Pub Date : 2025-03-02DOI: 10.1109/TRPMS.2025.3561406
{"title":"IEEE Transactions on Radiation and Plasma Medical Sciences Information for Authors","authors":"","doi":"10.1109/TRPMS.2025.3561406","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3561406","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 5","pages":"C3-C3"},"PeriodicalIF":4.6,"publicationDate":"2025-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10982370","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143900520","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-02DOI: 10.1109/TRPMS.2025.3566630
Jun Hou;Tianqi Chen;Yinchi Zhou;Xiongchao Chen;Huidong Xie;Qiong Liu;Menghua Xia;Vladimir Y. Panin;Takuya Toyonaga;Chi Liu;Bo Zhou
Attenuation correction (AC) is a critical step to ensure accurate quantitative Positron Emission Tomography (PET) imaging. To eliminate the radiation dose from CT, deep learning (DL)-based methods have been extensively investigated to generate the CT-equivalent attenuation map (<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-CT) directly from the PET signal. However, almost all previous studies only focus on <inline-formula> <tex-math>${}^{18}text {F-FDG}$ </tex-math></inline-formula> due to its extensive data availability which is suitable for DL model training. For other less common tracer types, it is generally believed that new models must be trained separately on these tracer-specific data to ensure reasonable performance. In this work, we explored the cross-tracer generalizability of <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-DL generation DL models - primarily focusing on whether a model trained on a commonly used tracer like <inline-formula> <tex-math>${}^{18}text {F-FDG}$ </tex-math></inline-formula> can be effectively applied to less common tracers, such as <inline-formula> <tex-math>${}^{68}text {Ga-DOTATE}$ </tex-math></inline-formula> and <inline-formula> <tex-math>${}^{18}text {F-Fluciclovine}$ </tex-math></inline-formula>, and vice versa. Unlike methods that directly generate attenuation-corrected (AC) PET images from nonattenuation corrected (NAC) PET images or maximum likelihood reconstruction of activity and attenuation (MLAA) reconstructions, we generate the CT-based DL attenuation maps (<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-DL) using MLAA reconstruction with the combined input of attenuation maps (<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-MLAA) and tracer activity (<inline-formula> <tex-math>$lambda $ </tex-math></inline-formula>-MLAA). This <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-DL is then used for AC to obtain the final AC PET image. Our comprehensive evaluations on both <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-CT generation and the PET reconstruction found that the DL model trained on one specific tracer can be adapted to other tracers with competitive performance when compared to the tracer-specific trained DL model. The <inline-formula> <tex-math>${}^{18}text {F-FDG}$ </tex-math></inline-formula>-trained model demonstrated the best generalizability when applied to less common tracer types which often have relatively fewer available data for training. Additionally, we investigated the role of the <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-MLAA and <inline-formula> <tex-math>$lambda $ </tex-math></inline-formula>-MLAA as inputs for the network performance. We found that combining both inputs resulted in the best performance, but the <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-MLAA contributed more significantly compared to the <inline-formula> <tex-math>$lambda $ </tex
{"title":"An Investigation on Cross-Tracer Generalizability of Deep Learning-Based PET Attenuation Correction","authors":"Jun Hou;Tianqi Chen;Yinchi Zhou;Xiongchao Chen;Huidong Xie;Qiong Liu;Menghua Xia;Vladimir Y. Panin;Takuya Toyonaga;Chi Liu;Bo Zhou","doi":"10.1109/TRPMS.2025.3566630","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3566630","url":null,"abstract":"Attenuation correction (AC) is a critical step to ensure accurate quantitative Positron Emission Tomography (PET) imaging. To eliminate the radiation dose from CT, deep learning (DL)-based methods have been extensively investigated to generate the CT-equivalent attenuation map (<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-CT) directly from the PET signal. However, almost all previous studies only focus on <inline-formula> <tex-math>${}^{18}text {F-FDG}$ </tex-math></inline-formula> due to its extensive data availability which is suitable for DL model training. For other less common tracer types, it is generally believed that new models must be trained separately on these tracer-specific data to ensure reasonable performance. In this work, we explored the cross-tracer generalizability of <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-DL generation DL models - primarily focusing on whether a model trained on a commonly used tracer like <inline-formula> <tex-math>${}^{18}text {F-FDG}$ </tex-math></inline-formula> can be effectively applied to less common tracers, such as <inline-formula> <tex-math>${}^{68}text {Ga-DOTATE}$ </tex-math></inline-formula> and <inline-formula> <tex-math>${}^{18}text {F-Fluciclovine}$ </tex-math></inline-formula>, and vice versa. Unlike methods that directly generate attenuation-corrected (AC) PET images from nonattenuation corrected (NAC) PET images or maximum likelihood reconstruction of activity and attenuation (MLAA) reconstructions, we generate the CT-based DL attenuation maps (<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-DL) using MLAA reconstruction with the combined input of attenuation maps (<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-MLAA) and tracer activity (<inline-formula> <tex-math>$lambda $ </tex-math></inline-formula>-MLAA). This <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-DL is then used for AC to obtain the final AC PET image. Our comprehensive evaluations on both <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-CT generation and the PET reconstruction found that the DL model trained on one specific tracer can be adapted to other tracers with competitive performance when compared to the tracer-specific trained DL model. The <inline-formula> <tex-math>${}^{18}text {F-FDG}$ </tex-math></inline-formula>-trained model demonstrated the best generalizability when applied to less common tracer types which often have relatively fewer available data for training. Additionally, we investigated the role of the <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-MLAA and <inline-formula> <tex-math>$lambda $ </tex-math></inline-formula>-MLAA as inputs for the network performance. We found that combining both inputs resulted in the best performance, but the <inline-formula> <tex-math>$mu $ </tex-math></inline-formula>-MLAA contributed more significantly compared to the <inline-formula> <tex-math>$lambda $ </tex","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 8","pages":"1025-1035"},"PeriodicalIF":3.5,"publicationDate":"2025-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435701","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-21DOI: 10.1109/TRPMS.2025.3542994
Jan Debus;Werner Lustermann;Afroditi Eleftheriou;Matthias Wyss;Bruno Weber;Günther Dissertori
SAFIR-II is a preclinical PET insert compatible with a Bruker BioSpec 70/30 magnetic resonance imaging (MRI) scanner. It was designed to acquire data at activities of up to 500 MBq, enabling truly simultaneous preclinical positron emission tomography magnetic resonance imaging for mice and rats using image acquisition times of as little as 5 s. We present a brief overview of the system’s design as well as the results of several performance evaluations. SAFIR-II features an axial field-of-view (FOV) of 145 mm, covered by lutetium-yttrium oxyorthosilicate crystals coupled to Hamamatsu silicon photomultiplier (SiPM) arrays. PETA8 application-specific integrated circuits are used to digitize the SiPM’s analog signals, and custom MR-compatible dc-dc converters condition the system’s internal voltages. The insert exhibits a coincidence timing resolution of 221-ps full width at half maximum (FWHM), a coincidence energy resolution of 12.1%, and a peak sensitivity of 3.89% observed following the NEMA-NU4 standard. It is capable of resolving 1.7-mm hot rods within a Derenzo phantom filled with $^{18}{mathrm { F}}$ and features a peak noise-equivalent count rate of 1.12 Mcps observed at an activity of 451 MBq using the NEMA rat-like phantom. We furthermore present an evaluation of the system’s image quality determined using a NEMA image quality phantom, an evaluation of its MRI-compatibility, as well as images from an initial in vivo measurement using a Sprague-Dawley rat injected with 283-MBq fluordesoxyglucose.
{"title":"SAFIR-II: Performance Evaluation of a High-Rate Preclinical PET-MR System","authors":"Jan Debus;Werner Lustermann;Afroditi Eleftheriou;Matthias Wyss;Bruno Weber;Günther Dissertori","doi":"10.1109/TRPMS.2025.3542994","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3542994","url":null,"abstract":"SAFIR-II is a preclinical PET insert compatible with a Bruker BioSpec 70/30 magnetic resonance imaging (MRI) scanner. It was designed to acquire data at activities of up to 500 MBq, enabling truly simultaneous preclinical positron emission tomography magnetic resonance imaging for mice and rats using image acquisition times of as little as 5 s. We present a brief overview of the system’s design as well as the results of several performance evaluations. SAFIR-II features an axial field-of-view (FOV) of 145 mm, covered by lutetium-yttrium oxyorthosilicate crystals coupled to Hamamatsu silicon photomultiplier (SiPM) arrays. PETA8 application-specific integrated circuits are used to digitize the SiPM’s analog signals, and custom MR-compatible dc-dc converters condition the system’s internal voltages. The insert exhibits a coincidence timing resolution of 221-ps full width at half maximum (FWHM), a coincidence energy resolution of 12.1%, and a peak sensitivity of 3.89% observed following the NEMA-NU4 standard. It is capable of resolving 1.7-mm hot rods within a Derenzo phantom filled with <inline-formula> <tex-math>$^{18}{mathrm { F}}$ </tex-math></inline-formula> and features a peak noise-equivalent count rate of 1.12 Mcps observed at an activity of 451 MBq using the NEMA rat-like phantom. We furthermore present an evaluation of the system’s image quality determined using a NEMA image quality phantom, an evaluation of its MRI-compatibility, as well as images from an initial in vivo measurement using a Sprague-Dawley rat injected with 283-MBq fluordesoxyglucose.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"951-958"},"PeriodicalIF":3.5,"publicationDate":"2025-02-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144996102","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Time resolution in time-of-flight positron emission tomography (TOF-PET) has improved significantly over the last decade due to advancements in scintillation materials, photodetectors, and readout electronics, which has increased the signal-to-noise ratio (SNR) compared to conventional positron emission tomography. Silicon photomultipliers (SiPMs) in TOF-PET detectors are often operated at high bias voltage to improve the time performance at the expense of increasing signal noise. SiPM noise, both correlated and uncorrelated, can cause baseline fluctuations, leading to time-walk effects when a leading edge trigger strategy is used, and thus limiting timing performance. We examined the effect of SiPM baseline fluctuations using the FastIC ASIC, a scalable multichannel readout for fast timing applications. We flagged noisy events by using a comparator signal triggered by dark counts before the actual scintillation event. We tested different classification and correction methods with scintillating crystals and Cherenkov radiators, coupled to analog SiPMs from Broadcom (NUV-MT) and Hamamatsu Photonics. We reduced the coincidence time resolution (CTR) in bismuth germanate $2times 2times $ 3 mm3 (BGO) crystals from $410~pm ~10$ to $388~pm ~10$ ps FWHM (5%) by correcting the time-walk on the noisy events. We measured an improvement from $107~pm 2$ to $93.5~pm ~0.6$ ps (11%) for LYSO $2times 2times $ 3 mm3 crystals by filtering the noisy events. An improvement of 9% on the CTR of the EJ232 plastic scintillator was also achieved by filtering noisy events, reducing it from $82.2~pm ~0.5$ to $75~pm ~1$ ps. This study presents a scalable method for flagging undesired events in a full TOF-PET system and discusses the impact of SiPM noise on the FastIC readout.
{"title":"Improving CTR With the FastIC ASIC for TOF-PET by Overcoming SiPM Noise With Baseline Correction","authors":"Afonso Silvério Xavier De Matos Pinto;Nicolaus Kratochwil;Sergio Gómez;David Gascón;Pedro Correia;João Veloso;Emilie Roncali;Ana Luísa Silva;Gerard Ariño-Estrada","doi":"10.1109/TRPMS.2025.3532794","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3532794","url":null,"abstract":"Time resolution in time-of-flight positron emission tomography (TOF-PET) has improved significantly over the last decade due to advancements in scintillation materials, photodetectors, and readout electronics, which has increased the signal-to-noise ratio (SNR) compared to conventional positron emission tomography. Silicon photomultipliers (SiPMs) in TOF-PET detectors are often operated at high bias voltage to improve the time performance at the expense of increasing signal noise. SiPM noise, both correlated and uncorrelated, can cause baseline fluctuations, leading to time-walk effects when a leading edge trigger strategy is used, and thus limiting timing performance. We examined the effect of SiPM baseline fluctuations using the FastIC ASIC, a scalable multichannel readout for fast timing applications. We flagged noisy events by using a comparator signal triggered by dark counts before the actual scintillation event. We tested different classification and correction methods with scintillating crystals and Cherenkov radiators, coupled to analog SiPMs from Broadcom (NUV-MT) and Hamamatsu Photonics. We reduced the coincidence time resolution (CTR) in bismuth germanate <inline-formula> <tex-math>$2times 2times $ </tex-math></inline-formula>3 mm3 (BGO) crystals from <inline-formula> <tex-math>$410~pm ~10$ </tex-math></inline-formula> to <inline-formula> <tex-math>$388~pm ~10$ </tex-math></inline-formula> ps FWHM (5%) by correcting the time-walk on the noisy events. We measured an improvement from <inline-formula> <tex-math>$107~pm 2$ </tex-math></inline-formula> to <inline-formula> <tex-math>$93.5~pm ~0.6$ </tex-math></inline-formula> ps (11%) for LYSO <inline-formula> <tex-math>$2times 2times $ </tex-math></inline-formula>3 mm3 crystals by filtering the noisy events. An improvement of 9% on the CTR of the EJ232 plastic scintillator was also achieved by filtering noisy events, reducing it from <inline-formula> <tex-math>$82.2~pm ~0.5$ </tex-math></inline-formula> to <inline-formula> <tex-math>$75~pm ~1$ </tex-math></inline-formula> ps. This study presents a scalable method for flagging undesired events in a full TOF-PET system and discusses the impact of SiPM noise on the FastIC readout.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"857-865"},"PeriodicalIF":3.5,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10893703","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144998050","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Radiopharmaceutical therapy (RPT) is an established treatment modality and is of increasing interest for different cancer types. A key unmet need, both in the wider adoption of RPT and in the improvement of outcomes with existing RPTs, is in treatment planning and optimization. Research efforts have been hindered by the incomplete understanding of the radiobiology RPTs. Modeling in RPT often mirrors external beam radiotherapy (EBRT), despite key differences. The dose rate (DR) is notably distinct between the two, influencing radiation responses. In EBRT, radiation is acutely delivered in discrete transient fractions of relatively short duration, with a near-constant DR. In RPT, by contrast, exposure is gradual, protracted, and characterized by temporal nonuniformities arising from organ-specific radio-pharmacokinetics. As a result, low-DR (LDR) radiobiology adapted for RPT (LDR-RPT) has emerged as a vibrant area of research. In this review, we discuss the state-of-the-art understanding of the etiological mechanisms underlying cellular and tissue-level dose responses in LDR-RPT, with a focus on how this radiobiological knowledge is codified in mathematical and computational models. We also describe current feasibility and future prospects for utilizing such quantitative radiobiological models to perform personalized RPT planning and highlight research directions that should be prioritized to accelerate clinical adoption.
{"title":"Shifting the Spotlight to Low-Dose Rate Radiobiology in Radiopharmaceutical Therapies: Mathematical Modeling, Challenges, and Future Directions","authors":"Hamid Abdollahi;Babak Saboury;Tahir Yusufaly;Ian Alberts;Carlos Uribe;Arman Rahmim","doi":"10.1109/TRPMS.2025.3540739","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3540739","url":null,"abstract":"Radiopharmaceutical therapy (RPT) is an established treatment modality and is of increasing interest for different cancer types. A key unmet need, both in the wider adoption of RPT and in the improvement of outcomes with existing RPTs, is in treatment planning and optimization. Research efforts have been hindered by the incomplete understanding of the radiobiology RPTs. Modeling in RPT often mirrors external beam radiotherapy (EBRT), despite key differences. The dose rate (DR) is notably distinct between the two, influencing radiation responses. In EBRT, radiation is acutely delivered in discrete transient fractions of relatively short duration, with a near-constant DR. In RPT, by contrast, exposure is gradual, protracted, and characterized by temporal nonuniformities arising from organ-specific radio-pharmacokinetics. As a result, low-DR (LDR) radiobiology adapted for RPT (LDR-RPT) has emerged as a vibrant area of research. In this review, we discuss the state-of-the-art understanding of the etiological mechanisms underlying cellular and tissue-level dose responses in LDR-RPT, with a focus on how this radiobiological knowledge is codified in mathematical and computational models. We also describe current feasibility and future prospects for utilizing such quantitative radiobiological models to perform personalized RPT planning and highlight research directions that should be prioritized to accelerate clinical adoption.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"843-856"},"PeriodicalIF":3.5,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10891667","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144998067","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-17DOI: 10.1109/TRPMS.2025.3542729
Liang Guo;Stephanie L. Thorn;Pedro Gil de Rubio Cruz;Zhao Liu;Jean-Dominique Gallezot;Qiong Liu;Eric Moulton;Richard E. Carson;Albert J. Sinusas;Chi Liu
Accurate assessment of regional flow in the lower extremities is crucial for managing peripheral arterial disease with critical limb ischemia. This study investigates dynamic 82Rb PET imaging with kinetic modeling for evaluating skeletal muscle flow in a porcine model of hindlimb ischemia. Five pigs with acute unilateral occlusion of the right common femoral artery were scanned at rest using two protocols: The first protocol involved two sequential injections to measure the image-derived input function (IDIF) in the left ventricle (LV) and leg blood flow. A three-parameter one-tissue compartment with spillover model estimated skeletal muscle flow in ischemic and nonischemic limbs. The effects of correcting delay and dispersion of LV-IDIF on model fitting were explored. For short axial field of view scanners, the feasibility of a single injection with shuttling between the heart and the leg was also assessed. Flow estimates ranged from 0.012 to 0.077 cm3/min/cm3 across animals and significantly decreased on ischemic muscles (p < 0.05). Delay and dispersion corrections yielded improved Akaike information criterion values and physiological consistency. However, accurate corrections were more difficult using the single injection and shuttling protocol. Future studies to optimize data acquisition are needed.
{"title":"Lower Extremity Flow Quantification Using Dynamic ⁸²Rb PET: A Preclinical Investigation","authors":"Liang Guo;Stephanie L. Thorn;Pedro Gil de Rubio Cruz;Zhao Liu;Jean-Dominique Gallezot;Qiong Liu;Eric Moulton;Richard E. Carson;Albert J. Sinusas;Chi Liu","doi":"10.1109/TRPMS.2025.3542729","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3542729","url":null,"abstract":"Accurate assessment of regional flow in the lower extremities is crucial for managing peripheral arterial disease with critical limb ischemia. This study investigates dynamic 82Rb PET imaging with kinetic modeling for evaluating skeletal muscle flow in a porcine model of hindlimb ischemia. Five pigs with acute unilateral occlusion of the right common femoral artery were scanned at rest using two protocols: The first protocol involved two sequential injections to measure the image-derived input function (IDIF) in the left ventricle (LV) and leg blood flow. A three-parameter one-tissue compartment with spillover model estimated skeletal muscle flow in ischemic and nonischemic limbs. The effects of correcting delay and dispersion of LV-IDIF on model fitting were explored. For short axial field of view scanners, the feasibility of a single injection with shuttling between the heart and the leg was also assessed. Flow estimates ranged from 0.012 to 0.077 cm3/min/cm3 across animals and significantly decreased on ischemic muscles (p < 0.05). Delay and dispersion corrections yielded improved Akaike information criterion values and physiological consistency. However, accurate corrections were more difficult using the single injection and shuttling protocol. Future studies to optimize data acquisition are needed.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"918-926"},"PeriodicalIF":3.5,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144998087","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-14DOI: 10.1109/TRPMS.2025.3542024
Zhixiang Zhao;Qiu Huang;Craig S. Levin
This study introduces and evaluates a new front-end electronics design for time-of-flight (TOF) 3-D position sensitive (TOF-3-DPS) detectors with a side-readout configuration. This design employs an RF amplifier and summing circuit-based timing multiplexing scheme to achieve 24:1 timing multiplexing. Additionally, complex programmable logic devices are utilized for precise energy measurement and 3-D positioning, accommodating both single and multiinteraction intercrystal scatter (ICS) events within a detector unit. Experimental results on a single $3times 3times $ 10 mm3 LYSO:Ce crystal side-coupled to three $3times $ 3 mm2 SiPMs in a $4times 6$ SiPM array demonstrated a $9.17pm 0.20$ % energy resolution, a $1.20pm 0$ .26 mm FWHM depth-of-interaction (DOI) resolution, and a $112.46pm 1.91$ ps FWHM coincidence time resolution (CTR) after DOI-related time skew correction. Further tests on a detector unit comprising a $4times 2$ array of $3times 3times $ 10 mm3 LYSO:Ce crystals, side-coupled with the same $4times 6$ SiPM array, yielded a $10.56pm 1.05$ % energy resolution and a $121.28pm 3.35$ ps FWHM DOI-calibrated CTR. The ICS event ratio for each crystal element within the detector unit was also preliminarily assessed. The front-end readout circuit consumes approximately 0.75 W per 24-SiPMs detector unit and features a compact $27times $ 95 mm2 footprint capable of reading out two units, enabling easy stacking of multiple units to form a complete TOF-3-DPS detector module.
{"title":"Front-End Electronics Design for 3-D Position Sensitive TOF-PET Detector That Achieves ~120-ps CTR and ~1.2-mm DOI Resolution","authors":"Zhixiang Zhao;Qiu Huang;Craig S. Levin","doi":"10.1109/TRPMS.2025.3542024","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3542024","url":null,"abstract":"This study introduces and evaluates a new front-end electronics design for time-of-flight (TOF) 3-D position sensitive (TOF-3-DPS) detectors with a side-readout configuration. This design employs an RF amplifier and summing circuit-based timing multiplexing scheme to achieve 24:1 timing multiplexing. Additionally, complex programmable logic devices are utilized for precise energy measurement and 3-D positioning, accommodating both single and multiinteraction intercrystal scatter (ICS) events within a detector unit. Experimental results on a single <inline-formula> <tex-math>$3times 3times $ </tex-math></inline-formula> 10 mm3 LYSO:Ce crystal side-coupled to three <inline-formula> <tex-math>$3times $ </tex-math></inline-formula> 3 mm2 SiPMs in a <inline-formula> <tex-math>$4times 6$ </tex-math></inline-formula> SiPM array demonstrated a <inline-formula> <tex-math>$9.17pm 0.20$ </tex-math></inline-formula>% energy resolution, a <inline-formula> <tex-math>$1.20pm 0$ </tex-math></inline-formula>.26 mm FWHM depth-of-interaction (DOI) resolution, and a <inline-formula> <tex-math>$112.46pm 1.91$ </tex-math></inline-formula> ps FWHM coincidence time resolution (CTR) after DOI-related time skew correction. Further tests on a detector unit comprising a <inline-formula> <tex-math>$4times 2$ </tex-math></inline-formula> array of <inline-formula> <tex-math>$3times 3times $ </tex-math></inline-formula> 10 mm3 LYSO:Ce crystals, side-coupled with the same <inline-formula> <tex-math>$4times 6$ </tex-math></inline-formula> SiPM array, yielded a <inline-formula> <tex-math>$10.56pm 1.05$ </tex-math></inline-formula>% energy resolution and a <inline-formula> <tex-math>$121.28pm 3.35$ </tex-math></inline-formula> ps FWHM DOI-calibrated CTR. The ICS event ratio for each crystal element within the detector unit was also preliminarily assessed. The front-end readout circuit consumes approximately 0.75 W per 24-SiPMs detector unit and features a compact <inline-formula> <tex-math>$27times $ </tex-math></inline-formula> 95 mm2 footprint capable of reading out two units, enabling easy stacking of multiple units to form a complete TOF-3-DPS detector module.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 6","pages":"736-746"},"PeriodicalIF":4.6,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10887305","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144597922","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Computed tomography (CT) is widely used to generate cross-sectional views of the internal anatomy of a subject. Conventional CT imaging with single energy is, however, incapable of providing material composition information for various clinical applications because different materials may lead to the same CT numbers. Dual-energy CT (DECT) with physical means of simultaneously generating and measuring photon signals of two different spectra is designed to break this degeneracy. While valuable, this approach adds an extra layer of complexity on top of the widely used single-energy CT (SECT) and increases system costs, hindering the use of DECT scanners in less developed regions. Leveraging the ability of deep learning in nonlinear mapping and prior knowledge extraction from routine clinical data, here we develop a data-driven, lightweight strategy of obtaining DECT images from SECT images using a physically constrained attention mechanism. The proposed strategy is evaluated comprehensively by using high-fidelity simulation datasets and clinical contrast-enhanced DECT datasets. In terms of both prediction accuracy and inference speed, our method exhibits notable advantages over a variety of existing approaches. This technique holds the potential to provide a fast and cost-effective solution for contrast-enhanced spectral CT, catering to a broad range of CT applications.
{"title":"Data-Driven Contrast-Enhanced Dual-Energy CT Imaging via Physically Constrained Attention","authors":"Wenwen Zhang;Tianling Lyu;Yongqing Li;Yang Chen;Baohua Sun;Wei Zhao","doi":"10.1109/TRPMS.2025.3541742","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3541742","url":null,"abstract":"Computed tomography (CT) is widely used to generate cross-sectional views of the internal anatomy of a subject. Conventional CT imaging with single energy is, however, incapable of providing material composition information for various clinical applications because different materials may lead to the same CT numbers. Dual-energy CT (DECT) with physical means of simultaneously generating and measuring photon signals of two different spectra is designed to break this degeneracy. While valuable, this approach adds an extra layer of complexity on top of the widely used single-energy CT (SECT) and increases system costs, hindering the use of DECT scanners in less developed regions. Leveraging the ability of deep learning in nonlinear mapping and prior knowledge extraction from routine clinical data, here we develop a data-driven, lightweight strategy of obtaining DECT images from SECT images using a physically constrained attention mechanism. The proposed strategy is evaluated comprehensively by using high-fidelity simulation datasets and clinical contrast-enhanced DECT datasets. In terms of both prediction accuracy and inference speed, our method exhibits notable advantages over a variety of existing approaches. This technique holds the potential to provide a fast and cost-effective solution for contrast-enhanced spectral CT, catering to a broad range of CT applications.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"905-917"},"PeriodicalIF":3.5,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10884935","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144998090","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-12DOI: 10.1109/TRPMS.2025.3541169
Jiping Wang;Hao Fan;Zhongyi Wu;Qiang Du;Ming Li;Jian Zheng;Greta S. P. Mok;Benjamin M. W. Tsui
Low-dose computed tomography (LDCT) denoising methods based on supervised learning with labeled simulation data have made significant progress. However, these methods usually struggle to directly process unlabeled LDCT images due to inherent biases. While unsupervised methods have been explored to utilize unlabeled LDCT images, they typically involve complex network structures with limited denoising performance. To address these issues, we propose a self-adaptive weight embedded lightweight semi-supervised network (SWELNet) for unlabeled LDCT image denoising, which integrates supervised and unsupervised learning in a lightweight architecture. Unlike other semi-supervised algorithms that only consider the correlations between labeled simulation data and unlabeled data, the proposed SWELNet not only takes into account correlations but also the differences between data. There are three modules in the proposed network, respectively, for feature extraction, refinement and self-adaptive weight. Specially, the multiscale convolution feature extraction module (MCFEM) and recursive module (RECM) extract and refine common representations from labeled simulation and unlabeled data with the well-designed. After that, the softmax feature fusion module (SFFM) with self-adaptive weighted learning for forming different feature spaces for two types of data. Extensive experiments using one simulation and two unlabeled datasets demonstrate that the proposed SWELNet outperforms several state-of-the-art baseline network methods in terms of robustness and generalization, as well as inference efficiency. The code is available at https://github.com/nightastars/SWELNet-main.git.
{"title":"Self-Adaptive Weight Embedded Lightweight Network Using Semi-Supervised Learning for Low-Dose CT Image Denoising","authors":"Jiping Wang;Hao Fan;Zhongyi Wu;Qiang Du;Ming Li;Jian Zheng;Greta S. P. Mok;Benjamin M. W. Tsui","doi":"10.1109/TRPMS.2025.3541169","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3541169","url":null,"abstract":"Low-dose computed tomography (LDCT) denoising methods based on supervised learning with labeled simulation data have made significant progress. However, these methods usually struggle to directly process unlabeled LDCT images due to inherent biases. While unsupervised methods have been explored to utilize unlabeled LDCT images, they typically involve complex network structures with limited denoising performance. To address these issues, we propose a self-adaptive weight embedded lightweight semi-supervised network (SWELNet) for unlabeled LDCT image denoising, which integrates supervised and unsupervised learning in a lightweight architecture. Unlike other semi-supervised algorithms that only consider the correlations between labeled simulation data and unlabeled data, the proposed SWELNet not only takes into account correlations but also the differences between data. There are three modules in the proposed network, respectively, for feature extraction, refinement and self-adaptive weight. Specially, the multiscale convolution feature extraction module (MCFEM) and recursive module (RECM) extract and refine common representations from labeled simulation and unlabeled data with the well-designed. After that, the softmax feature fusion module (SFFM) with self-adaptive weighted learning for forming different feature spaces for two types of data. Extensive experiments using one simulation and two unlabeled datasets demonstrate that the proposed SWELNet outperforms several state-of-the-art baseline network methods in terms of robustness and generalization, as well as inference efficiency. The code is available at <uri>https://github.com/nightastars/SWELNet-main.git</uri>.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"890-904"},"PeriodicalIF":3.5,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144998009","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-12DOI: 10.1109/TRPMS.2025.3539739
Jiaqi Cui;Yuanyuan Xu;Hanci Zheng;Xi Wu;Jiliu Zhou;Yuanjun Liu;Yan Wang
Survival prediction is crucial for cancer patients as it offers prognostic information for treatment planning. Recently, deep learning-based multimodal survival prediction models have demonstrated promising performance. However, current models face challenges in effectively utilizing heterogeneous multimodal data (e.g., positron emission tomography (PET)/computed tomography (CT) images and clinical tabular) and extracting essential information from tumor regions, resulting in suboptimal survival prediction accuracy. To tackle these limitations, in this article, we propose a novel hybrid multimodal transformer model (HMT), namely HMT, for survival prediction from PET/CT images and clinical tabular in Head and Neck (H&N) cancer. Specifically, we develop hybrid attention modules to capture intramodal information and intermodal correlations from multimodal PET/CT images. Moreover, we design hierarchical Tabular Affine transformation modules (TATMs) to integrate supplementary insights from the heterogenous tabular with images via affine transformations. The TATM dynamically emphasizes features contributing to the survival prediction while suppressing irrelevant ones during integration. To achieve finer feature fusion, TATMs are hierarchically embedded into the network, allowing for consistent interaction between tabular and multimodal image features across multiple scales. To mitigate interferences caused by irrelevant information, we introduce tumor segmentation as an auxiliary task to capture features related to tumor regions, thus enhancing prediction accuracy. Experiments demonstrate our superior performance. The code is available at https://github.com/gluucose/HMT.
{"title":"HMT: A Hybrid Multimodal Transformer With Multitask Learning for Survival Prediction in Head and Neck Cancer","authors":"Jiaqi Cui;Yuanyuan Xu;Hanci Zheng;Xi Wu;Jiliu Zhou;Yuanjun Liu;Yan Wang","doi":"10.1109/TRPMS.2025.3539739","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3539739","url":null,"abstract":"Survival prediction is crucial for cancer patients as it offers prognostic information for treatment planning. Recently, deep learning-based multimodal survival prediction models have demonstrated promising performance. However, current models face challenges in effectively utilizing heterogeneous multimodal data (e.g., positron emission tomography (PET)/computed tomography (CT) images and clinical tabular) and extracting essential information from tumor regions, resulting in suboptimal survival prediction accuracy. To tackle these limitations, in this article, we propose a novel hybrid multimodal transformer model (HMT), namely HMT, for survival prediction from PET/CT images and clinical tabular in Head and Neck (H&N) cancer. Specifically, we develop hybrid attention modules to capture intramodal information and intermodal correlations from multimodal PET/CT images. Moreover, we design hierarchical Tabular Affine transformation modules (TATMs) to integrate supplementary insights from the heterogenous tabular with images via affine transformations. The TATM dynamically emphasizes features contributing to the survival prediction while suppressing irrelevant ones during integration. To achieve finer feature fusion, TATMs are hierarchically embedded into the network, allowing for consistent interaction between tabular and multimodal image features across multiple scales. To mitigate interferences caused by irrelevant information, we introduce tumor segmentation as an auxiliary task to capture features related to tumor regions, thus enhancing prediction accuracy. Experiments demonstrate our superior performance. The code is available at <uri>https://github.com/gluucose/HMT</uri>.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 7","pages":"879-889"},"PeriodicalIF":3.5,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144997932","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：