Ryosuke Ota

Objective . Scintillation detectors with excellent timing resolution enable more precise localization of radiation sources in positron emission tomography, leading to substantial improvements in diagnostic capability for diseases such as cancer and dementia. At the timing precision required for such applications at the picosecond scale, detector performance is governed by the dynamics of scintillation photons generated within the detector and their subsequent detection processes. However, detector signals have conventionally been treated only as collective responses of many photons due to structural constraints inherent to photodetectors. In this study, we overcome this fundamental limitation using deep learning framework. Approach . The proposed method estimates photon-by-photon arrival times directly from detector waveforms without requiring any modification to the detector structure; the method operates on an event-by-event basis without ground-truth labels by integrating an unsupervised learning framework with a physically informed detector-response model. Main results . Through comprehensive validation combining Monte Carlo simulation and experimental measurements across various detector configurations, we experimentally demonstrate improved timing resolution, visualized depth-of-interaction-dependent photon transport, and classified Cherenkov and scintillation photons based on the estimated photon-level timing information using a unified deep learning-based framework. Significance . These results provide experimental access to photon dynamics, bridging the gap between theoretical modeling and experimental observation, and they open a new data-driven pathway for discovery in detector physics and optimization.

Objective. To enhance the image quality in time-of-flight positron emission tomography (TOF-PET) reconstruction by improving the accuracy of TOF estimation through an unsupervised waveform reshaping method based on cycle-consistent generative adversarial networks (CycleGANs). This method aims to transform distorted waveforms acquired from real detector responses into clean, virtual waveforms generated by Geant4 simulations, thereby significantly improving the signal-to-noise ratio in PET imaging. Approach. We developed a real event signal dataset using a ²² Na gamma source and LYSO–silicon photomultiplier detector pairs, and proposed an unsupervised waveform reshaping method based on CycleGAN. The framework learns to map experimentally acquired waveforms—distorted by noise, electronic delays, and scintillator response broadening—to clean, virtual waveforms generated by Geant4 simulations that model ideal detector responses. By using cycle-consistency constraints, the model performs this transformation without requiring paired or labeled data. The reshaped waveforms retain essential timing characteristics while reducing distortions, thereby enabling more accurate TOF estimation through standard timing pick-off methods such as constant fraction discrimination (CFD). Main results. After applying CFD timing to the reshaped waveforms, the results show a coincidence time resolution improvement of 41 ps (15.1%) over conventional CFD applied to raw signals, and 81 ps (26.0%) over leading edge discrimination. Moreover, the proposed method achieves lower edge bias compared to supervised convolutional neural network-based approaches. Significance. Our unsupervised TOF enhancement framework eliminates the need for large-scale labeled datasets and reduces hardware dependency by using simulated virtual waveforms. The CycleGAN-based architecture enables efficient, generalizable performance across source positions, offering a practical and scalable strategy for advancing high-precision TOF-PET systems.

In positron emission tomography (PET), time-of-flight (TOF) information localizes source positions along lines of response. Cherenkov-radiator-integrated microchannel-plate photomultiplier tubes have achieved 30 ps TOF resolution, demonstrating cross-sectional imaging without reconstruction. Such ultrafast TOF detectors would free PET from conventional ring geometries. Therefore, this study aimed at investigating imaging characteristics of a dual-panel PET with ultrafast TOF detectors using Geant4 simulation. Two detector panels (137 W 137 mm2), which consisted of 5.0 mm-thick bismuth germanate pixelized crystals with a 5.75 mm pitch, were placed face-to-face at a 300 mm distance. Imaging characteristics with various TOF resolutions from 30 to 90 ps were evaluated. Because degraded efficiency may cancel TOF gain in image quality, detection efficiency was also parameterized by reducing coincidence counts. Data acquisitions for a numerical multi-rod and uniform phantom (21 MBq) and a modified NEMA NU2 image quality phantom were simulated for 600 s. Results of the maximum likelihood expectation maximization (MLEM) reconstruction were compared with those of a backprojection (i.e., no reconstruction). The dual-panel PET required a 40 ps TOF resolution to have a similar spatial resolution to that of a non-TOF ring PET (300 mm in diameter) for the same detection efficiency. TOF showed benefit in the reconstruction of image quality phantom with 40% efficiency, and the image noise with 20% efficiency at 30 ps TOF was similar to the complete efficiency at 40 ps TOF. MLEM provided better imaging performance than backprojection, even at 30 ps TOF. The feasibility of the proposed dual-panel PET was shown.

Objective. Lutetium oxide (Lu ₂ O ₃ ), with its high density (9.4 g cm ⁻³ ), presents a compelling scintillation host for detecting 511 keV annihilation photons in positron emission tomography (PET). Despite its favorable density, the practical deployment of Lu ₂ O ₃ -based scintillators for PET has faced limitations due to difficulties in crystal growth and inappropriate decay time. Recent progress in ceramic processing has facilitated the development of transparent Lu ₂ O ₃ ceramics, while targeted doping strategies have significantly improved their luminescence performance. This study evaluates the performance of Lu ₂ O ₃ :Yb and a newly developed ceramic scintillator of (Lu,Y) ₂ O ₃ :La, a modified Lu ₂ O ₃ -based compound incorporating yttrium (Y) and doped with lanthanum (La). Approach. Various ceramic disks were fabricated and cut into 3 × 3 × 5 mm ³ samples. The performance of both Lu ₂ O ₃ :Yb and (Lu,Y) ₂ O ₃ :La ceramic samples in terms of decay time, energy resolution, and coincidence timing resolution (CTR) was assessed. Decay time measurements were conducted using waveform data collected from samples mounted on an H10580 photomultiplier tube (PMT) and irradiated with 511 keV photons from a ²² Na source. Energy and CTRs were evaluated using both PMT and silicon photomultiplier setups, arranged in coincidence with a reference lutetium–yttrium oxyorthosilicate (LYSO) detector of the same size. Main results. All three (Lu,Y) ₂ O ₃ :La ceramic scintillator samples exhibited a triple exponential decay profile and were dominated by a slow component ranging from 1379.3 to 1515.6 ns. The best energy resolution of 15.4% at 511 keV and the best CTR of 237.9 ps full width at half maximum (FWHM) were observed for the same sample. In contrast, a fast decay time of 1.6 ns was observed for the Lu ₂ O ₃ :Yb samples, which exhibited CTR values ranging from 237.9 ps to 261.4 ps FWHM, while the photopeak at 511 keV was difficult to distinguish. These CTR values were estimated between two identical ceramic samples, derived from coincidence measurements of each ceramic sample against the LYSO reference detector. The (Lu,Y) ₂ O ₃ :La samples achieved CTR values comparable to those of the Lu ₂ O ₃ :Yb samples, as their much higher light yield offsets the disadvantage associated with their slower decay time. Significance. These results highlight the promising potential of the (Lu,Y) ₂ O ₃ :La ceramic scintillators for PET applications, especially for time-of-flight PET.

Objective. Coincidence time resolution (CTR) in time-of-flight positron emission tomography (TOF-PET) has significantly improved with advancements in scintillators, photodetectors, and readout electronics. Achieving a CTR of 100 ps remains challenging due to the need for sufficiently thick scintillators—typically 20 mm—to ensure adequate sensitivity because the photon transit time spread within these thick scintillators impedes achieving 100 ps CTR. Therefore thinner scintillators are preferable for CTR better than 100 ps. To address the trade-off between TOF capability and sensitivity, we propose a readout scheme of PET detectors. Approach. The proposed scheme utilizes two orthogonally stacked one-dimensional PET detectors, enabling the thickness of the scintillators to be reduced to approximately 13 mm without compromising sensitivity. This is achieved by stacking the detectors along the depth-of-interaction (DOI) axis of a PET scanner. We refer to this design as the cross-stacked detector, or xDetector. Furthermore, the xDetector inherently provides DOI information using the same readout scheme. Main results. Experimental evaluations demonstrated that the xDetector achieved the best CTR of 175 ps full width at half maximum (FWHM) and an energy resolution of 11% FWHM at 511 keV with 3 × 3 × 12.8 mm³ lutetium oxyorthosilicate crystals, each coupled one-to-one with silicon photomultipliers. The CTRs are between the xDetector and reference detector with a single timing resolution of 111.2 ± 0.8 ps FWHM. In terms of xy-spatial resolution, the xDetector exhibited an asymmetric resolution due to its readout scheme: one resolution was defined by the 3.2 mm readout pitch, while the other was calculated using the center-of-gravity method. Significance. The xDetector effectively resolves the trade-off between TOF capability and sensitivity while offering scalability and DOI capability. By integrating state-of-the-art scintillators, photodetectors, and readout electronics with the xDetector scheme, achieving a CTR of 100 ps FWHM alongside high DOI resolution becomes a practical possibility.

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.

Bismuth germanate (BGO) has been receiving attention again because it is a potential scintillator for future time-of-flight positron emission tomography. Owing to its optical properties, BGO emits a relatively large number of Cherenkov photons after 511 keV gamma-ray interactions, which can enable good coincidence time resolution. Nonetheless, optimally exploiting the Cherenkov emissions can be confounded by scintillation emissions. Thus, we propose a method efficiently emphasizing Cherenkov photon from a detector waveform by deconvolving a single photon response of photodetector. As a proof-of-concept, we perform the deconvolution, and a probability density function (PDF) of BGO was obtained, which is compared to a conventional time correlated single photon counting method. Furthermore, we investigate if the proposed deconvolution can emphasize a faint Cherenkov signal. Consequently, the PDF obtained by the proposed deconvolution shows a good agreement with that obtained using a conventional time correlated single photon counting methods. A coincidence time resolution obtained using the proposed deconvolution is improved by 12% and 43% in FWHM compared to a voltage-based leading edge discriminator for with and without high-frequency readout electronics, respectively. Thus, the proposed deconvolution method can efficiently emphasize Cherenkov photon by lowering the threshold level and improve the timing performance of BGO-based detectors.

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 μ _Compton 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.