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

Predicting time-of-flight with Cerenkov light in BGO: a three-stage network approach with multiple timing kernels prior

Objective.In the quest for enhanced image quality in positron emission tomography (PET) reconstruction, the introduction of time-of-flight (TOF) constraints in TOF-PET reconstruction offers superior signal-to-noise ratio. By employing BGO detectors capable of simultaneously emitting prompt Cerenkov light and scintillation light, this approach combines the high time resolution of prompt photons with the high energy resolution of scintillation light, thereby presenting a promising avenue for acquiring more precise TOF information.Approach.In Stage One, we train a raw method capable of predicting TOF information based on coincidence waveform pairs. In Stage Two, the data is categorized into 25 classes based on signal rise time, and the pre-trained raw method is utilized to obtain TOF kernels for each of the 25 classes, thereby generating prior knowledge. Within Stage Three, our proposed deep learning (DL) module, combined with a bias fine-tuning module, utilizes the kernel prior to provide bias compensation values for the data, thereby refining the first-stage outputs and obtaining more accurate TOF predictions.Main results.The three-stage network built upon the LED method resulted in improvements of 11.7 ps and 41.8 ps for full width at half maximum (FWHM) and full width at tenth maximum (FWTM), respectively. Optimal performance was achieved with FWHM of 128.2 ps and FWTM of 286.6 ps when CNN and Transformer were utilized in Stages One and Three, respectively. Further enhancements of 2.3 ps and 3.5 ps for FWHM and FWTM were attained through data augmentation methods.Significance.This study employs neural networks to compensate for the timing delays in mixed (Cerenkov and scintillation photons) signals, combining multiple timing kernels as prior knowledge with DL models. This integration yields optimal predictive performance, offering a superior solution for TOF-PET research utilizing Cerenkov signals.

Holistic evaluation of a machine learning-based timing calibration for PET detectors under varying data sparsity

Objective.Modern PET scanners offer precise TOF information, improving the SNR of the reconstructed images. Timing calibrations are performed to reduce the worsening effects of the system components and provide valuable TOF information. Traditional calibration procedures often provide static or linear corrections, with the drawback that higher-order skews or event-to-event corrections are not addressed. Novel research demonstrated significant improvements in the reachable timing resolutions when combining conventional calibration approaches with machine learning, with the disadvantage of extensive calibration times infeasible for a clinical application. In this work, we made the first steps towards an in-system application and analyzed the effects of varying data sparsity on a machine learning timing calibration, aiming to accelerate the calibration time. Furthermore, we demonstrated the versatility of our calibration concept by applying the procedure for the first time to analog readout technology.Approach.We modified experimentally acquired calibration data used for training regarding their statistical and spatial sparsity, mimicking reduced measurement time and variability of the training data. Trained models were tested on unseen test data, characterized by fine spatial sampling and rich statistics. In total, 80 decision tree models with the same hyperparameter settings, were trained and holistically evaluated regarding data scientific, physics-based, and PET-based quality criteria.Main results.The calibration procedure can be heavily reduced from several days to some minutes without sacrificing quality and still significantly improving the timing resolution from(304±5)psto(216±1)pscompared to conventionally used analytical calibration methods.Significance.This work serves as the first step in making the developed machine learning-based calibration suitable for an in-system application to profit from the method's capabilities on the system level. Furthermore, this work demonstrates the functionality of the methodology on detectors using analog readout technology. The proposed holistic evaluation criteria here serve as a guideline for future evaluations of machine learning-based calibration approaches.

Transformer-CNN hybrid network for improving PET time of flight prediction

Objective.In positron emission tomography (PET) reconstruction, the integration of time-of-flight (TOF) information, known as TOF-PET, has been a major research focus. Compared to traditional reconstruction methods, the introduction of TOF enhances the signal-to-noise ratio of images. Precision in TOF is measured by full width at half maximum (FWHM) and the offset from ground truth, referred to as coincidence time resolution (CTR) and bias.Approach.This study proposes a network combining transformer and convolutional neural network (CNN) to utilize TOF information from detector waveforms, using event waveform pairs as inputs. This approach integrates the global self-attention mechanism of Transformer, which focuses on temporal relationships, with the local receptive field of CNN. The combination of global and local information allows the network to assign greater weight to the rising edges of waveforms, thereby extracting valuable temporal information for precise TOF predictions. Experiments were conducted using lutetium yttrium oxyorthosilicate (LYSO) scintillators and silicon photomultiplier (SiPM) detectors. The network was trained and tested using the waveform datasets after cropping.Main results.Compared to the constant fraction discriminator (CFD), CNN, CNN with attention, long short-term memory (LSTM) and Transformer, our network achieved an average CTR of 189 ps, reducing it by 82 ps (more than 30%), 13 ps (6.4%), 12 ps (6.0%), 16 ps (7.8%) and 9 ps (4.6%), respectively. Additionally, a reduction of 10.3, 8.7, 6.7 and 4 ps in average bias was achieved compared to CNN, CNN with attention, LSTM and Transformer.Significance.This work demonstrates the potential of applying the Transformer for PET TOF estimation using real experimental data. Through the integration of both CNN and Transformer with local and global attention, it achieves optimal performance, thereby presenting a novel direction for future research in this field.

Whole Reconstruction-Free System Design for Direct Positron Emission Imaging From Image Generation to Attenuation Correction

Direct positron emission imaging (dPEI), which does not require a mathematical reconstruction step, is a next-generation molecular imaging modality. To maximize the practical applicability of the dPEI system to clinical practice, we introduce a novel reconstruction-free image-formation method called direct μCompton imaging, which directly localizes the interaction position of Compton scattering from the annihilation photons in a three-dimensional space by utilizing the same compact geometry as that for dPEI, involving ultrafast time-of-flight radiation detectors. This unique imaging method not only provides the anatomical information about an object but can also be applied to attenuation correction of dPEI images. Evaluations through Monte Carlo simulation showed that functional and anatomical hybrid images can be acquired using this multimodal imaging system. By fusing the images, it is possible to simultaneously access various object data, which ensures the synergistic effect of the two imaging methodologies. In addition, attenuation correction improves the quantification of dPEI images. The realization of the whole reconstruction-free imaging system from image generation to quantitative correction provides a new perspective in molecular imaging.

Deep learning-based PET image denoising and reconstruction: a review

This review focuses on positron emission tomography (PET) imaging algorithms and traces the evolution of PET image reconstruction methods. First, we provide an overview of conventional PET image reconstruction methods from filtered backprojection through to recent iterative PET image reconstruction algorithms, and then review deep learning methods for PET data up to the latest innovations within three main categories. The first category involves post-processing methods for PET image denoising. The second category comprises direct image reconstruction methods that learn mappings from sinograms to the reconstructed images in an end-to-end manner. The third category comprises iterative reconstruction methods that combine conventional iterative image reconstruction with neural-network enhancement. We discuss future perspectives on PET imaging and deep learning technology.

Advancements in Positron Emission Tomography Detectors: From Silicon Photomultiplier Technology to Artificial Intelligence Applications

This review article focuses on PET detector technology, which is the most crucial factor in determining PET image quality. The article highlights the desired properties of PET detectors, including high detection efficiency, spatial resolution, energy resolution, and timing resolution. Recent advancements in PET detectors to improve these properties are also discussed, including the use of silicon photomultiplier technology, advancements in depth-of-interaction and time-of-flight PET detectors, and the use of artificial intelligence for detector development. The article provides an overview of PET detector technology and its recent advancements, which can significantly enhance PET image quality.

Simulation study on 3D convolutional neural networks for time-of-flight prediction in monolithic PET detectors using digitized waveforms

Objective. We investigate the use of 3D convolutional neural networks for gamma arrival time estimation in monolithic scintillation detectors. Approach. The required data is obtained by Monte Carlo simulation in GATE v8.2, based on a 50 × 50 × 16 mm3 monolithic LYSO crystal coupled to an 8 × 8 readout array of silicon photomultipliers (SiPMs). The electronic signals are simulated as a sum of bi-exponentional functions centered around the scintillation photon detection times. We include various effects of statistical fluctuations present in non-ideal SiPMs, such as dark counts and limited photon detection efficiency. The data was simulated for two distinct overvoltages of the SensL J-Series 60 035 SiPMs, in order to test the effects of different SiPM parameters. The neural network uses the array of detector waveforms, digitized at 10 GS s−1, to predict the time at which the gamma arrived at the crystal. Main results. Best results were achieved for an overvoltage of +6 V, at which point the SiPM reaches its optimal photon detection efficiency, resulting in a coincidence time resolution (CTR) of 141 ps full width at half maximum (FWHM). It is a 26% improvement compared to a simple averaging of the first few SiPM timestamps obtained by leading edge discrimination, which in comparison produced a CTR of 177 ps FWHM. In addition, better detector uniformity was achieved, although some degradation near the corners did remain. Significance. These improvements in time resolution can lead to higher signal-to-noise ratios in time-of-flight positron emission tomography, ultimately resulting in better diagnostic capabilities.