Pub Date : 2025-04-14DOI: 10.1109/TRPMS.2025.3560558
Elena Maria Zannoni;Can Yang;Ling Cai;Matthew D. Wilson;Chin-Tu Chen;Ling-Jian Meng
There is a rising interest in single-photon emission computed tomography (SPECT) imaging systems with improved energy resolution to facilitate multifunctional molecular imaging applications, such as alpha-emitter radiopharmaceutical therapy ($alpha $ -RPT). In this article, we report the design and evaluation of the Alpha-SPECT-Mini system that offers an ultrahigh energy resolution and high sensitivity for small animal studies. The Alpha-SPECT-Mini system is constructed based on small-pixel CdTe detectors that offers sub-1-keV full-width-half-maximum (FWHM) energy resolution for single pixel events and an average ~2.5-keV energy resolution at 122 keV and ~3.5 keV at 218 keV over 153 600 pixels in the system. This allows to easily identify X- and gamma-ray contributions in densely populated spectra, such as from the Ac-225 decay chain. The system uses a 96-loft-hole collimator and six stationary detection panels in a full ring geometry. Finally, the system performance is demonstrated using Tc-99m- and Ac-225-filled resolution and image quality (IQ) phantoms. We have experimentally demonstrated that the Alpha-SPECT-Mini is a high-performance imaging system capable of imaging alpha-emitters in preclinical applications.
{"title":"The Alpha-SPECT-Mini: A Small-Animal SPECT System Based on Hyperspectral Compound-Eye Gamma Cameras","authors":"Elena Maria Zannoni;Can Yang;Ling Cai;Matthew D. Wilson;Chin-Tu Chen;Ling-Jian Meng","doi":"10.1109/TRPMS.2025.3560558","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3560558","url":null,"abstract":"There is a rising interest in single-photon emission computed tomography (SPECT) imaging systems with improved energy resolution to facilitate multifunctional molecular imaging applications, such as alpha-emitter radiopharmaceutical therapy (<inline-formula> <tex-math>$alpha $ </tex-math></inline-formula>-RPT). In this article, we report the design and evaluation of the Alpha-SPECT-Mini system that offers an ultrahigh energy resolution and high sensitivity for small animal studies. The Alpha-SPECT-Mini system is constructed based on small-pixel CdTe detectors that offers sub-1-keV full-width-half-maximum (FWHM) energy resolution for single pixel events and an average ~2.5-keV energy resolution at 122 keV and ~3.5 keV at 218 keV over 153 600 pixels in the system. This allows to easily identify X- and gamma-ray contributions in densely populated spectra, such as from the Ac-225 decay chain. The system uses a 96-loft-hole collimator and six stationary detection panels in a full ring geometry. Finally, the system performance is demonstrated using Tc-99m- and Ac-225-filled resolution and image quality (IQ) phantoms. We have experimentally demonstrated that the Alpha-SPECT-Mini is a high-performance imaging system capable of imaging alpha-emitters in preclinical applications.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 8","pages":"1107-1117"},"PeriodicalIF":3.5,"publicationDate":"2025-04-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435706","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 : 2025-04-09DOI: 10.1109/TRPMS.2025.3559095
Zerui Yu;Zhenlei Lyu;Peng Fan;Jing Wu;Yaqiang Liu;Tianyu Ma
In nuclear medicine imaging systems, intrinsic spatial resolution of the detector is one of the most important performance metrics. In this work, we aim to develop a high-resolution single photon emission computed tomography (SPECT) detector using pixelated Ce-doped gadolinium aluminum gallium garnet (GAGG:Ce) scintillators and silicon photomultiplier (SiPM) arrays. Special attention is paid to improving the resolving capability of edge crystals. We propose to place optical barrier (OB) slits onto the light guide that enhances the difference in light distribution for edge crystals. We experimentally optimize OB designs for two scintillator arrays, named as Array-ESR and Array-BaSO4, which uses enhanced specular reflector (ESR) film and barium sulfate (BaSO4) as the reflectors, respectively. Both arrays have $31times 31~0$ .8 mm $times 0$ .8 mm $times $ 6 mm GAGG:Ce crystals. We introduce the flood map quality (FMQ) parameter to assess the separation of responses of neighboring crystals. The results demonstrate that for Array-ESR, an optimal light guide with two 7° OB slits and two 11° OB slits resolves 92.40% crystals with an energy resolution of 13.19% $pm ~0.68$ %. The FMQ is $1.52~pm ~0.38$ . For Array-BaSO4, the optimal design is a light guide with four 7° OB slits. 98.75% crystals are resolvable with an energy resolution of 15.33% $pm ~0.96$ % and FMQ parameter of $1.81~pm ~0.45$ . Overall, Array-BaSO4 is more suitable for building SPECT detector for its good crystal resolving performance and fabrication convenience. This study proposes a practical submillimeter pixelated SPECT detector design with no detection dead space and compact electronics. It is promising for being used to build large-scale detectors for high resolution SPECT systems.
在核医学成像系统中，探测器的固有空间分辨率是最重要的性能指标之一。在这项工作中，我们的目标是使用像素化掺Ce钆铝镓石榴石（GAGG:Ce）闪烁体和硅光电倍增管（SiPM）阵列开发高分辨率单光子发射计算机断层扫描（SPECT）探测器。特别注意提高边缘晶体的分辨能力。我们建议在光导上放置光学屏障（OB）狭缝，以增强边缘晶体的光分布差异。本文通过实验优化了两种闪烁体阵列（Array-ESR和Array-BaSO4）的OB设计，这两种闪烁体阵列分别使用增强镜面反射器（ESR）薄膜和硫酸钡（BaSO4）作为反射器。两个数组都有$31乘以31~0$。8 mm $乘以0$。8毫米$乘以6毫米$ GAGG:Ce晶体。我们引入洪水图质量（FMQ）参数来评估相邻晶体的分离响应。结果表明，对于Array-ESR，具有两个7°OB狭缝和两个11°OB狭缝的最优光导可以分辨92.40%的晶体，能量分辨率为13.19% ~0.68美元%。FMQ为1.52~ 0.38美元。对于Array-BaSO4，最优设计是具有四个7°OB狭缝的光导。98.75%的晶体可分辨，能量分辨率为15.33% ~0.96美元%，FMQ参数为1.81~ 0.45美元。综上所述，阵列- baso4具有良好的晶体分辨性能和制作方便，更适合用于构建SPECT探测器。本研究提出一种实用的亚毫米像素化SPECT探测器设计，无检测死区，电子元件紧凑。它有望用于构建高分辨率SPECT系统的大规模探测器。
{"title":"Submillimeter Pixelated SPECT Detector Using GAGG:Ce and Light Guide With Optical Barrier Slits","authors":"Zerui Yu;Zhenlei Lyu;Peng Fan;Jing Wu;Yaqiang Liu;Tianyu Ma","doi":"10.1109/TRPMS.2025.3559095","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3559095","url":null,"abstract":"In nuclear medicine imaging systems, intrinsic spatial resolution of the detector is one of the most important performance metrics. In this work, we aim to develop a high-resolution single photon emission computed tomography (SPECT) detector using pixelated Ce-doped gadolinium aluminum gallium garnet (GAGG:Ce) scintillators and silicon photomultiplier (SiPM) arrays. Special attention is paid to improving the resolving capability of edge crystals. We propose to place optical barrier (OB) slits onto the light guide that enhances the difference in light distribution for edge crystals. We experimentally optimize OB designs for two scintillator arrays, named as Array-ESR and Array-BaSO4, which uses enhanced specular reflector (ESR) film and barium sulfate (BaSO4) as the reflectors, respectively. Both arrays have <inline-formula> <tex-math>$31times 31~0$ </tex-math></inline-formula>.8 mm <inline-formula> <tex-math>$times 0$ </tex-math></inline-formula>.8 mm <inline-formula> <tex-math>$times $ </tex-math></inline-formula> 6 mm GAGG:Ce crystals. We introduce the flood map quality (FMQ) parameter to assess the separation of responses of neighboring crystals. The results demonstrate that for Array-ESR, an optimal light guide with two 7° OB slits and two 11° OB slits resolves 92.40% crystals with an energy resolution of 13.19% <inline-formula> <tex-math>$pm ~0.68$ </tex-math></inline-formula>%. The FMQ is <inline-formula> <tex-math>$1.52~pm ~0.38$ </tex-math></inline-formula>. For Array-BaSO4, the optimal design is a light guide with four 7° OB slits. 98.75% crystals are resolvable with an energy resolution of 15.33% <inline-formula> <tex-math>$pm ~0.96$ </tex-math></inline-formula>% and FMQ parameter of <inline-formula> <tex-math>$1.81~pm ~0.45$ </tex-math></inline-formula>. Overall, Array-BaSO4 is more suitable for building SPECT detector for its good crystal resolving performance and fabrication convenience. This study proposes a practical submillimeter pixelated SPECT detector design with no detection dead space and compact electronics. It is promising for being used to build large-scale detectors for high resolution SPECT systems.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 8","pages":"1015-1024"},"PeriodicalIF":3.5,"publicationDate":"2025-04-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435702","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}
This study aims to develop a compact, low-cost, and high-performance benchtop small-animal PET/MRI scanner that achieves functional and anatomical image fusion. The system is designed to address challenges in cost reduction, spatial resolution, sensitivity, image quality (IQ), and quantitative accuracy. The PET/MRI system was developed with a parallel configuration, integrating a custom-designed PET scanner and a 0.5-T permanent magnet MRI system. Quantitative assessments included spatial resolution, sensitivity, IQ, and quantitative accuracy, as well as signal-to-noise ratio (SNR), geometric distortion (GD), and image uniformity (IU) for MRI. The spatial resolution at the axial center is 1.31 (axial), 1.26 (radial), and 1.22 mm (tangential), with a center sensitivity of 8.05% under a wide energy window. Image quality (IQ) tests using an IQ phantom demonstrated a uniformity of 10.08% standard deviation, recovery coefficients (RC) ranging from 0.23 to 0.96, and spill-over ratios (SOR) of 0.08 and 0.18 in air and water regions, respectively. The MRI system achieved an SNR of 14.16 in phantom tests, a GD of less than 1%, and IU of 90.13%. Fusion imaging of PET and MRI demonstrated high registration accuracy in both phantom and mouse studies, with complementary functional and anatomical information. The proposed PET/MRI system achieves high spatial resolution, sensitivity, IQ, and quantitative accuracy while maintaining a simple, low-cost design. The parallel configuration facilitates precise PET/MRI image fusion and allows for efficient multianimal imaging. The results highlight the potential of this system for preclinical research and its feasibility for future in-vehicle imaging applications. Further optimization of the MRI system and data transmission methods will enhance its performance in high-activity studies and broaden its application scope, with potential applications in preclinical research and in-vehicle imaging.
{"title":"Development and Performance Evaluation of a Benchtop Small-Animal PET/MRI Scanner","authors":"Xin Yu;Zhijun Zhao;Han Liu;Da Liang;Wenjing Zhu;Ying Lin;Jiayang Zeng;Chenxuan Liu;Jianfeng Xu;Siwei Xie;Weimin Wang;Qiyu Peng","doi":"10.1109/TRPMS.2025.3557789","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3557789","url":null,"abstract":"This study aims to develop a compact, low-cost, and high-performance benchtop small-animal PET/MRI scanner that achieves functional and anatomical image fusion. The system is designed to address challenges in cost reduction, spatial resolution, sensitivity, image quality (IQ), and quantitative accuracy. The PET/MRI system was developed with a parallel configuration, integrating a custom-designed PET scanner and a 0.5-T permanent magnet MRI system. Quantitative assessments included spatial resolution, sensitivity, IQ, and quantitative accuracy, as well as signal-to-noise ratio (SNR), geometric distortion (GD), and image uniformity (IU) for MRI. The spatial resolution at the axial center is 1.31 (axial), 1.26 (radial), and 1.22 mm (tangential), with a center sensitivity of 8.05% under a wide energy window. Image quality (IQ) tests using an IQ phantom demonstrated a uniformity of 10.08% standard deviation, recovery coefficients (RC) ranging from 0.23 to 0.96, and spill-over ratios (SOR) of 0.08 and 0.18 in air and water regions, respectively. The MRI system achieved an SNR of 14.16 in phantom tests, a GD of less than 1%, and IU of 90.13%. Fusion imaging of PET and MRI demonstrated high registration accuracy in both phantom and mouse studies, with complementary functional and anatomical information. The proposed PET/MRI system achieves high spatial resolution, sensitivity, IQ, and quantitative accuracy while maintaining a simple, low-cost design. The parallel configuration facilitates precise PET/MRI image fusion and allows for efficient multianimal imaging. The results highlight the potential of this system for preclinical research and its feasibility for future in-vehicle imaging applications. Further optimization of the MRI system and data transmission methods will enhance its performance in high-activity studies and broaden its application scope, with potential applications in preclinical research and in-vehicle imaging.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 8","pages":"1118-1126"},"PeriodicalIF":3.5,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435709","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}
Proton therapy is one of the most advanced radiotherapy techniques. Despite its advantages in dose delivery, it has not yet achieved significant clinical benefits for patients due to uncertainties in proton range. Accurate, real-time monitoring of proton dose and range is crucial for ensuring the precision of proton therapy. In prior work, a dual-head prompt gamma imaging system was proposed and evaluated through Monte Carlo simulations, demonstrating high spatial resolution and sufficient detection efficiency for proton pencil beam imaging at clinical doses. This study focuses on the assembly, calibration, and testing of one of the detectors in this system. Spatial resolution and detection efficiency were evaluated using a 22Na point source, while range shift detection and accuracy were assessed with 60 and 100 MeV proton beams under low proton count conditions. The single-head system achieved a detection efficiency of 0.22% and a full-width at half-maximum (FWHM) spatial resolution of 2.8 mm at the center of the field of view (FOV). The system was able to detect a 1 mm range shift by identifying the most distal edge position (MDEP) of the prompt gamma profile. The detector demonstrated a range accuracy of less than 1 mm at typical count levels for a single spot in proton pencil beam scanning. The results suggest that this system performs well in terms of both detection efficiency and spatial resolution, and the system could achieve real-time range verification with high accuracy.
{"title":"Proton Range Verification Realized via a Multislit Prompt Gamma Imaging System","authors":"Hongyang Zhang;Bo Zhao;Peng Fan;Shi Wang;Wenzhuo Lu;Yancheng Yu;Zhaoxia Wu;Tianyu Ma;Hui Liu;Yaqiang Liu","doi":"10.1109/TRPMS.2025.3553133","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3553133","url":null,"abstract":"Proton therapy is one of the most advanced radiotherapy techniques. Despite its advantages in dose delivery, it has not yet achieved significant clinical benefits for patients due to uncertainties in proton range. Accurate, real-time monitoring of proton dose and range is crucial for ensuring the precision of proton therapy. In prior work, a dual-head prompt gamma imaging system was proposed and evaluated through Monte Carlo simulations, demonstrating high spatial resolution and sufficient detection efficiency for proton pencil beam imaging at clinical doses. This study focuses on the assembly, calibration, and testing of one of the detectors in this system. Spatial resolution and detection efficiency were evaluated using a 22Na point source, while range shift detection and accuracy were assessed with 60 and 100 MeV proton beams under low proton count conditions. The single-head system achieved a detection efficiency of 0.22% and a full-width at half-maximum (FWHM) spatial resolution of 2.8 mm at the center of the field of view (FOV). The system was able to detect a 1 mm range shift by identifying the most distal edge position (MDEP) of the prompt gamma profile. The detector demonstrated a range accuracy of less than 1 mm at typical count levels for a single spot in proton pencil beam scanning. The results suggest that this system performs well in terms of both detection efficiency and spatial resolution, and the system could achieve real-time range verification with high accuracy.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 8","pages":"1127-1134"},"PeriodicalIF":3.5,"publicationDate":"2025-04-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435694","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 : 2025-04-02DOI: 10.1109/TRPMS.2025.3552178
{"title":">Member Get-a-Member (MGM) Program","authors":"","doi":"10.1109/TRPMS.2025.3552178","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3552178","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 4","pages":"529-529"},"PeriodicalIF":4.6,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10947670","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143761356","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 : 2025-04-02DOI: 10.1109/TRPMS.2025.3552176
{"title":"IEEE DataPort","authors":"","doi":"10.1109/TRPMS.2025.3552176","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3552176","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 4","pages":"528-528"},"PeriodicalIF":4.6,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10947674","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143761466","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 : 2025-04-02DOI: 10.1109/TRPMS.2025.3552150
{"title":"IEEE Transactions on Radiation and Plasma Medical Sciences Publication Information","authors":"","doi":"10.1109/TRPMS.2025.3552150","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3552150","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 4","pages":"C2-C2"},"PeriodicalIF":4.6,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10947672","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143761439","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 : 2025-04-02DOI: 10.1109/TRPMS.2025.3552148
{"title":"IEEE Transactions on Radiation and Plasma Medical Sciences Information for Authors","authors":"","doi":"10.1109/TRPMS.2025.3552148","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3552148","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 4","pages":"C3-C3"},"PeriodicalIF":4.6,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10947675","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143761399","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}
Recent advances in deep-learning-based methods have shown great potential in improving low-dose CT image quality. Meanwhile, these methods are constructed based on a large, centralized, and diverse CT dataset from multiple institutions that is difficult to collect and share due to the high-cost acquisition and data privacy regulations. Previously developed federated learning (FL)-based methods enable collaborative and decentralized training without exchanging local data to preserve data privacy. In this work, we focus on analyzing the robustness of FL-based methods against dataset shifts (i.e., the datasets among multiple institutions are from different scanners, different protocols, or different sampling conditions). The results show that the FL-based CT reconstruction methods are sensitive to domain shifts, which can be attributed to the data heterogeneity among multiple institutions. Based on these findings, we propose a unified CT reconstruction method that leverages high-quality metadata (e.g., low-dose images and their corresponding normal-dose counterparts) stored on the cloud server to address the challenge of multi-institutional domain shifts. For simplicity, we refer to the proposed method as FM-iRadonMAP, representing federated metadata learning (FMDL) with a personalized condition-modulated iRadonMAP (CM-iRadonMAP). Specifically, the FM-iRadonMAP consists of two modules, i.e., CM-iRadonMAP and FMDL. CM-iRadonMAP introduces the knowledge of client-specific sampling conditions, i.e., imaging geometries and scan protocols, into iRadonMAP reconstruction network at each client to modulate the reconstruction effectively. FMDL trains a supervised meta model using high-quality metadata in an additional round and then adaptively unifies the network parameters of the meta model with those of the local models from all clients for broadcasting, addressing the issue of data heterogeneity. A large-scale multi-institutional CT dataset is used to validate and evaluate the reconstruction performance of the FM-iRadonMAP. The experimental results demonstrate the feasibility of the FM-iRadonMAP for multi-institutional CT reconstruction with severe data heterogeneity.
{"title":"Toward Unified CT Reconstruction: Federated Metadata Learning With Personalized Condition-Modulated iRadonMAP","authors":"Hao Wang;Mingqiang Li;Shixuan Chen;Mingqiang Meng;Ji He;Jianhua Ma;Dong Zeng","doi":"10.1109/TRPMS.2025.3574209","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3574209","url":null,"abstract":"Recent advances in deep-learning-based methods have shown great potential in improving low-dose CT image quality. Meanwhile, these methods are constructed based on a large, centralized, and diverse CT dataset from multiple institutions that is difficult to collect and share due to the high-cost acquisition and data privacy regulations. Previously developed federated learning (FL)-based methods enable collaborative and decentralized training without exchanging local data to preserve data privacy. In this work, we focus on analyzing the robustness of FL-based methods against dataset shifts (i.e., the datasets among multiple institutions are from different scanners, different protocols, or different sampling conditions). The results show that the FL-based CT reconstruction methods are sensitive to domain shifts, which can be attributed to the data heterogeneity among multiple institutions. Based on these findings, we propose a unified CT reconstruction method that leverages high-quality metadata (e.g., low-dose images and their corresponding normal-dose counterparts) stored on the cloud server to address the challenge of multi-institutional domain shifts. For simplicity, we refer to the proposed method as FM-iRadonMAP, representing federated metadata learning (FMDL) with a personalized condition-modulated iRadonMAP (CM-iRadonMAP). Specifically, the FM-iRadonMAP consists of two modules, i.e., CM-iRadonMAP and FMDL. CM-iRadonMAP introduces the knowledge of client-specific sampling conditions, i.e., imaging geometries and scan protocols, into iRadonMAP reconstruction network at each client to modulate the reconstruction effectively. FMDL trains a supervised meta model using high-quality metadata in an additional round and then adaptively unifies the network parameters of the meta model with those of the local models from all clients for broadcasting, addressing the issue of data heterogeneity. A large-scale multi-institutional CT dataset is used to validate and evaluate the reconstruction performance of the FM-iRadonMAP. The experimental results demonstrate the feasibility of the FM-iRadonMAP for multi-institutional CT reconstruction with severe data heterogeneity.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 2","pages":"169-180"},"PeriodicalIF":3.5,"publicationDate":"2025-03-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11017337","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146116838","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}
Four-dimensionalcone-beam computed tomography (4-D CBCT) provides respiration-resolved images and facilitates image-guided radiation therapy. However, the ability to reveal respiratory motion comes at the cost of image artifacts. As raw projection data are sorted into multiple respiratory phases, the reconstructed 4-D CBCT images are covered by severe streak artifacts. Although several deep learning-based methods have been proposed to address this issue, most algorithms formulate it as a 2-D image enhancement task, neglecting the dynamic nature of 4-D CBCT. In this article, we first identify the origin and appearance of streak artifacts in 4-D CBCT images. We find that streak artifacts exhibit a unique “rotational motion” along with the patient’s respiration, distinguishable from diaphragm-driven respiratory motion in 4-D space. Therefore, we introduce RSTAR4D-Net, a 4-D model that performs rotational streak artifact reduction by exploring the dynamic prior of 4-D CBCT images. Specifically, we overcome the computational and training difficulties of a 4-D neural network. The specially designed model decomposes the 4-D convolutions into multiple lower-dimensional operations and thus efficiently processes a whole 4-D image. Additionally, a Tetris training strategy is proposed to effectively train the model using limited 4-D data. Extensive experiments substantiate the superior performance of RSTAR4D-Net compared to existing methods.
{"title":"RSTAR4D: Rotational Streak Artifact Reduction in 4-D CBCT Using Separable 4-D Convolutions","authors":"Ziheng Deng;Hua Chen;Yongzheng Zhou;Haibo Hu;Zhiyong Xu;Tianling Lyu;Yan Xi;Yang Chen;Jiayuan Sun;Jun Zhao","doi":"10.1109/TRPMS.2025.3553866","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3553866","url":null,"abstract":"Four-dimensionalcone-beam computed tomography (4-D CBCT) provides respiration-resolved images and facilitates image-guided radiation therapy. However, the ability to reveal respiratory motion comes at the cost of image artifacts. As raw projection data are sorted into multiple respiratory phases, the reconstructed 4-D CBCT images are covered by severe streak artifacts. Although several deep learning-based methods have been proposed to address this issue, most algorithms formulate it as a 2-D image enhancement task, neglecting the dynamic nature of 4-D CBCT. In this article, we first identify the origin and appearance of streak artifacts in 4-D CBCT images. We find that streak artifacts exhibit a unique “rotational motion” along with the patient’s respiration, distinguishable from diaphragm-driven respiratory motion in 4-D space. Therefore, we introduce RSTAR4D-Net, a 4-D model that performs rotational streak artifact reduction by exploring the dynamic prior of 4-D CBCT images. Specifically, we overcome the computational and training difficulties of a 4-D neural network. The specially designed model decomposes the 4-D convolutions into multiple lower-dimensional operations and thus efficiently processes a whole 4-D image. Additionally, a Tetris training strategy is proposed to effectively train the model using limited 4-D data. Extensive experiments substantiate the superior performance of RSTAR4D-Net compared to existing methods.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"9 8","pages":"1094-1106"},"PeriodicalIF":3.5,"publicationDate":"2025-03-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145435696","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：