Depth of Interaction Estimation in a Preclinical PET Scanner Equipped with Monolithic Crystals Coupled to SiPMs Using a Deep Neural Network

A deep neural network for positioning and inter-crystal scatter identification in multiplexed PET detectors: a simulation study

Objective.High-resolution positron emission tomography (PET) relies on the accurate positioning of annihilation photons impinging the crystal array. However, conventional positioning algorithms in light-sharing PET detectors are often limited due to edge effects and/or the absence of additional information for identifying and correcting scattering within the crystal array (known as inter-crystal scattering). This study explores the feasibility of deep neural network (DNN) techniques for more precise event positioning in finely segmented and highly multiplexed PET detectors with light-sharing.Approach.Initially, a Geant4 Application for Tomographic Emission (GATE) simulation was used to study the spatial and statistical properties of inter-crystal scatter (ICS) events in finely segmented LYSO PET detectors. Next, a DNN for crystal localisation was designed, trained and tested with light distributions of photoelectric (P) and Compton + photoelectric (CP) events simulated using optical GATE and an analytical method to speed up data generation. Using the statistical properties of ICS events, an energy-guided positioning algorithm was then built into the DNN. The positioning algorithm enables selection of the unique or first crystal of interaction in P and CP events, respectively. Performance of the DNN was compared with Anger logic using light distributions from simulated 511 keV point sources placed at different locations around a single PET detector module.Main results. The fraction of events forward and backward scattered in the LYSO detector was 0.54 and 0.46, respectively, whereas naïve application of the Klein-Nishina formulation predicts 70% forward scatter. Despite coarse photodetector data due to signal multiplexing, the DNN demonstrated a crystal classification accuracy of 90% for P events and 82% for CP events. For crystal positioning, the DNN outperformed Anger logic by at least 34% and 14% for P and CP events, respectively. Further improvement is somewhat constrained by the physics-specifically, the ratio of backward to forward scattering of gamma rays within the crystal array being close to 1. This prevents selecting the first crystal of interaction in CP events with a high degree of certainty.Significance.Light sharing and multiplexed PET detectors are common in high-resolution PET, yet their traditional positioning algorithms often underperform due to edge effects and/or the difficulty in correcting ICS events. Our study indicates that DNN-based event positioning has the potential to enhance 2D coincidence event positioning accuracy by nearly a factor of 3 compared to Anger logic. However, further improvements are difficult to foresee without additional information such as event timing.

Evaluation of monolithic crystal detector with dual-ended readout utilizing multiplexing method

Objective.Monolithic crystal detectors are increasingly being applied in positron emission tomography (PET) devices owing to their excellent depth-of-interaction (DOI) resolution capabilities and high detection efficiency. In this study, we constructed and evaluated a dual-ended readout monolithic crystal detector based on a multiplexing method.Approach.We employed two 12 × 12 silicon photomultiplier (SiPM) arrays for readout, and the signals from the 12 × 12 array were merged into 12 X and 12 Y channels using channel multiplexing. In 2D reconstruction, three methods based on the centre of gravity (COG) were compared, and the concept of thresholds was introduced. Furthermore, a light convolutional neural network (CNN) was employed for testing. To enhance depth localization resolution, we proposed a method by utilizing the mutual information from both ends of the SiPMs. The source width and collimation effect were simulated using GEANT4, and the intrinsic spatial resolution was separated from the measured values.Main results.At an operational voltage of 29 V for the SiPM, an energy resolution of approximately 12.5 % was achieved. By subtracting a 0.8 % threshold from the total energy in every channel, a 2D spatial resolution of approximately 0.90 mm full width at half maximum (FWHM) can be obtained. Furthermore, a higher level of resolution, approximately 0.80 mm FWHM, was achieved using a CNN, with some alleviation of edge effects. With the proposed DOI method, a significant 1.36 mm FWHM average DOI resolution can be achieved. Additionally, it was found that polishing and black coating on the crystal surface yielded smaller edge effects compared to a rough surface with a black coating.Significance.The introduction of a threshold in COG method and a dual-ended readout scheme can lead to excellent spatial resolution for monolithic crystal detectors, which can help to develop PET systems with both high sensitivity and high spatial resolution.

The quest for multifunctional and dedicated PET instrumentation with irregular geometries

We focus on reviewing state-of-the-art developments of dedicated PET scanners with irregular geometries and the potential of different aspects of multifunctional PET imaging. First, we discuss advances in non-conventional PET detector geometries. Then, we present innovative designs of organ-specific dedicated PET scanners for breast, brain, prostate, and cardiac imaging. We will also review challenges and possible artifacts by image reconstruction algorithms for PET scanners with irregular geometries, such as non-cylindrical and partial angular coverage geometries and how they can be addressed. Then, we attempt to address some open issues about cost/benefits analysis of dedicated PET scanners, how far are the theoretical conceptual designs from the market/clinic, and strategies to reduce fabrication cost without compromising performance.

Applications of machine learning and deep learning in SPECT and PET imaging: General overview, challenges and future prospects

The integration of positron emission tomography (PET) and single-photon emission computed tomography (SPECT) imaging techniques with machine learning (ML) algorithms, including deep learning (DL) models, is a promising approach. This integration enhances the precision and efficiency of current diagnostic and treatment strategies while offering invaluable insights into disease mechanisms. In this comprehensive review, we delve into the transformative impact of ML and DL in this domain. Firstly, a brief analysis is provided of how these algorithms have evolved and which are the most widely applied in this domain. Their different potential applications in nuclear imaging are then discussed, such as optimization of image adquisition or reconstruction, biomarkers identification, multimodal fusion and the development of diagnostic, prognostic, and disease progression evaluation systems. This is because they are able to analyse complex patterns and relationships within imaging data, as well as extracting quantitative and objective measures. Furthermore, we discuss the challenges in implementation, such as data standardization and limited sample sizes, and explore the clinical opportunities and future horizons, including data augmentation and explainable AI. Together, these factors are propelling the continuous advancement of more robust, transparent, and reliable systems.

Deep-TOF-PET: Deep learning-guided generation of time-of-flight from non-TOF brain PET images in the image and projection domains

We aim to synthesize brain time-of-flight (TOF) PET images/sinograms from their corresponding non-TOF information in the image space (IS) and sinogram space (SS) to increase the signal-to-noise ratio (SNR) and contrast of abnormalities, and decrease the bias in tracer uptake quantification. One hundred forty clinical brain 18 F-FDG PET/CT scans were collected to generate TOF and non-TOF sinograms. The TOF sinograms were split into seven time bins (0, ±1, ±2, ±3). The predicted TOF sinogram was reconstructed and the performance of both models (IS and SS) compared with reference TOF and non-TOF. Wide-ranging quantitative and statistical analysis metrics, including structural similarity index metric (SSIM), root mean square error (RMSE), as well as 28 radiomic features for 83 brain regions were extracted to evaluate the performance of the CycleGAN model. SSIM and RMSE of 0.99 ± 0.03, 0.98 ± 0.02 and 0.12 ± 0.09, 0.16 ± 0.04 were achieved for the generated TOF-PET images in IS and SS, respectively. They were 0.97 ± 0.03 and 0.22 ± 0.12, respectively, for non-TOF-PET images. The Bland & Altman analysis revealed that the lowest tracer uptake value bias (-0.02%) and minimum variance (95% CI: -0.17%, +0.21%) were achieved for TOF-PET images generated in IS. For malignant lesions, the contrast in the test dataset was enhanced from 3.22 ± 2.51 for non-TOF to 3.34 ± 0.41 and 3.65 ± 3.10 for TOF PET in SS and IS, respectively. The implemented CycleGAN is capable of generating TOF from non-TOF PET images to achieve better image quality.

Active-PET: a multifunctional PET scanner with dynamic gantry size featuring high-resolution and high-sensitivity imaging: a Monte Carlo simulation study

Organ-specific PET scanners have been developed to provide both high spatial resolution and sensitivity, although the deployment of several dedicated PET scanners at the same center is costly and space-consuming. Active-PET is a multifunctional PET scanner design exploiting the advantages of two different types of detector modules and mechanical arms mechanisms enabling repositioning of the detectors to allow the implementation of different geometries/configurations. Active-PET can be used for different applications, including brain, axilla, breast, prostate, whole-body, preclinical and pediatrics imaging, cell tracking, and image guidance for therapy. Monte Carlo techniques were used to simulate a PET scanner with two sets of high resolution and high sensitivity pixelated Lutetium Oxyorthoscilicate (LSO(Ce)) detector blocks (24 for each group, overall 48 detector modules for each ring), one with large pixel size (4 × 4 mm2) and crystal thickness (20 mm), and another one with small pixel size (2 × 2 mm2) and thickness (10 mm). Each row of detector modules is connected to a linear motor that can displace the detectors forward and backward along the radial axis to achieve variable gantry diameter in order to image the target subject at the optimal/desired resolution and/or sensitivity. At the center of the field-of-view, the highest sensitivity (15.98 kcps MBq−1) was achieved by the scanner with a small gantry and high-sensitivity detectors while the best spatial resolution was obtained by the scanner with a small gantry and high-resolution detectors (2.2 mm, 2.3 mm, 2.5 mm FWHM for tangential, radial, and axial, respectively). The configuration with large-bore (combination of high-resolution and high-sensitivity detectors) achieved better performance and provided higher image quality compared to the Biograph mCT as reflected by the 3D Hoffman brain phantom simulation study. We introduced the concept of a non-static PET scanner capable of switching between large and small field-of-view as well as high-resolution and high-sensitivity imaging.

The Use of Artificial Neural Networks for Determining Values of Selected Strength Parameters of Miscanthus × Giganteus

The aim of this paper is to develop neural models enabling the determination of biomechanical parameters for giant miscanthus stems. The static three-point bending test is used to determine the bending strength parameters of the miscanthus stem. In this study, we assume the modulus of elasticity bending and maximum stress in bending as the dependent variables. As independent variables (inputs of the neural network) we assume water content, internode number, maximum bending force value and dimensions characterizing the cross-section of miscanthus stem: maximum and minimum stem diameter and stem wall thickness. The four developed neural models, enabling the determination of the value of the modulus of elasticity in bending and the maximum stress in bending, demonstrate sufficient and even very high accuracy. The neural networks have an average relative error of 2.18%, 2.21%, 3.24% and 0.18% for all data subsets, respectively. The results of the sensitivity analysis confirmed that all input variables are important for the accuracy of the developed neural models—correct semantic models.

Classification of Alzheimer’s Disease and Mild-Cognitive Impairment Base on High-Order Dynamic Functional Connectivity at Different Frequency Band

Functional brain connectivity networks obtained from resting-state functional magnetic resonance imaging (rs-fMRI) have been extensively utilized for the diagnosis of Alzheimer’s disease (AD). However, the traditional correlation analysis technique only explores the pairwise relation, which may not be suitable for revealing sufficient and proper functional connectivity links among brain regions. Additionally, previous literature typically focuses on only lower-order dynamics, without considering higher-order dynamic networks properties, and they particularly focus on single frequency range time series of rs-fMRI. To solve these problems, in this article, a new diagnosis scheme is proposed by constructing a high-order dynamic functional network at different frequency level time series (full-band (0.01–0.08 Hz); slow-4 (0.027–0.08 Hz); and slow-5 (0.01–0.027 Hz)) data obtained from rs-fMRI to build the functional brain network for all brain regions. Especially, to tune the precise analysis of the regularized parameters in the Support Vector Machine (SVM), a nested leave-one-out cross-validation (LOOCV) technique is adopted. Finally, the SVM classifier is trained to classify AD from HC based on these higher-order dynamic functional brain networks at different frequency ranges. The experiment results illustrate that for all bands with a LOOCV classification accuracy of 94.10% with a 90.95% of sensitivity, and a 96.75% of specificity outperforms the individual networks. Utilization of the given technique for the identification of AD from HC compete for the most state-of-the-art technology in terms of the diagnosis accuracy. Additionally, results obtained for the all-band shows performance further suggest that our proposed scheme has a high-rate accuracy. These results have validated the effectiveness of the proposed methods for clinical value to the identification of AD.

Unbiased TOF estimation using leading-edge discriminator and convolutional neural network trained by single-source-position waveforms

Objective. Convolutional neural networks (CNNs) are a strong tool for improving the coincidence time resolution (CTR) of time-of-flight (TOF) positron emission tomography detectors. However, several signal waveforms from multiple source positions are required for CNN training. Furthermore, there is concern that TOF estimation is biased near the edge of the training space, despite the reduced estimation variance (i.e. timing uncertainty). Approach. We propose a simple method for unbiased TOF estimation by combining a conventional leading-edge discriminator (LED) and a CNN that can be trained with waveforms collected from one source position. The proposed method estimates and corrects the time difference error calculated by the LED rather than the absolute time difference. This model can eliminate the TOF estimation bias, as the combination with the LED converts the distribution of the label data from discrete values at each position into a continuous symmetric distribution. Main results. Evaluation results using signal waveforms collected from scintillation detectors show that the proposed method can correctly estimate all source positions without bias from a single source position. Moreover, the proposed method improves the CTR of the conventional LED. Significance. We believe that the improved CTR will not only increase the signal-to-noise ratio but will also contribute significantly to a part of the direct positron emission imaging.

TCAN-IDS: Intrusion Detection System for Internet of Vehicle Using Temporal Convolutional Attention Network

Intrusion detection systems based on recurrent neural network (RNN) have been considered as one of the effective methods to detect time-series data of in-vehicle networks. However, building a model for each arbitration bit is not only complex in structure but also has high computational overhead. Convolutional neural network (CNN) has always performed excellently in processing images, but they have recently shown great performance in learning features of normal and attack traffic by constructing message matrices in such a manner as to achieve real-time monitoring but suffer from the problem of temporal relationships in context and inadequate feature representation in key regions. Therefore, this paper proposes a temporal convolutional network with global attention to construct an in-vehicle network intrusion detection model, called TCAN-IDS. Specifically, the TCAN-IDS model continuously encodes 19-bit features consisting of an arbitration bit and data field of the original message into a message matrix, which is symmetric to messages recalling a historical moment. Thereafter, the feature extraction model extracts its spatial-temporal detail features. Notably, global attention enables global critical region attention based on channel and spatial feature coefficients, thus ignoring unimportant byte changes. Finally, anomalous traffic is monitored by a two-class classification component. Experiments show that TCAN-IDS demonstrates high detection performance on publicly known attack datasets and is able to accomplish real-time monitoring. In particular, it is anticipated to provide a high level of symmetry between information security and illegal intrusion.

Rural Built-Up Area Extraction from Remote Sensing Images Using Spectral Residual Methods with Embedded Deep Neural Network

A rural built-up area is one of the most important features of rural regions. Rapid and accurate extraction of rural built-up areas has great significance to rural planning and urbanization. In this paper, the spectral residual method is embedded into a deep neural network to accurately describe the rural built-up areas from large-scale satellite images. Our proposed method is composed of two processes: coarse localization and fine extraction. Firstly, an improved Faster R-CNN (Regions with Convolutional Neural Network) detector is trained to obtain the coarse localization of the candidate built-up areas, and then the spectral residual method is used to describe the accurate boundary of each built-up area based on the bounding boxes. In the experimental part, we firstly explored the relationship between the sizes of built-up areas and the kernels in the spectral residual method. Then, the comparing experiments demonstrate that our proposed method has better performance in the extraction of rural built-up areas.

S3D: Squeeze and Excitation 3D Convolutional Neural Networks for a Fall Detection System

Because of the limitations of previous studies on a fall detection system (FDS) based on wearable and ambient devices and visible light and depth cameras, the research using thermal cameras has recently been conducted. However, they also have the problem of deteriorating the accuracy of FDS depending on various environmental changes. Given these facts, in this study, we newly propose an FDS method based on the squeeze and excitation (SE) 3D convolutional neural networks (S3D). In our method, keyframes are extracted from input thermal videos using the optical flow vectors, and the fall detection is carried out based on the output of the proposed S3D, using the extracted keyframes as input. Comparative experiments were carried out on three open databases of thermal videos with different image resolutions, and our proposed method obtained F1 scores of 97.14%, 95.30%, and 98.89% in the Thermal Simulated Fall, Telerobotics and Control Lab fall detection, and eHomeSeniors datasets, respectively (the F1 score is a harmonic mean of recall and precision; it was confirmed that these are superior results to those obtained using the state-of-the-art methods of a thermal camera-based FDS.

A universal methodology for reliable predicting the non-steroidal anti-inflammatory drug solubility in supercritical carbon dioxide

Understanding the drug solubility behavior is likely the first essential requirement for designing the supercritical technology for pharmaceutical processing. Therefore, this study utilizes different machine learning scenarios to simulate the solubility of twelve non-steroidal anti-inflammatory drugs (NSAIDs) in the supercritical carbon dioxide (SCCO₂). The considered NSAIDs are Fenoprofen, Flurbiprofen, Ibuprofen, Ketoprofen, Loxoprofen, Nabumetone, Naproxen, Nimesulide, Phenylbutazone, Piroxicam, Salicylamide, and Tolmetin. Physical characteristics of the drugs (molecular weight and melting temperature), operating conditions (pressure and temperature), and solvent property (SCCO₂ density) are effectively used to estimate the drug solubility. Monitoring and comparing the prediction accuracy of twelve intelligent paradigms from three categories (artificial neural networks, support vector regression, and hybrid neuro-fuzzy) approves that adaptive neuro-fuzzy inference is the best tool for the considered task. The hybrid optimization strategy adjusts the cluster radius of the subtractive clustering membership function to 0.6111. This model estimates 254 laboratory-measured solubility data with the AAPRE = 3.13%, MSE = 2.58 × 10^–9, and R² = 0.99919. The leverage technique confirms that outliers may poison less than four percent of the experimental data. In addition, the proposed hybrid paradigm is more reliable than the equations of state and available correlations in the literature. Experimental measurements, model predictions, and relevancy analyses justified that the drug solubility in SCCO₂ increases by increasing temperature and pressure. The results show that Ibuprofen and Naproxen are the most soluble and insoluble drugs in SCCO₂, respectively.

Reducing Calibration Time in PET Systems Based on Monolithic Crystals

In the past years, the gamma-ray detector designs based on the monolithic crystals have demonstrated to be excellent candidates for the design of high-performance PET systems. The monolithic crystals allow to achieve the intrinsic detector resolutions well below state-of-the-art; to increase packing fraction thus, increasing the system sensitivity; and to improve lesion detectability at the edges of the scanner field of view (FOV) because of their intrinsic depth of interaction (DOI) capabilities. The bottleneck to translate to the clinical PET systems based on a large number of monolithic detectors is eventually the requirement of mechanically complex and time-consuming calibration processes. To mitigate this drawback, several methods have been already proposed, such as using non-physically collimated radioactive sources or implementing the neuronal networks (NN) algorithms trained with simulated data. In this work, we aimed to simplify and fasten a calibration process of the monolithic based systems. The Normal procedure consists of individually acquiring a 11 × 11 ²²Na source array for all the detectors composing the PET system and obtaining the calibration map for each module using a method based on the Voronoi diagrams. Two reducing time methodologies are presented: (i) TEST1, where the calibration map of one detector is estimated and shared among all others, and (ii) TEST2, where the calibration map is slightly modified for each module as a function of their detector uniformity map. The experimental data from a dedicated prostate PET system was used to compare the standard calibration procedure with both the proposed methods. A greater similarity was exhibited between the TEST2 methodology and the Normal procedure; obtaining spatial resolution variances within 0.1 mm error bars and count rate deviations as small as 0.2%. Moreover, the negligible reconstructed image differences (13% deviation at most in the contrast-to-noise ratio) and almost identical contrast values were reported. Therefore, this proposed method allows us to calibrate the PET systems based on the monolithic crystals reducing the calibration time by approximately 80% compared with the Normal procedure.

Application of Artificial Intelligence in PET Instrumentation

Artificial intelligence (AI) has been widely used throughout medical imaging, including PET, for data correction, image reconstruction, and image processing tasks. However, there are number of opportunities for the application of AI in photon detector performance or the data collection process, such as to improve detector spatial resolution, time-of-flight information, or other PET system performance characteristics. This review outlines current topics, research highlights, and future directions of AI in PET instrumentation.

DeepTOFSino: A deep learning model for synthesizing full-dose time-of-flight bin sinograms from their corresponding low-dose sinograms

PURPOSE

Reducing the injected activity and/or the scanning time is a desirable goal to minimize radiation exposure and maximize patients' comfort. To achieve this goal, we developed a deep neural network (DNN) model for synthesizing full-dose (FD) time-of-flight (TOF) bin sinograms from their corresponding fast/low-dose (LD) TOF bin sinograms.

METHODS

Clinical brain PET/CT raw data of 140 normal and abnormal patients were employed to create LD and FD TOF bin sinograms. The LD TOF sinograms were created through 5% undersampling of FD list-mode PET data. The TOF sinograms were split into seven time bins (0, ±1, ±2, ±3). Residual network (ResNet) algorithms were trained separately to generate FD bins from LD bins. An extra ResNet model was trained to synthesize FD images from LD images to compare the performance of DNN in sinogram space (SS) vs implementation in image space (IS). Comprehensive quantitative and statistical analysis was performed to assess the performance of the proposed model using established quantitative metrics, including the peak signal-to-noise ratio (PSNR), structural similarity index metric (SSIM) region-wise standardized uptake value (SUV) bias and statistical analysis for 83 brain regions.

RESULTS

SSIM and PSNR values of 0.97 ± 0.01, 0.98 ± 0.01 and 33.70 ± 0.32, 39.36 ± 0.21 were obtained for IS and SS, respectively, compared to 0.86 ± 0.02and 31.12 ± 0.22 for reference LD images. The absolute average SUV bias was 0.96 ± 0.95% and 1.40 ± 0.72% for SS and IS implementations, respectively. The joint histogram analysis revealed the lowest mean square error (MSE) and highest correlation (R² = 0.99, MSE = 0.019) was achieved by SS compared to IS (R² = 0.97, MSE= 0.028). The Bland & Altman analysis showed that the lowest SUV bias (-0.4%) and minimum variance (95% CI: -2.6%, +1.9%) were achieved by SS images. The voxel-wise t-test analysis revealed the presence of voxels with statistically significantly lower values in LD, IS, and SS images compared to FD images respectively.

CONCLUSION

The results demonstrated that images reconstructed from the predicted TOF FD sinograms using the SS approach led to higher image quality and lower bias compared to images predicted from LD images.

Development of dual-ended depth-of-interaction detectors using laser-induced crystals for small animal PET systems

Sensitivity and spatial resolution of positron emission tomography (PET) scanners can be improved by using thicker scintillation crystals with depth-of-interaction (DOI) encoding. Subsurface laser engraving (SSLE) can be used to segment crystals of a scintillation detector in order to fabricate a DOI detector. We previously applied SSLE to crystal bars of 3 × 3 × 20 mm³and 1.5 × 1.5 × 20 mm³and developed two dual-ended detectors with DOI segments of 3 mm and 1.5 mm, respectively. To further improve the DOI resolution, our SSLE detector design can be used with smaller pitch crystal bars, making them excellent detector candidates for small animal PET scanners with submillimetre resolution. In the present study, three small crystal bars of 1 × 1 × 20 mm³, 2 × 1 × 20 mm³, and 2 × 1 × 40 mm³were laser engraved to 12, 20 and 40 segments, respectively, by applying SSLE in their height directions. The segmented crystal bars were characterised in three prototype detector arrangements. First, the 1 × 1 × 20 mm³crystal bars were characterised in an 8 × 8 crystal array designed for DOI encoding along crystal height in a conventional small animal PET design. Second, a 4 × 8 crystal array of 2 × 1 × 20 mm³crystal bars was characterised for using the DOI information for crystal interaction positioning along the axial axis of a small animal PET scanner. Finally, the third part of the study was performed on a single 2 × 1 × 40 mm³crystal bar with 40 segments to investigate the feasibility of DOI estimation in longer crystals for application in a system with extended axial length. We evaluated the capability of segment identification and energy resolution of theses detectors. The 3D position maps of the detectors were obtained using the Anger-type calculation and the crystal identification performance was evaluated for each detector. Clear segment separation was obtained for the crystal arrays with 12 (segment pitch of 1.67 mm) and 20 (segment pitch of 1 mm) segments. Mean energy resolutions of 8.8% ± 0.4% and 9.6% ± 0.8% at 511 keV were obtained for the segments in the central regions of the 8 × 8 array with 12 segments and the 4 × 8 array with 20 segments, respectively. Clear segment identification was found to be difficult for the detector with 40 segments, especially for the segments at the middle of the crystal. Energy and interaction positioning characterisation results suggest that both prototype detectors with 12 and 20 segments are well suited for small animal PET scanners with high spatial resolution.

Artificial Intelligence in PET: An Industry Perspective

Artificial intelligence (AI) has significant potential to positively impact and advance medical imaging, including positron emission tomography (PET) imaging applications. AI has the ability to enhance and optimize all aspects of the PET imaging chain from patient scheduling, patient setup, protocoling, data acquisition, detector signal processing, reconstruction, image processing, and interpretation. AI poses industry-specific challenges which will need to be addressed and overcome to maximize the future potentials of AI in PET. This article provides an overview of these industry-specific challenges for the development, standardization, commercialization, and clinical adoption of AI and explores the potential enhancements to PET imaging brought on by AI in the near future. In particular, the combination of on-demand image reconstruction, AI, and custom-designed data-processing workflows may open new possibilities for innovation which would positively impact the industry and ultimately patients.

Deep residual-convolutional neural networks for event positioning in a monolithic annular PET scanner

PET scanners based on monolithic pieces of scintillator can potentially produce superior performance characteristics (high spatial resolution and detection sensitivity, for example) compared to conventional PET scanners. Consequently, we initiated development of a preclinical PET system based on a single 7.2 cm long annulus of LYSO, called AnnPET. While this system could facilitate creation of high-quality images, its unique geometry results in optics that can complicate estimation of event positioning in the detector. To address this challenge, we evaluated deep-residual convolutional neural networks (DR-CNN) to estimate the three-dimensional position of annihilation photon interactions. Monte Carlo simulations of the AnnPET scanner were used to replicate the physics, including optics, of the scanner. It was determined that a ten-layer-DR-CNN was most suited to application with AnnPET. The errors between known event positions, and those estimated by this network and those calculated with the commonly used center-of-mass algorithm (COM) were used to assess performance. The mean absolute errors (MAE) for the ten-layer-DR-CNN-based event positions were 0.54 mm, 0.42 mm and 0.45 mm along thex(axial)-,y(transaxial)- andz- (depth-of-interaction) axes, respectively. For COM estimates, the MAEs were 1.22 mm, 1.04 mm and 2.79 mm in thex-,y- andz-directions, respectively. Reconstruction of the network-estimated data with the 3D-FBP algorithm (5 mm source offset) yielded spatial resolutions (full-width-at-half-maximum (FWHM)) of 0.8 mm (radial), 0.7 mm (tangential) and 0.71 mm (axial). Reconstruction of the COM-derived data yielded spatial resolutions (FWHM) of 1.15 mm (radial), 0.96 mm (tangential) and 1.14 mm (axial). These findings demonstrated that use of a ten-layer-DR-CNN with a PET scanner based on a monolithic annulus of scintillator has the potential to produce excellent performance compared to standard analytical methods.

Quantitative Molecular Positron Emission Tomography Imaging Using Advanced Deep Learning Techniques

The widespread availability of high-performance computing and the popularity of artificial intelligence (AI) with machine learning and deep learning (ML/DL) algorithms at the helm have stimulated the development of many applications involving the use of AI-based techniques in molecular imaging research. Applications reported in the literature encompass various areas, including innovative design concepts in positron emission tomography (PET) instrumentation, quantitative image reconstruction and analysis techniques, computer-aided detection and diagnosis, as well as modeling and prediction of outcomes. This review reflects the tremendous interest in quantitative molecular imaging using ML/DL techniques during the past decade, ranging from the basic principles of ML/DL techniques to the various steps required for obtaining quantitatively accurate PET data, including algorithms used to denoise or correct for physical degrading factors as well as to quantify tracer uptake and metabolic tumor volume for treatment monitoring or radiation therapy treatment planning and response prediction.This review also addresses future opportunities and current challenges facing the adoption of ML/DL approaches and their role in multimodality imaging.

Polaroid-PET: a PET scanner with detectors fitted with Polaroid for filtering unpolarized optical photons-a Monte Carlo simulation study

We propose and evaluate the performance of an improved preclinical positron emission tomography (PET) scanner design, referred to as Polaroid-PET, consisting of a detector equipped with a layer of horizontal Polaroid to filter scintillation photons with vertical polarization. This makes it possible to improve the spatial resolution of PET scanners based on monolithic crystals. First, a detector module based on a lutetium-yttrium orthosilicate monolithic crystal with 10 mm thickness and silicon photomultipliers (SiPMs) was implemented in the GEANT4 Monte Carlo toolkit. Subsequently, a layer of Polaroid was inserted between the crystal and the SiPMs. Two preclinical PET scanners based on ten detector modules with and without Polaroid were simulated. The performance of the proposed detector modules and corresponding PET scanner for the two configurations (with and without Polaroid) was assessed using standard performance parameters, including spatial resolution, sensitivity, optical photon ratio detected for positioning, and image quality. The detector module fitted with Polaroid led to higher spatial resolution (1.05 mm FWHM) in comparison with a detector without Polaroid (1.30 mm FHWM) for a point source located at the center of the detector module. From 100% of optical photons produced in the scintillator crystal, 65% and 66% were used for positioning in the detectors without and with Polaroid, respectively. Polaroid-PET resulted in higher axial spatial resolution (0.83 mm FWHM) compared to the scanner without Polaroid (1.01 mm FWHM) for a point source at the center of the field of view (CFOV). The absolute sensitivity at the CFOV was 4.37% and 4.31% for regular and Polaroid-PET, respectively. Planar images of a grid phantom demonstrated the potential of the detector with a Polaroid in distinguishing point sources located at close distances. Our results indicated that Polaroid-PET may improve spatial resolution by filtering the reflected optical photons according to their polarization state, while retaining the high sensitivity expected with monolithic crystal detector blocks.

Applications of artificial intelligence and deep learning in molecular imaging and radiotherapy

This brief review summarizes the major applications of artificial intelligence (AI), in particular deep learning approaches, in molecular imaging and radiation therapy research. To this end, the applications of artificial intelligence in five generic fields of molecular imaging and radiation therapy, including PET instrumentation design, PET image reconstruction quantification and segmentation, image denoising (low-dose imaging), radiation dosimetry and computer-aided diagnosis, and outcome prediction are discussed. This review sets out to cover briefly the fundamental concepts of AI and deep learning followed by a presentation of seminal achievements and the challenges facing their adoption in clinical setting.