EJNMMI Physics最新文献

{"title":"Validation of a data-driven motion-compensated PET brain image reconstruction algorithm in clinical patients using four radiotracers.","authors":"Ole L Munk, Anders B Rodell, Patricia B Danielsen, Josefine R Madsen, Mie T Sørensen, Niels Okkels, Jacob Horsager, Katrine B Andersen, Per Borghammer, Joel Aanerud, Judson Jones, Inki Hong, Sven Zuehlsdorff","doi":"10.1186/s40658-025-00723-w","DOIUrl":"10.1186/s40658-025-00723-w","url":null,"abstract":"<p><strong>Purpose: </strong>Patients with dementia symptoms often struggle to limit movements during PET examinations, necessitating motion compensation in brain PET imaging to ensure the high image quality needed for diagnostic accuracy. This study validates a data-driven motion-compensated (MoCo) PET brain image reconstruction algorithm that corrects head motion by integrating the detected motion frames and their associated rigid body transformations into the iterative image reconstruction. Validation was conducted using phantom scans, healthy volunteers, and clinical patients using four radiotracers with distinct tracer activity distributions.</p><p><strong>Methods: </strong>We conducted technical validation experiments of the algorithm using Hoffman brain phantom scans during a series of controlled movements, followed by two blinded reader studies assessing image quality between standard images and MoCo images in 38 clinical patients receiving dementia scans with [¹⁸F]Fluorodeoxyglucose, [¹⁸F]N-(3-iodopro-2E-enyl)-2beta-carbomethoxy-3beta-(4'-methylphenyl)-nortropane, [¹⁸F]flutemetamol, and a research group comprising 25 elderly subjects scanned with [¹⁸F]fluoroethoxybenzovesamicol.</p><p><strong>Results: </strong>The Hoffman brain phantom study demonstrated the algorithm's capability to detect and correct for even minimal movements, 1-mm translations and 1⁰ rotations, applied to the phantom. Within the clinical cohort, where standard images were deemed suboptimal or non-diagnostic, all MoCo images were classified as having acceptable diagnostic quality. In the research cohort, MoCo images consistently matched or surpassed the standard image quality even in cases with minimal head movement, and the MoCo algorithm never led to degraded image quality.</p><p><strong>Conclusion: </strong>The PET brain MoCo reconstruction algorithm was robust and worked well for four different tracers with markedly different uptake patterns. Moco images markedly improved the image quality for patients who were unable to lie still during a PET examination and obviated the need for any repeat scans. Thus, the method was clinically feasible and has the potential for improving diagnostic accuracy.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"11"},"PeriodicalIF":3.0,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11790531/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143079168","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Phantom-based investigation of block sequential regularised expectation maximisation (BSREM) reconstruction for zirconium-89 PET-CT for varied count levels.","authors":"Lara M Bonney, Matthew D Walker, Daniel R McGowan","doi":"10.1186/s40658-025-00722-x","DOIUrl":"10.1186/s40658-025-00722-x","url":null,"abstract":"<p><strong>Background: </strong>Zirconium-89 (Zr-89) PET tracers have become increasingly significant in the field of nuclear medicine due to their 3-day physical half-life, allowing for the study of dynamic biological processes over relatively long timeframes. To date there has been limited publication of studies focused on optimisation of acquisition parameters for Zr-89 PET. This paper outlines a short phantom study investigating the optimal beta regularization parameter for quantitation and noise in block sequential regularised expectation maximisation (BSREM) also known as Bayesian penalized likelihood (BPL) reconstruction, for varying image noise characteristics (acquisition duration).</p><p><strong>Results: </strong>The choice of the beta regularisation parameter substantially impacts image quality and quantitation. For larger volumes, BSREM reconstruction enhanced image quality (lower noise) and maintained quantitation, whereas for smaller volumes quantitation worsened as compared to OSEM for high regularisation parameters.</p><p><strong>Conclusion: </strong>Where BSREM reconstruction is used for Zr-89 images, careful attention must be paid to the choice of weighting factor, especially for quantitative clinical studies. The effect of varying beta on several measures of image quality was characterised for the case of a phantom, with the results indicating that the value of beta for optimal Zr-89 quantitation is lower than what is generally used for optimal visualisation. This work demonstrates the need for careful attention to the reconstruction methods used for quantitative imaging studies, such as those required for theragnostic imaging.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"10"},"PeriodicalIF":3.0,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11790545/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143079167","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A review of state-of-the-art resolution improvement techniques in SPECT imaging.","authors":"Zhibiao Cheng, Ping Chen, Jianhua Yan","doi":"10.1186/s40658-025-00724-9","DOIUrl":"10.1186/s40658-025-00724-9","url":null,"abstract":"<p><p>Single photon emission computed tomography (SPECT), a technique capable of capturing functional and molecular information, has been widely adopted in theranostics applications across various fields, including cardiology, neurology, and oncology. The spatial resolution of SPECT imaging is relatively poor, which poses a significant limitation, especially the visualization of small lesions. The main factors affecting the limited spatial resolution of SPECT include projection sampling techniques, hardware and software. Both hardware and software innovations have contributed substantially to improved SPECT imaging quality. The present review provides an overview of state-of-the-art methods for improving spatial resolution in clinical and pre-clinical SPECT systems. It delves into advancements in detector design and modifications, projection sampling techniques, traditional reconstruction algorithm development and optimization, and the emerging role of deep learning. Hardware enhancements can result in SPECT systems that are both lighter and more compact, while also improving spatial resolution. Software innovations can mitigate the costs of hardware modifications. This survey offers a thorough overview of the rapid advancements in resolution enhancement techniques within the field of SPECT, with the objective of identifying the most recent trends. This is anticipated to facilitate further optimization and improvement of clinical systems, enabling the visualization of small lesions in the early stages of tumor detection, thereby enhancing accurate localization and facilitating both diagnostic imaging and radionuclide therapy, ultimately benefiting both clinicians and patients.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"9"},"PeriodicalIF":3.0,"publicationDate":"2025-01-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11782768/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143064494","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Count-rate management in ¹³¹I SPECT/CT calibration.","authors":"Staffan Jacobsson Svärd, Cecilia Hindorf, Joachim N Nilsson","doi":"10.1186/s40658-025-00718-7","DOIUrl":"10.1186/s40658-025-00718-7","url":null,"abstract":"<p><strong>Background: </strong>System calibration is essential for accurate SPECT/CT dosimetry. However, count losses due to dead time and pulse pileup may cause calibration errors, in particular for ¹³¹I, where high count rates may be encountered. Calibration at low count rates should also be avoided to minimise detrimental effects from e.g. background counts and statistical fluctuations. This paper aims to present experimental data illustrating count-rate dependencies and to propose practical routines to mitigate errors in the ¹³¹I calibration procedure without needing advanced analysis tools.</p><p><strong>Results: </strong>The sensitivities of two General Electric (GE) Discovery 670 Pro systems were assessed using two Jaszczak phantom geometries. SPECT/CT data were collected over two months, starting with an initial ¹³¹I content of > 2 GBq, decaying to approximately 20 MBq. This allowed for a detailed analysis of count losses due to dead time and pulse pileup. From the sensitivity analysis, it was shown that robust calibration was obtained for ¹³¹I phantom activities ranging between 250 and 1500 MBq.</p><p><strong>Conclusions: </strong>The results show that adequate corrections for dead-time and pulse-pileup counting losses are essential for accurate calibration. It is argued that loss corrections should be based on total spectrum count rates in projections and not only on the 364.5 keV energy window data. The measurement campaigns presented in this paper, using basic tools and equipment, may serve as a model for establishing routines for count-loss corrections as well as for system calibration and regular control of system sensitivity. The data suggest that analysis of source and count concentration in a homogeneous Jaszczak phantom offers robust calibration, whereas analysis of source strength and counts in a delineated phantom insert offers a practical and robust method for regular quality control.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"8"},"PeriodicalIF":3.0,"publicationDate":"2025-01-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11769881/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143037613","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exhalation of Rn-219 by patients treated with Radium-223.","authors":"Carsten Wanke, Joerg Pinkert, Lilli Geworski, Bastian Szermerski","doi":"10.1186/s40658-025-00719-6","DOIUrl":"10.1186/s40658-025-00719-6","url":null,"abstract":"<p><strong>Background: </strong>Treatment with Ra-223 dichloride is approved for the therapy of castration resistant prostate cancer (CRPC) with symptomatic bone metastases and no known visceral metastases in Europe since 2013, and Ra-223 is under discussion for labelling other molecules and nanoparticles. The direct progeny of Ra-223 is Rn-219, also known as actinon, a radioactive noble gas with a half-life of 3.98 s. This Rn-219 can be exhaled by patients while Ra-223 is present in the blood. Hence, direct measurements for assessing the exhalation of Rn-219 were performed for the first time in the context of the non-interventional multicenter study \"RAPSODY\", a substudy to the international early access program, which aimed at assessing the radiation exposure of relatives of patients suffering from castration resistant prostate cancer with bone metastases and treated with Ra-223 dichloride in an outpatient setting, in order to investigate if this kind of method is functional and tolerated by the patients.</p><p><strong>Methods: </strong>Rn-219 was measured directly in patients' exhalations using Alphaguard radon monitors (Saphymo, formerly Genitron, Frankfurt, Germany), originally intended for the measurement of Rn-222, and custom-made breath-test kits. Measurements were performed 20-30 min p. i. and 3-4 h p. i. In total, datasets from 21 administrations in 14 patients were obtained.</p><p><strong>Results: </strong>Although 75% of the measurement data 20-30 min p. i. and 35% of the measurement data 3-4 h p. i., respectively, were censored due to exceedance of the upper limit of the Alphaguards' measurement range in the applied measurement setup, statistical data were derived based on the assumption of lognormal distributions. For measurements 3-4 h p. i., mean activity concentrations of Rn-219 in exhaled breath of approx. 4.4 kBq/l were obtained. In measurements 20-30 min p. i., the expectation value of the activity concentration of approx. 6 kBq/l, derived by statistical methods, was higher.</p><p><strong>Conclusions: </strong>Direct measurements using Alphaguard instruments are suitable for assessing the exhalation of Rn-219 by patients treated with Ra-223. The measurement method is well tolerated by the patients. Rn-219 is present in patients' exhalations. Our results are in accordance with published data obtained using other, indirect measurement methods.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"6"},"PeriodicalIF":3.0,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11757652/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143022769","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction: Radiopharmacokinetic modelling and radiation dose assessment of ²²³Ra used for treatment of metastatic castration-resistant prostate cancer.","authors":"Vera Höllriegl, Nina Petoussi-Henss, Kerstin Hürkamp, Juan Camilo Ocampo Ramos, Wei Bo Li","doi":"10.1186/s40658-024-00708-1","DOIUrl":"10.1186/s40658-024-00708-1","url":null,"abstract":"","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"7"},"PeriodicalIF":3.0,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11757809/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143022760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantitative accuracy of preclinical in ovo PET/MRI: influence of attenuation and quantification methods.","authors":"Theresa Balber, Katarína Benčurová, Manuela Mayrhofer, Joachim Friske, Martin Haas, Claudia Kuntner, Thomas H Helbich, Marcus Hacker, Markus Mitterhauser, Ivo Rausch","doi":"10.1186/s40658-024-00714-3","DOIUrl":"10.1186/s40658-024-00714-3","url":null,"abstract":"<p><strong>Aim: </strong>The combination of positron emission tomography (PET) and magnetic resonance imaging (MRI) provides an innovation leap in the use of fertilized chicken eggs (in ovo model) in preclinical imaging as PET/MRI enables the investigation of the chick embryonal organ-specific distribution of PET-tracers. However, hybrid PET/MRI inheres technical challenges in quantitative in ovo PET such as attenuation correction (AC) for the object as well as for additional hardware parts present in the PET field-of-view, which potentially contribute to quantification biases in the PET images if not accounted for. This study aimed to investigate the influence of the different sources of attenuation on in ovo PET/MRI and assess the accuracy of MR-based AC for in ovo experiments.</p><p><strong>Method: </strong>An in-house made chicken egg phantom was used to investigate the magnitude of self-attenuation and the influence of the MRI hardware on the PET signal. The phantom was placed in a preclinical PET/MRI system and PET acquisitions were performed without, and after subsequently adding the different hardware parts to the setup. Reconstructions were performed without any AC for the different setups and with subsequently incorporating the hardware parts into the AC. In addition, in ovo imaging was performed using [¹⁸F]FDG and [⁶⁸Ga]Ga-Pentixafor, and PET data was reconstructed with the different AC combinations. Quantitative accuracy was assessed for the phantom and the in ovo measurements.</p><p><strong>Results: </strong>In general, not accounting for the self-attenuation of the egg and the hardware parts caused an underestimation of the PET signal of around 49% within the egg. Accounting for all sources of attenuation allowed a proper quantification with global offsets of 2% from the true activity. Quantification based on % injected dose per cc (%ID/cc) was similar for the in ovo measurements, regardless of whether hardware parts were included in AC or not, when the injected activity was extracted from the PET images. However, substantial quantification biases were found when the self-attenuation of the egg was not taken into account.</p><p><strong>Conclusion: </strong>Self-attenuation of the egg and PET signal attenuation within the hardware parts of the MRI substantially influence quantitative accuracy in in ovo measurements. However, when compensating for the self-attenuation of the egg by a respective AC, a reliable quantification using %ID/cc can be performed even if not accounting for the attenuation of the hardware parts.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"5"},"PeriodicalIF":3.0,"publicationDate":"2025-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11753441/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143002281","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"On the Effect of the Patient Table on Attenuation in Myocardial Perfusion Imaging SPECT.","authors":"Tamino Huxohl, Gopesh Patel, Wolfgang Burchert","doi":"10.1186/s40658-024-00713-4","DOIUrl":"10.1186/s40658-024-00713-4","url":null,"abstract":"<p><strong>Background: </strong>The topic of the effect of the patient table on attenuation in myocardial perfusion imaging (MPI) SPECT is gaining new relevance due to deep learning methods. Existing studies on this effect are old, rare and only consider phantom measurements, not patient studies. This study investigates the effect of the patient table on attenuation based on the difference between reconstructions of phantom scans and polar maps of patient studies.</p><p><strong>Methods: </strong>Jaszczak phantom scans are acquired according to quality control and MPI procedures. An algorithm is developed to automatically remove the patient table from the CT for attenuation correction. The scans are then reconstructed with attenuation correction either with or without the patient table in the CT. The reconstructions are compared qualitatively and on the basis of their percentage difference. In addition, a small retrospective cohort of 15 patients is examined by comparing the resulting polar maps. Polar maps are compared qualitatively and based on the segment perfusion scores.</p><p><strong>Results: </strong>The phantom reconstructions look qualitatively similar in both the quality control and MPI procedures. The percentage difference is highest in the lower part of the phantom, but it always remains below 17.5%. Polar maps from patient studies also look qualitatively similar. Furthermore, the segment scores are not significantly different (p=0.83).</p><p><strong>Conclusions: </strong>The effect of the patient table on attenuation in MPI SPECT is negligible.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"3"},"PeriodicalIF":3.0,"publicationDate":"2025-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11746982/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143002278","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI-based automatic patient positioning in a digital-BGO PET/CT scanner: efficacy and impact.","authors":"John A Kennedy, Tala Palchan-Hazan, Zohar Keidar","doi":"10.1186/s40658-025-00715-w","DOIUrl":"10.1186/s40658-025-00715-w","url":null,"abstract":"<p><strong>Background: </strong>A recently released digital solid-state positron emission tomography/x-ray CT (PET/CT) scanner with bismuth germanate (BGO) scintillators provides an artificial intelligence (AI) based system for automatic patient positioning. The efficacy of this digital-BGO system in patient placement at the isocenter and its impact on image quality and radiation exposure was evaluated.</p><p><strong>Method: </strong>The digital-BGO PET/CT with AI-based auto-positioning was compared (χ², Mann-Whitney tests) to a solid-state lutetium-yttrium oxyorthosilicate (digital-LYSO) PET/CT with manual patient positioning (n = 432 and 343 studies each, respectively), with results split into groups before and after the date of a recalibration of the digital-BGO auto-positioning camera. To measure the transverse displacement of the patient center from the scanner isocenter (off-centering), CT slices were retrospectively selected and automatically analyzed using in-house software. Noise was measured as the coefficient of variation within the liver of absolute Hounsfield units referenced to air. Radiation exposure was recorded as dose-length product (DLP). Off-centering measurements were validated by a phantom study.</p><p><strong>Results: </strong>The phantom validation study gave < 1.6 mm error in 15 off-centering measurements. Patient off-centering was biased 1.92 ± 1.79 cm (mean ± standard deviation) in the posterior direction which was significantly different from the 0.22 ± 1.21 cm bias in the left lateral direction (p < 0.0001, Wilcoxon). After recalibration, 27% (38/140) of the studies had off-centering results > 2.5cm for the digital-BGO, which was significantly better than the 49% (143/292, p < 0.001) before recalibration and better than for the digital-LYSO: 54% (119/222, p < 0.001) before and 55% (66/121, p < 0.001) after. On average, CT image quality was superior for non-obese patients who were most closely aligned with the isocenter: noise increased by 3.2 ± 0.1% for every 1 cm increase in off-centering. DLP increased by 144 ± 22 Gy cm for every 1 cm increase in anterior off-centering.</p><p><strong>Conclusion: </strong>AI-based automatic patient positioning in a digital-BGO PET/CT scanner significantly reduces patient off-centering, thereby improving image quality and ensuring proper radiation exposure.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"4"},"PeriodicalIF":3.0,"publicationDate":"2025-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11746997/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143002273","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving timing resolution of BGO for TOF-PET: a comparative analysis with and without deep learning.","authors":"Francis Loignon-Houle, Nicolaus Kratochwil, Maxime Toussaint, Carsten Lowis, Gerard Ariño-Estrada, Antonio J Gonzalez, Etiennette Auffray, Roger Lecomte","doi":"10.1186/s40658-024-00711-6","DOIUrl":"10.1186/s40658-024-00711-6","url":null,"abstract":"<p><strong>Background: </strong>The renewed interest in BGO scintillators for TOF-PET is driven by the improved Cherenkov photon detection with new blue-sensitive SiPMs. However, the slower scintillation light from BGO causes significant time walk with leading edge discrimination (LED), which degrades the coincidence time resolution (CTR). To address this, a time walk correction (TWC) can be done by using the rise time measured with a second threshold. Deep learning, particularly convolutional neural networks (CNNs), can also enhance CTR by training with digitized waveforms. It remains to be explored how timing estimation methods utilizing one (LED), two (TWC), or multiple (CNN) waveform data points compare in CTR performance of BGO scintillators.</p><p><strong>Results: </strong>In this work, we compare classical experimental timing estimation methods (LED, TWC) with a CNN-based method using the signals from BGO crystals read out by NUV-HD-MT SiPMs and high-frequency electronics. For <math> <mrow><mrow><mn>2</mn> <mo>×</mo> <mn>2</mn> <mo>×</mo> <mn>3</mn></mrow> <mspace></mspace> <msup><mtext>mm</mtext> <mn>3</mn></msup> </mrow> </math> crystals, implementing TWC results in a CTR of 129 ± 2 ps FWHM, while employing the CNN yields 115 ± 2 ps FWHM, marking improvements of 18 % and 26 %, respectively, relative to the standard LED estimator. For <math> <mrow><mrow><mn>2</mn> <mo>×</mo> <mn>2</mn> <mo>×</mo> <mn>20</mn></mrow> <mspace></mspace> <msup><mtext>mm</mtext> <mn>3</mn></msup> </mrow> </math> crystals, both methods yield similar CTR (around 240 ps FWHM), offering a <math><mo>∼</mo></math> 15 % gain over LED. The CNN, however, exhibits better tail suppression in the coincidence time distribution.</p><p><strong>Conclusions: </strong>The higher complexity of waveform digitization needed for CNNs could potentially be mitigated by adopting a simpler two-threshold approach, which appears to currently capture most of the essential information for improving CTR in longer BGO crystals. Other innovative deep learning models and training strategies may nonetheless contribute further in a near future to harnessing increasingly discernible timing features in TOF-PET detector signals.</p>","PeriodicalId":11559,"journal":{"name":"EJNMMI Physics","volume":"12 1","pages":"2"},"PeriodicalIF":3.0,"publicationDate":"2025-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11739447/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143002276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}