases. In this work, the performance of a 20 mm thick long rectangular semi-monolithic scintillator PET detector was measured by using both single-ended and dual-ended readouts with silicon photomultiplier (SiPM) arrays to provide a high detection efficiency. The semi-monolithic scintillator detector consists of nine lutetium–yttrium oxyorthosilicate slices measuring 1.37 × 51.2 × 20 mm3 with erythrocyte sedimentation rate foils of 0.065 mm thickness in between the slices. The SiPM array at each end of the scintillator detector consists of 16 × 4 SiPMs with a pixel size of 3.0 × 3.0 mm2 and a pitch of 3.2 mm. The 64 signals of each SiPM array are processed by using the TOFPET2 application-specific integrated circuit individually. All but the edge slices can be clearly resolved for the detectors with both single-ended and dual-ended readouts. The single-ended readout detector provides an average full width at half maximum (FWHM) Y (continuous direction) position resolution of 2.43 mm, Z (depth direction) position resolution of 4.77 mm, energy resolution of 25.7% and timing resolution of 779 ps. The dual-ended readout detector significantly improves the Y and Z position resolutions, slightly improves the energy and timing resolution at the cost of two photodetectors required for one detector module and provides an average FWHM Y position resolution of 1.97 mm, Z position resolution of 2.60 mm, energy resolution of 21.7% and timing resolution of 718 ps. The energy and timing resolution of the semi-monolithic scintillator detector in this work are worse than those of the segmented scintillator array detector and need to be further improved. The semi-monolithic scintillator detector described in this work reduces costs as compared to the traditional segmented scintillator array detector and reduces the edge effect as compared to the monolithic scintillator detector.

BACKGROUND: Statistical reconstruction methods based on penalized maximum likelihood (PML) are being increasingly used in positron emission tomography (PET) imaging to reduce noise and improve image quality. Wang and Qi proposed a patch-based edge-preserving penalties algorithm that can be implemented in three simple steps: a maximum-likelihood expectation-maximization (MLEM) image update, an image smoothing step, and a pixel-by-pixel image fusion step. The pixel-by-pixel image fusion step, which fuses the MLEM updated image and the smoothed image, involves a trade-off between preserving the fine structural features of an image and suppressing noise. Particularly when reconstructing images from low-count data, this step cannot preserve fine structural features in detail. To better preserve these features and accelerate the algorithm convergence, we proposed to improve the patch-based regularization reconstruction method. METHODS: Our improved method involved adding a total variation (TV) regularization step following the MLEM image update in the patch-based algorithm. A feature refinement (FR) step was then used to extract the lost fine structural features from the residual image between the TV regularized image and the fused image based on patch regularization. These structural features would then be added back to the fused image. With the addition of these steps, each iteration of the image should gain more structural information. A brain phantom simulation experiment and a mouse study were conducted to evaluate our proposed improved method. Brain phantom simulation with added noise were used to determine the feasibility of the proposed algorithm and its acceleration of convergence. Data obtained from the mouse study were divided into event count sets to validate the performance of the proposed algorithm when reconstructing images from low-count data. Five criteria were used for quantitative evaluation: signal-to-noise ratio (SNR), covariance (COV), contrast recovery coefficient (CRC), regional relative bias, and relative variance. RESULTS: The bias and variance of the phantom brain image reconstructed using the patch-based method were 0.421 and 5.035, respectively, and this process took 83.637 seconds. The bias and variance of the image reconstructed by the proposed improved method, however, were 0.396 and 4.568, respectively, and this process took 41.851 seconds. This demonstrates that the proposed algorithm accelerated the reconstruction convergence. The CRC of the phantom brain image reconstructed using the patch-based method was iterated 20 times and reached 0.284, compared with the proposed method, which reached 0.446. When using a count of 5,000 K data obtained from the mouse study, both the patch-based method and the proposed method reconstructed images similar to the ground truth image. The intensity of the ground truth image was 88.3, and it was located in the 102(nd) row and the 116(th) column. However, when the count was reduced to below 40 K, and the patch-based method was used, image quality was significantly reduced. This effect was not observed when the proposed method was used. When a count of 40 K was used, the image intensity was 58.79 when iterated 100 times by the patch-based method, and it was located in the 102(nd) row and the 116(th) column, while the intensity when iterated 50 times by the proposed method was 63.83. This suggests that the proposed method improves image reconstruction from low-count data. CONCLUSIONS: This improved method of PET image reconstruction could potentially improve the quality of PET images faster than other methods and also produce better reconstructions from low-count data.

We investigate the reconstruction of low-count positron emission tomography (PET) projection, which is an important, but challenging, task. Using the texture feature extraction method of radiomics, i.e. the gray-level co-occurrence matrix (GLCM), texture features can be extracted from magnetic resonance imaging images with high-spatial resolution. In this work, we propose a kernel reconstruction method combining autocorrelation texture features derived from the GLCM. The new kernel function includes the correlations of both the intensity and texture features from the prior image. By regarding the GLCM as a discrete approximation of a probability density function, the asymptotically gray-level-invariant autocorrelation texture feature is generated, which can maintain the accuracy of texture features extracted from small image regions by reducing the number of quantized image gray levels. A computer simulation shows that the proposed method can effectively reduce the noise in the reconstructed image compared to the maximum likelihood expectation maximum method and improve the image quality and tumor region accuracy compared to the original kernel method for low-count PET reconstruction. A simulation study on clinical patient images also shows that the proposed method can improve the whole image quality and that the reconstruction of a high-uptake lesion is more accurate than that achieved by the original kernel method.

In this work, a small animal PET scanner named SIAT aPET was developed using dual-ended readout depth encoding detectors to simultaneously achieve high spatial resolution and high sensitivity. The scanner consists of four detector rings with 12 detector modules per ring; the ring diameter is 111 mm and the axial field of view (FOV) is 105.6 mm. The images are reconstructed using an ordered subset expectation maximization (OSEM) algorithm. The spatial resolution of the scanner was measured by using a 22Na point source at the center axial FOV with different radial offsets. The sensitivity of the scanner was measured at center axis of the scanner with different axial positions. The count rate performance of the system was evaluated by scanning mouse-sized and rat-sized phantoms. An ultra-micro hot-rods phantom and two mice injected with 18F-NaF and 18F-FDG were scanned on the scanner. An average depth of interaction (DOI) resolution of 1.96 mm, energy resolution of 19.1% and timing resolution of 1.20 ns were obtained for the detector. Average spatial resolutions of 0.82 mm and 1.16 mm were obtained up to a distance of 30 mm radially from the center of the FOV when reconstructing a point source in 1% and 10% warm backgrounds, respectively, using OSEM reconstruction with 16 subsets and 10 iterations. Sensitivities of 16.0% and 11.9% were achieved at center of the scanner for energy windows of 250–750 keV and 350–750 keV respectively. Peak noise equivalent count rates (NECRs) of 324 kcps and 144 kcps were obtained at an activity of 26.4 MBq for the mouse-sized and rat-sized phantoms. Rods of 1.0 mm diameter can be visually resolved from the image of the ultra-micro hot-rods phantom. The capability of the scanner was demonstrated by high quality in-vivo mouse images.

In this work, a GPU-accelerated fully 3D ordered-subset expectation maximization (OSEM) image reconstruction with point spread function (PSF) modeling was developed for a small animal PET scanner with a long axial field of view (FOV). Dual-ended readout detectors that provided high depth of interaction (DOI) resolution were used for the small animal PET scanner to simultaneously achieve uniform high spatial resolution and high sensitivity. First, we developed a novel sinogram generation method, in which the dimension of the sinogram was determined first and then an event was assigned to a few neighboring sinogram elements by using weights that are inversely proportional to the distance from the measured line of response (LOR) to the LOR of the sinogram elements. System geometric symmetry, precomputation of LOR-driven ray-tracing and texture memory were applied to accelerate the GPU-based reconstruction. We developed a spatially variant PSF model where the PSF parameters were obtained by using point source images measured at 18 positions in the FOV and a spatial invariant PSF model where the PSF parameters were obtained by using only one image measured at the center FOV. The performance of the image reconstruction method was evaluated by using simulated phantom data as well as phantom and in-vivo mouse data acquired on the scanner. The results showed that the proposed reconstruction method provided better spatial resolution, a higher contrast recovery coefficient and lower noise than the OSEM reconstruction and was more than 1000 times faster than the CPU-based reconstruction. The spatially variant PSF model did not result in any spatial resolution improvement compared to the spatial invariant PSF model, and thus, the latter that is much easier to implement in image reconstruction and can be used in a small animal PET scanner using detectors with very high DOI resolution. A whole body 18F-FDG mouse image with high resolution and a high contrast to noise ratio was obtained by using the proposed reconstruction method.

Purpose Segmentation of magnetic resonance images (MRI) of the left ventricle (LV) plays a key role in quantifying the volumetric functions of the heart, such as the area, volume, and ejection fraction. Traditionally, LV segmentation is performed manually by experienced experts, which is both time-consuming and prone to subjective bias. This study aims to develop a novel capsule-based automated segmentation method to automatically segment the LV from images obtained by cardiac MRI. Method The technique applied for segmentation uses Fourier analysis and the circular Hough transform (CHT) to indicate the approximate location of the LV and a network capsule to precisely segment the LV. The neurons of the capsule network output a vector and preserve much of the information about the input by replacing the largest pooling layer with convolutional strides and dynamic routing. Finally, the segmentation result is postprocessed by threshold segmentation and morphological processing to increase the accuracy of LV segmentation. Results We fully exploit the capsule network to achieve the segmentation goal and combine LV detection and capsule concepts to complete LV segmentation. In the experiments, the tested methods achieved LV Dice scores of 0.922±0.05 end-diastolic (ED) and 0.898±0.11 end-systolic (ES) on the ACDC 2017 data set. The experimental results confirm that the algorithm can effectively perform LV segmentation from a cardiac magnetic resonance image. To verify the performance of the proposed method, visual and quantitative comparisons are also performed, which show that the proposed method exhibits improved segmentation accuracy compared with the traditional method. Conclusions The evaluation metrics of medical image segmentation indicate that the proposed method in combination with postprocessing and feature detection effectively improves segmentation accuracy for cardiac MRI. To the best of our knowledge, this study is the first to use a deep learning model based on capsule networks to systematically evaluate end-to-end LV segmentation.

PET scanners using SiPMs as photodetectors could have tens of thousands of SiPMs. To simplify the readout electronics, analog signal multiplexing readouts are always preferred to be used as early as possible. In this paper, two simple analog signal multiplexing readouts, a capacitive charge-division readout, and a resistive charge-division readout were evaluated and compared using dual-ended readout detectors based on 10 × 10 arrays of SensL MicroFJ-30035 SiPMs coupled to both ends of a 20 × 20 LYSO array with a pitch size of 1.5 mm and a length of 20 mm. The performance of the detectors were evaluated at different bias voltages (from 27.0 V to 30.5 V with an interval of 0.5 V) and a temperature of 22.8 °C. The flood histograms show that all the crystals in the LYSO array were clearly identified, whilst better flood histogram was obtained using the resistive charge-division readout. At a bias voltage of 29.5V, the flood histogram quality, energy resolution, DOI resolution, and timing resolution of the detector obtained using the capacitive charge-division readout were 3.28 ± 0.85, 18.9% ± 6.2%, 1.93 ± 0.20 mm, 1.25 ± 0.11 ns respectively, and those obtained using the resistive charge-division readout were 3.57 ± 0.81, 16.9% ± 6.5%, 1.96 ± 0.23 mm and 1.23 ± 0.07 ns, respectively. Overall, the detector with the resistive charge-division readout provided better performance.

Positron emission tomography (PET) imaging can be used for early detection, diagnosis and postoperative patient monitoring of many diseases. Traditional PET imaging requires not only additional computed tomography (CT) imaging or magnetic resonance imaging (MR) to provide anatomical information but also attenuation correction (AC) map calculation based on CT images or MR images for accurate quantitative estimation. During a patient's treatment, PET/CT or PET/MR scans are inevitably repeated many times, leading to additional doses of ionizing radiation (CT scans) and additional economic and time costs (MR scans). To reduce adverse effects while obtaining high-quality PET/MR images in the course of a patient's treatment, especially in the stage of evaluating the effect of postoperative treatment, in this work, we propose a new method based on deep learning, which can directly obtain synthetic attenuation-corrected PET (sAC PET) and synthetic T1-weighted MR (sMR) images based only on non-attenuation-corrected PET (NAC PET) images. Our model, based on the Wasserstein generative adversarial network, first removes noise and artifacts from the NAC PET images to generate sAC PET images and then generates sMR images from the obtained sAC PET images. To evaluate the performance of this generative model, we evaluated it on paired PET/MR images from a total of eighty clinical patients. Based on qualitative and quantitative analysis, the generated sAC PET and sMR images showed a high degree of similarity to the real AC PET and real MR images. These results indicated that our proposed method can reduce the frequency of additional anatomical imaging scans during PET imaging and has great potential in improving doctors' clinical diagnosis efficiency, saving patients' economic expenditure and reducing the radiation risk brought by CT scanning.

A depth encoding PET detector module using semi-monolithic scintillation crystal single-ended readout by a SiPM array was built and its performance was measured. The semi-monolithic scintillator detector consists of 11 polished LYSO slices measuring 1 × 11.6 × 10 mm3. The slices are glued together with enhanced specular reflector (ESR) in between and outside of the slices. The bottom surface of the slices is coupled to a 4 × 4 SiPM array with a 1 mm light guide and silicon grease between them. No reflector is used on the top surface and two sides of the slices to reduce the scintillation photon reflection. The signals of the 4 × 4 SiPM array are grouped along rows and columns separately into eight signals. Four SiPM column signals are used to identify the slices according to the center of the gravity of the scintillation photon distribution in the pixelated direction. Four SiPM row signals are used to estimate the y (monolithic direction) and z (depth of interaction) positions according to the center of the gravity and the width of the scintillation photon distribution in the monolithic direction, respectively. The detector was measured with 1 mm sampling interval in both the y and z directions with electronic collimation by using a 0.25 mm diameter 22Na point source and a 1 × 1 × 20 mm3 LYSO crystal detector. An average slice based energy resolution of 14.9% was obtained. All slices of 1 mm thick were clearly resolved and a detector with even thinner slices could be used. The y positions calculated with the center of gravity method are different for interactions happening at the same y, but different z positions due to depth dependent edge effects. The least-square minimization and the maximum likelihood positioning algorithms were developed and both methods improved the spatial resolution at the edges of the detector as compared with the center of gravity method. A mean absolute error (MAE) which is defined as the probability-weighted mean of the absolute value of the positioning error is used to evaluate the spatial resolution. An average MAE spatial resolution of ~1.15 mm was obtained in both y and z directions without rejection of the multiple scattering events. The average MAE spatial resolution was ~0.7 mm in both y and z directions after the multiple scattering events were rejected. The timing resolution of the detector is 575 ps. In the next step, long rectangle detector will be built to reduce edge effects and improve the spatial resolution of the semi-monolithic detector. Thick detector up to 20 mm will be explored and the positioning algorithms will be further optimized.

Small animal positron emission tomography (PET) is a well-established imaging modality in preclinical biomedical research. The performance of current small animal PET scanners is mainly limited by the detector performance and depth-encoding detectors are required to simultaneously achieve high spatial resolution and high sensitivity. In this work, the performance of a high-resolution dual-ended readout lutetium-yttrium oxyorthosilicate (LYSO) array using barium sulfate powder (BaSO4) as the inter-crystal reflector was measured for the first time and compared to that of a LYSO array using the most commonly used enhanced specular reflector (ESR). Both LYSO arrays have 18 × 18 crystals and the crystal size is about 0.62 × 0.62 × 20 mm3. The LYSO arrays are readout by two position-sensitive photomultiplier tubes (PSPMTs) from both ends. The flood histograms, energy resolution, depth of interaction (DOI) resolution and timing resolution were measured. The flood histograms of the LYSO array with BaSO4 reflector is much better than that of the LYSO array with ESR reflector. For the BaSO4 array, all crystals can be clearly resolved. For the ESR array, all crystals in one direction can be clearly resolved, but the edge 2-3 columns of the crystals in the other direction cannot be resolved. The average energy resolution of the BaSO4 and ESR arrays are 15.2% and 15.3%, respectively. The average DOI resolution of the BaSO4 array is 2.19 mm, which is 24% worse than the 1.76 mm DOI resolution of the ESR array. The timing resolution of both arrays is ~1.6 ns. The LYSO array with the new BaSO4 reflector provided an much better flood histogram in a high resolution dual-ended readout PET detectors as compared to the ESR array, and will be used to develop a small animal PET scanner that can simultaneously achieve uniform high spatial resolution, high sensitivity and low cost.

Detectors with depth encoding capability, good energy resolution and good timing resolution are required to develop high performance whole body and total body PET scanners. In this work, an 8 × 8 LYSO array with crystal size of 3 × 3 × 20 mm3 was fabricated by using coupling materials that included 5 mm optical glue, 9 mm triangle ESR reflector and 6 mm rectangle ESR reflector from the top to the bottom in between adjacent crystals in one direction. The LYSO array was single-ended read out by an 8 × 8 SiPM array with one-to-one coupling. Row and column summing circuit was used to read out the SiPM signals. The depth of interaction (DOI) of the detector was measured with the depth dependent light sharing between adjacent crystals while maintaining acceptable energy resolution and timing resolution. The performance of the detector module was evaluated. All crystals can be clearly resolved from the measured flood histogram. An average DOI resolution of 4.62 mm and an average timing resolution of 518 ps are obtained for events with E > 400 keV. The average energy resolution of the detector is 13.5%. This work provides a cost-effective depth encoding detector module that can be used to build high performance whole body and total body PET scanners in the future.

Dynamic myocardial perfusion computed tomography (DMP-CT) is an effective medical imaging technique for coronary artery disease diagnosis and therapy guidance. However, the radiation dose received by the patient during repeated CT scans is a widespread concern of radiologists because of the increased risk of cancer. The sparse few-view CT scanning protocol can be a feasible approach to reduce the radiation dose of DMP-CT imaging; however, an advanced reconstruction algorithm is needed. In this paper, a temporal feature prior-aided separated reconstruction method (TFP-SR) for low-dose DMP-CT images reconstruction from sparse few-view sinograms is proposed. To implement the proposed method, the objective perfusion image is divided into the baseline fraction and the enhancement fraction introduced by the arrival of the contrast agent. The core of the proposed TFP-SR method is the utilization of the temporal evolution information that naturally exists in the DMP-CT image sequence to aid the enhancement image reconstruction from limited data. The temporal feature vector of an image pixel is defined by the intensities of this pixel in the pre-reconstructed enhancement sequence, and the connection between two related features is calculated via a zero-mean Gaussian function. A prior matrix is constructed based on the connections between the extracted temporal features and used in the iterative reconstruction of the enhancement images. To evaluate the proposed method, the conventional filtered back-projection algorithm, the total variation regularized PWLS (PWLS-TV) and the prior image constrained compressed sensing are compared in this paper based on studies on a digital extended cardiac-torso (XCAT) thoracic phantom and a preclinical porcine DMP-CT data set that take image misregistration into account. The experimental results demonstrate that the proposed TFP-SR method has superior performance in sparse DMP-CT images reconstruction in terms of image quality and the analyses of the time attenuation curve and hemodynamic parameters.

Many deep learning (DL)-based image restoration methods for low-dose CT (LDCT) problems directly employ the end-to-end networks on low-dose training data without considering dose differences. However, the radiation dose difference has a great impact on the ultimate results, and lower doses increase the difficulty of restoration. Moreover, there is increasing demand to design and estimate acceptable scanning doses for patients in clinical practice, necessitating dose-aware networks embedded with adaptive dose estimation. In this paper, we consider these dose differences of input LDCT images and propose an adaptive dose-aware network. First, considering a large dose distribution range for simulation convenience, we coarsely define five dose levels in advance as lowest, lower, mild, higher and highest radiation dose levels. Instead of directly building the end-to-end mapping function between LDCT images and high-dose CT counterparts, the dose level is primarily estimated in the first stage. In the second stage, the adaptively learned low-dose level is used to guide the image restoration process as the pattern of prior information through the channel feature transform. We conduct experiments on a simulated dataset based on original high dose parts of American Association of Physicists in Medicine challenge datasets from the Mayo Clinic. Ablation studies validate the effectiveness of the dose-level estimation, and the experimental results show that our method is superior to several other DL-based methods. Specifically, our method provides obviously better performance in terms of the peak signal-to-noise ratio and visual quality reflected in subjective scores. Due to the dual-stage process, our method may suffer limitations under more parameters and coarse dose-level definitions, and thus, further improvements in clinical practical applications with different CT equipment vendors are planned in future work.

Positron emission tomography (PET) imaging plays an indispensable role in early disease detection and postoperative patient staging diagnosis. However, PET imaging requires not only additional computed tomography (CT) imaging to provide detailed anatomical information but also attenuation correction (AC) maps calculated from CT images for precise PET quantification, which inevitably demands that patients undergo additional doses of ionizing radiation. To reduce the radiation dose and simultaneously obtain high-quality PET/CT images, in this work, we present an alternative based on deep learning that can estimate synthetic attenuation corrected PET (sAC PET) and synthetic CT (sCT) images from non-attenuation corrected PET (NAC PET) scans for whole-body PET/CT imaging. Our model consists of two stages: the first stage removes noise and artefacts in the NAC PET images to generate sAC PET images, and the second stage synthesizes CT images from the sAC PET images obtained in the first stage. Both stages employ the same deep Wasserstein generative adversarial network and identical loss functions, which encourage the proposed model to generate more realistic and satisfying output images. To evaluate the performance of the proposed algorithm, we conducted a comprehensive study on a total of 45 sets of paired PET/CT images of clinical patients. The final experimental results demonstrated that both the generated sAC PET and sCT images showed great similarity to true AC PET and true CT images based on both qualitative and quantitative analyses. These results also indicate that in the future, our proposed algorithm has tremendous potential for reducing the need for additional anatomic imaging in hybrid PET/CT systems or the need for lengthy MR sequence acquisition in hybrid PET/MRI systems.

A prototype small animal positron emission tomography (PET) scanner for mouse brain imaging has been developed at UC Davis. The new scanner uses tapered detector arrays with depth of interaction (DOI) measurement. In this paper, we present an efficient system model for the tapered PET scanner using matrix factorization and a virtual scanner geometry. The factored system matrix mainly consists of two components: a sinogram blurring matrix and a geometrical matrix. The geometric matrix is based on a virtual scanner geometry. The sinogram blurring matrix is estimated by matrix factorization. We investigate the performance of different virtual scanner geometries. Both simulation study and real data experiments are performed in the fully 3D mode to study the image quality under different system models. The results indicate that the proposed matrix factorization can maintain image quality while substantially reduce the image reconstruction time and system matrix storage cost. The proposed method can be also applied to other PET scanners with DOI measurement.

Detectors with depth-encoding capability and good timing resolution are required to develop high-performance whole-body or total-body PET scanners. In this work, depth-encoding PET detectors that use light sharing between two discrete crystals and single-ended readout with silicon photomultipliers (SiPMs) were manufactured and evaluated. The detectors consisted of two unpolished 3 × 3 × 20 mm3 LYSO crystals with different coupling materials between them and were read out by Hamamatsu 3 × 3 mm2 SiPMs with one-to-one coupling. The ratio of the energy of one SiPM to the total energy of two SiPMs was used to measure the depth of interaction (DOI). Detectors with different coupling materials in-between the crystals were measured in the singles mode in an effort to obtain detectors that can provide good DOI resolution. The DOI resolution and energy resolution of three types of detector were measured and the timing resolution was measured for the detector with the best DOI and energy resolution. The optimum detector, with 5 mm optical glue, a 9 mm triangular ESR and a 6 mm rectangular ESR in-between the unpolished crystals, provides a DOI resolution of 2.65 mm, an energy resolution of 10.0% and a timing resolution of 427 ps for events of E > 400 keV. The detectors simultaneously provide good DOI and timing resolution, and show great promise for the development of high-performance whole-body and total-body PET scanners.

The probability of inter-crystal scattering (ICS) events for 511 keV gamma rays in all current scintillation crystals is high and the ICS events degrade the spatial resolution of PET scanners. In this work, Monte Carlo simulations were performed to study the effects of ICS events on the sensitivity and spatial resolution of PET detectors. LaBr3, LYSO, and PWO that represent scintillation crystals of low, medium and high density, respectively, were used. For a point source placed in the middle of two scintillation detectors of 50 × 50 × 20 mm3 and a lower energy threshold (LET) of 350 keV, the probabilities that at least one gamma ray undergoes ICS are 94%, 84% and 76% for LaBr3, LYSO, and PWO, respectively. The ICS events still provide useful spatial information. The full width at half maximum (FWHM), the full width at tenth maximum (FWTM) and the mean absolute error (MAE) of the curve of the mispositioning of a point source caused by ICS events are 0.45, 3.0 and 0.9 mm if the most popular PET scintillator LYSO is used. The MAE is smaller than the spatial resolution of most current PET scanners. The effect of ICS increases as the detector LET increases, scintillator density decreases, and crystal size decreases. The intrinsic spatial resolutions of a pair of LYSO detectors were calculated using curves of the coincidence counts between one column of the crystals in the two detectors and the sum of the coincidence counts between two opposite crystals of the columns in the two detectors that are in line with the point source changing with the source positions. The latter method removes almost all of ICS events. The FWHM (FWTM) intrinsic spatial resolutions obtained by the two methods are 0.40 (2.0) mm and 0.33 (0.8) mm if the crystal size is 0.5 mm, and are 0.8 (3.0) and 0.68 (1.5) mm if the crystal size is 1.0 mm. ICS events have much bigger contributions to the FWTM rather than the FWHM of the intrinsic spatial resolution of PET detectors. The spatial resolution of a PET scanner can still be improved by decreasing the crystal size to as small as 0.5 mm.

UnlabelledWe developed a prototype small-animal PET scanner based on depth-encoding detectors using dual-ended readout of small scintillator elements to produce high and uniform spatial resolution suitable for imaging the mouse brain.MethodsThe scanner consists of 16 tapered dual-ended-readout detectors arranged in a 61-mm-diameter ring. The axial field of view (FOV) is 7 mm, and the transaxial FOV is 30 mm. The scintillator arrays consist of 14 × 14 lutetium oxyorthosilicate elements, with a crystal size of 0.43 × 0.43 mm at the front end and 0.80 × 0.43 mm at the back end, and the crystal elements are 13 mm long. The arrays are read out by 8 × 8 mm and 13 × 8 mm position-sensitive avalanche photodiodes (PSAPDs) placed at opposite ends of the array. Standard nuclear-instrumentation-module electronics and a custom-designed multiplexer are used for signal processing.ResultsThe detector performance was measured, and all but the crystals at the very edge could be clearly resolved. The average intrinsic spatial resolution in the axial direction was 0.61 mm. A depth-of-interaction resolution of 1.7 mm was achieved. The sensitivity of the scanner at the center of the FOV was 1.02% for a lower energy threshold of 150 keV and 0.68% for a lower energy threshold of 250 keV. The spatial resolution within a FOV that can accommodate the entire mouse brain was approximately 0.6 mm using a 3-dimensional maximum-likelihood expectation maximization reconstruction. Images of a hot-rod microphantom showed that rods with a diameter of as low as 0.5 mm could be resolved. The first in vivo studies were performed using (18)F-fluoride and confirmed that a 0.6-mm resolution can be achieved in the mouse head in vivo. Brain imaging studies with (18)F-FDG were also performed.ConclusionWe developed a prototype PET scanner that can achieve a spatial resolution approaching the physical limits of a small-bore PET scanner set by positron range and detector interaction. We plan to add more detector rings to extend the axial FOV of the scanner and increase sensitivity.

The aim of this study is to design and test a new high-resolution hybrid depth of interaction (DOI) detector for a dedicated breast PET/CT scanner. Two detectors have been designed and built. The completed detectors are based on a 14 × 14 array of 1.5 × 1.5 × 20 mm(3) unpolished lutetium orthosilicate scintillation crystals, with each element coated in a 50 μm layer of reflective material. The detector is read out from both ends using a position-sensitive photomultiplier tube (PSPMT) and a large active area (20 × 20 mm(2)) avalanche photodiode (APD) to enable acquisition of DOI information. Nuclear instrumentation modules were used to characterize the detectors' performances in terms of timing, intrinsic spatial resolution (ISR) and energy resolution, as well as DOI resolution with a dual-ended readout configuration. Measurements with the APD were performed at a temperature of 10 °C. All crystals were identified at all depths, even though the signal amplitude from the PSPMT decreases with depth away from it. We measured a timing resolution of 2.4 ns, and an average energy resolution of 19%. The mean ISR was measured to be 1.2 mm for crystals in the central row of the array for detectors in the face-to-face position. Two off-center positions were measured corresponding to 26° and 51° oblique photon incidence, and the mean ISR at these positions was 1.5 and 1.7 mm, respectively. The average DOI resolution across all crystals and depths was measured to be 2.9 mm (including the beam width of 0.6 mm). This detector design shows good promise as a high-resolution detector for a dedicated breast PET/CT scanner.

Improvements to current small animal PET scanners can be made by improving the sensitivity and the spatial resolution of the scanner. In the past, efforts have been made to minimize the crystal dimensions in the axial and transaxial directions to improve the spatial resolution and to increase the crystal length to improve the sensitivity of the scanner. We have designed tapered PET detectors with the purpose of reducing the gaps between detector modules and optimizing the sensitivity of a future-generation small animal PET scanner. In this work, we investigate spatial resolution and sensitivity of a scanner based on tapered detector elements using Monte Carlo simulations. For tapered detector elements more scintillation material is used per detector resulting in a higher sensitivity of the scanner. However, since the detector elements are not uniform in size, degradation in spatial resolution is also expected. To investigate characteristics of tapered PET detectors, the spatial resolution and sensitivity of a one-ring scanner were simulated for a system based on traditional cuboid detectors and a scanner based on tapered detectors. Additionally, the effect of depth of interaction (DOI) resolution on the spatial resolution for the traditional and tapered detectors was evaluated. All simulations were performed using the Monte Carlo simulation package GATE. Using the tapered arrays, a 64% improvement in the sensitivity across the field of view was found compared with traditional detectors for the same ring diameter. The level of DOI encoding was found to be the dominating factor in determining the radial spatial resolution and not the detector shape. For all levels of DOI encoding, no significant difference was found for the spatial resolution when comparing the tapered and the cuboid detectors. Detectors employing the tapered crystal design along with excellent DOI resolution will lead to PET scanners with higher sensitivity and uniform spatial resolution across the field of view.

In this paper, we study position-dependent timing shifts and timing resolution in position sensitive avalanche photodiodes (PSAPDs) and their effects on the coincidence window used in positron emission tomography (PET) systems using these devices. There is a delay in PSAPD signals that increases as the excitation position moves from the corner to the center of the device and the timing resolution concurrently worsens. The difference in timing between the center and the corner can be up to 30.7 ns for a 14 × 14 mm^2 area PSAPD. This means that a PSAPD-based PET system could require a very wide coincidence timing window (>60 ns) if this effect is not corrected, although the individual crystal pairs still have full-width half-maximum (FWHM) timing resolutions better than 7.4 ns. In addition to characterizing the timing properties of PSAPDs, two correction methods were developed and applied to data from a pair of PSAPD detectors. These two timing offset corrections reduced the timing shift of a crystal pair from 52.4 ns to 9.7 ns or 1.3 ns, improved the FWHM timing resolution of the detector pair from 24.6 ns to 9.5 ns or 6.0 ns and reduced the timing window (sufficient to cover at least twice the FWHM for all crystal pairs) from 65.1 ns to 22.0 ns or 15.2 ns, respectively. A two-step timing alignment method is proposed for a PET system consisting of multiple PSAPDs. Lastly, the effect of PSAPD size on the timing performance was also evaluated.

We investigate a novel detector using a lutetium oxyorthosilicate (LSO) scintillator and YGG (yttrium-aluminum-gallium oxide:cerium, Y(3)(Al,Ga)(5)O(12):Ce) phosphor to construct a detector with continuous depth-of-interaction (DOI) information. The far end of the LSO scintillator is coated with a thin layer of YGG phosphor powder which absorbs some fraction of the LSO scintillation light and emits wavelength-shifted photons with a characteristic decay time of approximately 50 ns. The near end of the LSO scintillator is directly coupled to a photodetector. The photodetector detects a mixture of the LSO light and the light emitted by YGG. With appropriate placement of the coating, the ratio of the light converted from the YGG coating with respect to the unconverted LSO light can be made to depend on the interaction depth. DOI information can then be estimated by inspecting the overall light pulse decay time. Experiments were conducted to optimize the coating method. 19 ns decay time differences across the length of the detector were achieved experimentally when reading out a 1.5 x 1.5 x 20 mm(3) LSO crystal with unpolished surfaces and half-coated with YGG phosphor. The same coating scheme was applied to a 4 x 4 LSO array. Pulse shape discrimination (PSD) methods were studied to extract DOI information from the pulse shape changes. The DOI full-width-half-maximum (FWHM) resolution was found to be approximately 8 mm for this 2 cm thick array.

Many laboratories develop depth-encoding detectors to improve the trade-off between spatial resolution and sensitivity in positron emission tomography (PET) scanners. One challenge in implementing these detectors is the need to calibrate the depth of interaction (DOI) response for the large numbers of detector elements in a scanner. In this work, we evaluate two different methods, a linear detector calibration and a linear crystal calibration, for determining DOI calibration parameters. Both methods can use measurements from any source distribution and location, or even the intrinsic lutetium oxyorthosilicate (LSO) background activity, and are therefore well suited for use in a depth-encoding PET scanner. The methods were evaluated by measuring detector and crystal DOI responses for all eight detectors in a prototype depth-encoding PET scanner. The detectors utilize dual-ended readout of LSO scintillator arrays with position-sensitive avalanche photodiodes (PSAPDs). The LSO arrays have 7 x 7 elements, with a crystal size of 0.92 x 0.92 x 20 mm(3) and pitch of 1.0 mm. The arrays are read out by two 8 x 8 mm(2) area PSAPDs placed at opposite ends of the arrays. DOI is measured by the ratio of the amplitude of the total energy signals measured by the two PSAPDs. Small variations were observed in the DOI responses of different crystals within an array as well as DOI responses for different arrays. A slightly nonlinear dependence of the DOI ratio on depth was observed and the nonlinearity was larger for the corner and edge crystals. The DOI calibration parameters were obtained from the DOI responses measured in a singles mode. The average error between the calibrated DOI and the known DOI was 0.8 mm if a linear detector DOI calibration was used and 0.5 mm if a linear crystal DOI calibration was used. A line source phantom and a hot rod phantom were scanned on the prototype PET scanner. DOI measurement significantly improved the image spatial resolution no matter which DOI calibration method was used. A linear crystal DOI calibration provided slightly better image spatial resolution compared with a linear detector DOI calibration.

We explore dual-ended read out of LSO arrays with two position sensitive avalanche photodiodes (PSAPDs) as a high resolution, high efficiency depth-encoding detector for PET applications. Flood histograms, energy resolution and depth of interaction (DOI) resolution were measured for unpolished LSO arrays with individual crystal sizes of 1.0, 1.3 and 1.5 mm, and for a polished LSO array with 1.3 mm pixels. The thickness of the crystal arrays was 20 mm. Good flood histograms were obtained for all four arrays, and crystals in all four arrays can be clearly resolved. Although the amplitude of each PSAPD signal decreases as the interaction depth moves further from the PSAPD, the sum of the two PSAPD signals is essentially constant with irradiation depth for all four arrays. The energy resolutions were similar for all four arrays, ranging from 14.7% to 15.4%. A DOI resolution of 3-4 mm (including the width of the irradiation band which is approximately 2 mm) was obtained for all the unpolished arrays. The best DOI resolution was achieved with the unpolished 1 mm array (average 3.5 mm). The DOI resolution for the 1.3 mm and 1.5 mm unpolished arrays was 3.7 and 4.0 mm respectively. For the polished array, the DOI resolution was only 16.5 mm. Summing the DOI profiles across all crystals for the 1 mm array only degraded the DOI resolution from 3.5 mm to 3.9 mm, indicating that it may not be necessary to calibrate the DOI response separately for each crystal within an array. The DOI response of individual crystals in the array confirms this finding. These results provide a detailed characterization of the DOI response of these PSAPD-based PET detectors which will be important in the design and calibration of a PET scanner making use of this detector approach.

MicroPET II is a second-generation animal PET scanner designed for high-resolution imaging of small laboratory rodents. The system consists of 90 scintillation detector modules arranged in three contiguous axial rings with a ring diameter of 16.0 cm and an axial length of 4.9 cm. Each detector module consists of a 14 x 14 array of lutetium oxyorthosilicate (LSO) crystals coupled to a multi-channel photomultiplier tube (MC-PMT) through a coherent optical fibre bundle. Each LSO crystal element measures 0.975 mm x 0.975 mm in cross section by 12.5 mm in length. A barium sulphate reflector material was used between LSO elements leading to a detector pitch of 1.15 mm in both axial and transverse directions. Fused optical fibre bundles were made from 90 microm diameter glass fibres with a numerical aperture of 0.56. Interstitial extramural absorber was added between the fibres to reduce optical cross talk. A charge-division readout circuit was implemented on printed circuit boards to decode the 196 crystals in each array from the outputs of the 64 anode signals of the MC-PMT. Electronics from Concorde Microsystems Inc. (Knoxville, TN) were used for signal amplification, digitization, event qualification, coincidence processing and data capture. Coincidence data were passed to a host PC that recorded events in list mode. Following acquisition, data were sorted into sinograms and reconstructed using Fourier rebinning and filtered hackprojection algorithms. Basic evaluation of the system has been completed. The absolute sensitivity of the microPET II scanner was 2.26% at the centre of the field of view (CFOV) for an energy window of 250-750 keV and a timing window of 10 ns. The intrinsic spatial resolution of the detectors in the system averaged 1.21 mm full width at half maximum (FWHM) when measured with a 22Na point source 0.5 mm in diameter. Reconstructed image resolution ranged from 0.83 mm FWHM at the CFOV to 1.47 mm FWHM in the radial direction, 1.17 mm FWHM in the tangential direction and 1.42 mm FWHM in the axial direction at 1 cm offset from the CFOV. These values represent highly significant improvements over our earlier microPET scanner (approximately fourfold sensitivity increase and 25-35% improvement in linear spatial resolution under equivalent operating conditions) and are expected to be further improved when the system is fully optimized.

A rapid radiochemical separation procedure has been finished by using of radiochemical separation and γ-spectrum measurement methods. In this isolation procedure, tantalum nuclides produced by the irradiation of natural tungsten targets with 14 MeV neutrons were extracted by methyl isobutyl ketone (MIBK) in the system of HF-HNO3 mixed solutions. It has been found that tantalum was rapidly and efficiently separated in the isolation procedure and the decontamination factors of tungsten were more than 104. The procedure is simultaneously used to separate tantalum from tungsten bombarded with intermediate energy (47 MeV/u)18O beams. The measured γ-spectra showed that tantalum can be separated from most elements produced in the nuclear reactions except for a few elements.

Our laboratory has developed PET detectors with depth-encoding accuracy of ∼2 mm based on finely pixelated crystals with a tapered geometry, readout at both ends with position-sensitive avalanche photodiodes (PSAPDs). These detectors are currently being used in our laboratory to build a one-ring high resolution PET scanner for mouse brain imaging studies. Due to the inactive areas around the PSAPDs, large gaps exist between the detector modules which can degrade the image spatial resolution obtained using analytical reconstruction with filtered backprojection (FBP). In this work, the Geant4-based GATE Monte Carlo package was used to assist in determining whether gantry rotation was necessary and to assess the expected spatial resolution of the system. The following factors were investigated: rotating vs. static gantry modes with and without compensation of missing data using the discrete cosine transform (DCT) method, two levels of depth-encoding, and positron annihilation effects for 18F. Our results indicate that while the static scanner produces poor quality FBP images with streak and ring artifacts, the image quality was greatly improved after compensation of missing data. The simulation indicates that the expected FWHM system spatial resolution is 0.70 ± 0.05 mm, which approaches the predicted limit of 0.5 mm FWHM due to positron range, photon non-colinearity and physical detector element size effects. We conclude that excellent reconstructed resolution without gantry rotation is possible even using FBP if the gaps are appropriately handled and that this design can approach the resolution limits set by positron annihilation physics.

This paper evaluates the performance of two large-area position-sensitive solid-state photomultipliers (PS-SSPM) for use in small animal PET detector designs. Both PS-SSPM device designs are 1 cm2 in area, the first being a 2 × 2 tiled array of 5 mm × 5 mm PS-SSPMs and the second being a 10 mm × 10 mm continuous PS-SSPM. Signal-to-noise measurements were performed to investigate the optimal operating parameters for each device and to compare the performance of the two PS-SSPM designs. A maximum signal-to-noise ratio of 29.3 was measured for the 5 mm PS-SSPM array and 15.1 for the 10 mm PS-SSPM, both measurements were made at 0 °C and at the optimal bias voltage. The best energy resolution measured with an array of 1.3 mm polished LSO crystals was 16% for the 5 mm PS-SSPM array and 18% for the 10 mm PS-SSPM. The timing properties of both devices were similar, with a best timing resolution (in coincidence with an LSO/PMT detector) of 6.8 ns (range 6.8–8.9 ns) and 7.1 ns (range 7.1–9.6 ns) for the 5 mm PS-SSPM and 10 mm PS-SSPM respectively. The 2 × 2 array of 5 mm PS-SSPMs was able to visually resolve the elements in an 0.5 × 0.5 × 20 mm LYSO scintillator array (unpolished, diffuse reflector) with an average peak-to-valley ratio in the flood histograms of ~11 indicating clear separation of the crystals. Advantages and drawbacks of PET detector designs using PS-SSPM photodetectors are addressed and comparisons to other small-animal PET detector designs using position-sensitive avalanche photodiodes are made.

After many years of development, position-sensitive avalanche photodiodes (PSAPDs) are now being incorporated into a range of scintillation detector systems, including those used in high-resolution small-animal PET and PET/MR scanners. In this work, the signal, noise, signal-to-noise ratio (SNR), flood histogram and timing resolution were measured for lutetium oxyorthosilicate (LSO) scintillator arrays coupled to PSAPDs ranging in size from 10 to 20 mm, and the optimum bias voltage and working temperature were determined. Variations in the SNR performance of PSAPDs with the same dimensions were small, but the SNR decreased significantly with increasing PSAPD size and increasing temperature. Smaller PSAPDs (10 mm and 15 mm in width) produced acceptable flood histograms at 24 °C, and cooling lower than 16 °C produced little improvement. The optimum bias voltage was about 25 V below the break down voltage. The larger 20 mm PSAPDs have lower SNR and require cooling to 0–7 °C for acceptable performance. The optimum bias voltage is also lower (35 V or more below the break down voltage depending on the temperature). Significant changes in the timing resolution were observed as the bias voltage and temperature varied. Higher bias voltages provided better timing resolution. The best timing resolution obtained for individual crystals was 2.8 ns and 3.3 ns for the 10 mm and 15 mm PSAPDs, respectively. The results of this work provide useful guidance for selecting the bias voltage and working temperature for scintillation detectors that incorporate PSAPDs as the photodetector.

We have proposed a zoom-in positron emission tomography (PET) system that incorporates a high-resolution detector into an existing PET scanner to obtain high-resolution images of a region of interest. Previously we have shown by computer simulations that the high-resolution detector can improve the overall system performance in terms of spatial resolution and lesion detectability. In this study, we assessed the resolution improvement in a real system by incorporating a high-resolution detector into our existing microPET II scanner. The high-resolution detector consists of a 14 × 28 array of 0.5 × 0.5 × 10 mm3 lutetium oxyorthosilicate scintillator elements and is placed near the center of the microPET II scanner. It is coupled to two 64-channel photomultiplier tubes (PMTs) via tapered optical fiber bundles. The PMT signals were read out by the electronics in the microPET II scanner. A 15 µCi Na-22 point source was positioned at various locations above the high-resolution detector. Images were reconstructed using the data measured by the microPET II scanner alone and the microPET II data combined with the high-resolution detector data. Profiles taken through the reconstructed point sources show substantial reduction in full-width-at-half-maximum along the direction parallel to the face of the high-resolution detector.

Integrated PET and 3-dimensional (3D) fluorescence optical tomography (FOT) imaging has unique and attractive features for in vivo molecular imaging applications. We have designed, built, and evaluated a simultaneous PET and 3D FOT system. The design of the FOT system is compatible with many existing small-animal PET scanners. Methods: The 3D FOT system comprises a novel conical mirror that is used to view the whole-body surface of a mouse with an electron-multiplying charge-coupled device camera when a collimated laser beam is projected on the mouse to stimulate fluorescence. The diffusion equation was used to model the propagation of optical photons inside the mouse body, and 3D fluorescence images were reconstructed iteratively from the fluorescence intensity measurements measured from the surface of the mouse. Insertion of the conical mirror into the gantry of a small-animal PET scanner allowed simultaneous PET and 3D FOT imaging. Results: The mutual interactions between PET and 3D FOT were evaluated experimentally. PET has negligible effects on 3D FOT performance. The inserted conical mirror introduces a reduction in the sensitivity and noise-equivalent count rate of the PET system and increases the scatter fraction. PET–FOT phantom experiments were performed. An in vivo experiment using both PET and FOT was also performed. Conclusion: Phantom and in vivo experiments demonstrate the feasibility of simultaneous PET and 3D FOT imaging. The first in vivo simultaneous PET–FOT results are reported.

By using detectors with good depth encoding accuracy (~ 2 mm), an animal PET scanner can be built with a small ring diameter and thick crystals to simultaneously obtain high spatial resolution and high sensitivity. However, there will be large wedge-shaped gaps between detector modules in such a scanner if traditional cuboid crystal arrays are used in a polygonal arrangement. The gaps can be minimized by using tapered scintillator arrays enabling the sensitivity of the scanner to be further improved. In this work, tapered lutetium oxyorthosilicate (LSO) arrays with different crystal dimensions and different combinations of inter-crystal reflector and crystal surface treatments were manufactured and their performance was evaluated. Arrays were read out from both ends by position-sensitive avalanche photodiodes (PSAPDs). In the optimal configuration, arrays consisting of 0.5 mm LSO elements could be clearly resolved and a depth of interaction resolution of 2.6 mm was obtained for a 20-mm thick array. For this tapered array, the intrinsic spatial is degraded from 0.67 mm to 0.75 mm compared to a standard cuboidal array with similar dimensions, while the increase in efficiency is 41%. Tapered scintillator arrays offer the prospect of improvements in sensitivity and sampling for small-bore scanners, without large increases in manufacturing complexity.

Small animal PET scanners may be improved by increasing the sensitivity, improving the spatial resolution and improving the uniformity of the spatial resolution across the field of view. This may be achieved by using PET detectors based on crystal elements that are thin in the axial and transaxial directions and long in the radial direction, and by employing depth of interaction (DOI) encoding to minimize the parallax error. With DOI detectors, the diameter of the ring of the PET scanner may also be decreased. This minimizes the number of detectors required to achieve the same solid angle coverage as a scanner with a larger ring diameter and minimizes errors due to non-collinearity of the annihilation photons. In this study, we characterize prototype PET detectors that are finely pixelated with individual LSO crystal element sizes of 0.5 mm x 0.5 mm x 20 mm and 0.7 mm x 0.7 mm x 20 mm, read out at both ends by position sensitive avalanche photodiodes (PSAPDs). Both a specular reflector and a diffuse reflector were evaluated. The detectors were characterized based on the ability to clearly resolve the individual crystal elements, the DOI resolution and the energy resolution. Our results indicate that a scanner based on any of the four detector designs would offer improved spatial resolution and more uniform spatial resolution compared to present day small animal PET scanners. The greatest improvements to spatial resolution will be achieved when the detectors employing the 0.5 mm x 0.5 mm x 20 mm crystals are used. Monte Carlo simulations were performed to demonstrate that 2 mm DOI resolution is adequate to ensure uniform spatial resolution for a small animal PET scanner geometry using these detectors. The sensitivity of such a scanner was also simulated using Monte Carlo simulations and was shown to be greater than 10% for a four ring scanner with an inner diameter of 6 cm, employing 20 detectors per scanner ring.

Over the past decade, many small animal PET scanners have been developed (1–12), and this technology has played a very important role in the rapidly growing field of molecular imaging. High sensitivity is needed to increase signal to noise ratio of the images in order to reliably detect lower levels of radiotracer uptake and to reduce the injected dose (reducing radiation dose to the subject) (13) and scan time (increasing temporal resolution for dynamic studies). High spatial resolution is required to detect small structures and lesions and to improve quantification by reducing the partial volume effect. A compromise between sensitivity and spatial resolution always exists in small animal PET scanner design due to depth of interaction (DOI) effects. Most designs use a large detector ring radius (relative to the subject to be studied) and/or a short crystal length in order to reduce the DOI effects to acceptable levels. But in both cases the sensitivity of the scanner is sacrificed. For this reason, much attention in recent years has focused on detector designs with depth-encoding ability. Multi-layer detectors consisting of crystal layers with different scintillation light decay times (14–16), with different reflector arrangements (17, 18), and using a position shift of half a crystal for different layers (19), have been proposed and successfully implemented to measure DOI. The DOI resolution obtained by these methods is limited by the number of layers that can be decoded, and is typically on the order of 5–10 mm. As a rule of thumb, to eliminate the DOI effect completely, the DOI resolution needs to be on the order of the detector width itself. For small animal PET, the detector widths are typically 0.5–2 mm (8, 10, 20, 21). Therefore, although these designs help reduce DOI-related degradation of spatial resolution, they are still not good enough to allow detectors to be placed in very close proximity to the subject. Thus the classic trade-off between resolution and sensitivity remains, with scanners utilizing thinner detectors and larger ring diameters than desirable. The larger ring diameter also has the undesirable effect of increasing the cost per unit sensitivity. It is therefore important to develop detectors for small animal PET imaging with DOI resolution that approaches the detector width. A promising approach for obtaining higher DOI resolution is to read out the scintillation light by placing photodetectors at both ends of the crystal array and use the ratio of the two photodetector signals to measure the DOI (22–26). An alternative approach being used to estimate DOI for monolithic scintillator crystals measures the distribution of the scintillation light signal on a position-sensitive photodetector arrays (27, 28). In this case events occurring nearer the photodetector produce a narrower spread of scintillation light on the detectors than events occurring far from the photodetector, thus providing a basis for discriminating based on depth. Both of these approaches provide continuous DOI measurement rather than the discrete DOI information provided by the layered approach described earlier. Our approach is to measure DOI by using dual-ended readout of finely pixelated lutetium oxyorthosilicate (LSO) arrays with two position sensitive avalanche photodiodes (PSAPDs) (29–31). Our previous results on individual detector modules have shown that with such a detector at room temperature, crystals arrays down to 1 mm could be resolved and that an energy resolution of about 15% FWHM and a uniform DOI resolution of 3–4 mm FWHM (including the radiation beam width of more than 2 mm) could be achieved (32). In this work the imaging performance of a prototype PET scanner consisting of eight detector modules arranged in two detector plates is evaluated. First, results obtained from individual detectors used within the scanner, including measurements of the flood histograms calculated by using the energy signals from just one PSAPD or both PSAPDs, the performance of the detectors at different temperatures, DOI resolution measured with a smaller radiation beam width of 0.7 mm, and the timing resolution of the PSAPD are presented. Secondly imaging studies are performed with the prototype scanner to measure the effect of the DOI information on reconstructed image resolution.

We propose detectors for a laboratory positron emission tomography (PET) scanner specific for mouse imaging that utilizes fewer detectors and channels of electronics compared with existing designs. The detectors are based on lutetium oxyorthosilicate (LSO) arrays, read out by orthogonal optical fibers placed on the top and bottom of the arrays. Depth of interaction (DOI) information is obtained from the ratio of the signals at either end of the array. Four different detector modules were evaluated, using different reflector materials and two types of optical fibers (wavelength shifting (WLS) fibers and clear optical fibers). The modules were compared in terms of flood histograms, energy resolution, DOI resolution and timing resolution. Energy resolution for single crystals at one irradiation depth was around 65% full-width-half-maximum (FWHM). A DOI resolution of ∼6 mm was obtained for the modules. Timing resolution was in the range of 5.1 ∼ 7.8 ns. An array assembled in the laboratory and coupled with WLS fibers had the best DOI resolution; the same array with clear fibers had the best timing resolution and a commercially manufactured array and coupled with WLS fibers had the best energy resolution.

A microCT scanner was designed, fabricated and integrated with a previously reported microPET II scanner (Tai et al 2003 Phys. Med. Biol. 48 1519, Yang et al 2004 Phys. Med. Biol. 49 2527), forming a dual modality system for in vivo anatomic and molecular imaging of the mouse. The system was designed to achieve high-spatial-resolution and high-sensitivity PET images with adequate CT image quality for anatomic localization and attenuation correction with low x-ray dose. The system also has relatively high throughput for screening, and a flexible gantry and user interface. X-rays were produced by a 50 kVp, 1.5 mA fixed tungsten anode tube, with a focal spot size of 70 microm. The detector was a 5 x 5 cm(2) photodiode detector incorporating 48 microm pixels on a CMOS array and a fast gadolinium oxysulfide (GOS) intensifying screen. The microCT system has a flexible C-arm gantry design with adjustable detector positioning, which acquires CT projection images around the common microPET/CT bed. The design and the initial characterization of the microCT system is described, and images of the first mouse scans with microPET/CT scanning protocols are shown.

Gating firmware and software were developed for the microPET II small animal scanner. The measured cardiac and respiratory signals were collected and converted to TTL gating signals by a Biopac MP150 data acquisition system and sent to microPET II through two BNC connectors on the front panel. During acquisition, the coincidence monitor takes the average of the last eight gate input cycles and inserts this into the list mode data stream on the falling edge of the gating pulse. This value is then used to determine the current time interval of the next gate cycle when the list mode data are sorted into sinograms. The gating firmware and software were validated by an experiment using a rotating point source. Mouse heart (18F-FDG) and bone (18F(-)) imaging was performed with simultaneous cardiac and respiratory gating. It was clearly demonstrated that the contractile function of the mouse heart can be studied by cardiac-gated imaging with microPET II. The left ventricular volumes at different times of the cardiac cycle were measured and the ejection fraction was calculated. In the bone scan, no detectable movement caused by heart contraction was observed. Respiratory motion was more subtle with virtually no motion for more than 75% of the respiratory cycle. The motion of the mouse heart and bones in the thorax caused by respiration was less than 1 mm. It appears with the current resolution of PET, and the small fraction of the respiratory cycle in which motion occurs, that respiratory gating is probably not necessary for most mouse cardiac studies.

MicroPET II is a newly developed PET (positron emission tomography) scanner designed for high-resolution imaging of small animals. It consists of 17,640 LSO crystals each measuring 0.975 x 0.975 x 12.5 mm3, which are arranged in 42 contiguous rings, with 420 crystals per ring. The scanner has an axial field of view (FOV) of 4.9 cm and a transaxial FOV of 8.5 cm. The purpose of this study was to carefully evaluate the performance of the system and to optimize settings for in vivo mouse and rat imaging studies. The volumetric image resolution was found to depend strongly on the reconstruction algorithm employed and averaged 1.1 mm (1.4 microl) across the central 3 cm of the transaxial FOV when using a statistical reconstruction algorithm with accurate system modelling. The sensitivity, scatter fraction and noise-equivalent count (NEC) rate for mouse- and rat-sized phantoms were measured for different energy and timing windows. Mouse imaging was optimized with a wide open energy window (150-750 keV) and a 10 ns timing window, leading to a sensitivity of 3.3% at the centre of the FOV and a peak NEC rate of 235,000 cps for a total activity of 80 MBq (2.2 mCi) in the phantom. Rat imaging, due to the higher scatter fraction, and the activity that lies outside of the field of view, achieved a maximum NEC rate of 24,600 cps for a total activity of 80 MBq (2.2 mCi) in the phantom, with an energy window of 250-750 keV and a 6 ns timing window. The sensitivity at the centre of the FOV for these settings is 2.1%. This work demonstrates that different scanner settings are necessary to optimize the NEC count rate for different-sized animals and different injected doses. Finally, phantom and in vivo animal studies are presented to demonstrate the capabilities of microPET II for small-animal imaging studies.