IEEE Transactions on Radiation and Plasma Medical Sciences最新文献

{"title":"Feature Importance Guided Network With Missing-Value Learning for Pneumonia Pathogen Classification and Biomarker Discovery","authors":"Di Xu;Ru Wen;Yaqi Wang;Can Han;Haochao Ying;Dahong Qian;Chen Liu;Xudong Jiang;Jun Wang","doi":"10.1109/TRPMS.2025.3614231","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3614231","url":null,"abstract":"Discovering clinical biomarkers through multimodal modeling is crucial for precise pneumonia diagnosis and treatment. However, existing methods often fall short in either their representational capacity or interpretability. Moreover, these methods are highly sensitive to some missing values, leading to a significant performance degradation. In this article, we propose a multimodal feature importance guided network (MFIG-Net) for pneumonia classification and biomarker discovery. MFIG-Net features two novel modules that operate synergistically: a feature reconstruction module (FRM) for handling missing values and a feature selection module (FSM) for identifying biomarkers. The FSM learns coefficients indicative of variable importance, facilitating biomarker discovery and providing real-time feedback to guide the FRM. The dynamic interaction between them enables the FRM to concentrate on learning the most important features by masking and reconstructing less important variables, thereby significantly boosting model performance and robustness. Rigorous validation using over 4200 subject samples from multiple hospitals demonstrates that our model achieves an AUC score of 95.40%±0.41%, significantly outperforming existing methods. Even at 50% data missing rate, it still achieves an AUC score of 93.94%±0.35%, while most other methods fall below 70%. Our experiments further demonstrate that MFIG-Net can also be extended to other domains, such as multiomics modeling, thereby highlighting its strong broad applicability. The code is available at: <uri>https://github.com/xingzhidaoren/MFIG-Net</uri>.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"712-726"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808123","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}

{"title":"Comprehensive Evaluation of Quantitative Measurements From Automated Deep Segmentations of PSMA PET/CT Images","authors":"Obed Korshie Dzikunu;Amirhossein Toosi;Shadab Ahamed;Sara Harsini;François Bénard;Xiaoxiao Li;Arman Rahmim","doi":"10.1109/TRPMS.2025.3625611","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3625611","url":null,"abstract":"This study performs a comprehensive evaluation of quantitative measurements as extracted from automated deep-learning-based segmentation methods, beyond traditional dice similarity coefficient (DSC) assessments, focusing on six quantitative metrics, namely SUVmax, SUVmean, total lesion activity (TLA), tumor volume (TMTV), lesion count, and lesion spread. We analyzed 380 prostate-specific membrane antigen (PSMA) targeted [18F]DCFPyL positron emission tomography (PET)/CT scans of patients with biochemical recurrence of prostate cancer, training deep neural networks, U-Net, Attention U-Net and SegResNet with four loss functions: 1) dice loss (DL); 2) dice cross entropy (DCE); 3) dice focal loss (DFL); and 4) our proposed L1-weighted DFL (L1DFL). Evaluations indicated that Attention U-Net paired with L1DFL achieved the strongest correlation with the ground truth (GT) (Lin’s concordance correlation <inline-formula> <tex-math>${=}~0.90$ </tex-math></inline-formula>–0.99 for SUVmax and TLA), whereas models employing the DL and the other two compound losses, particularly with SegResNet, underperformed. Equivalence testing (TOST, <inline-formula> <tex-math>$alpha ~{=}~0.05$ </tex-math></inline-formula>, <inline-formula> <tex-math>$Delta ~{=}~20%$ </tex-math></inline-formula>) confirmed high performance for SUV metrics, lesion count and TLA, with L1DFL yielding the best performance. By contrast, tumor volume and lesion spread exhibited greater variability. Bland–Altman, coverage probability, and total deviation index analyzes further highlighted that our proposed L1DFL minimizes variability in quantification of the GT clinical measures. However, differences in DSC across the various loss functions did not reach statistical significance based on p-value analysis. The code is publicly available at: <uri>https://github.com/ObedDzik/pca_segment.git</uri>.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"727-742"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808163","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}

{"title":"IEEE Transactions on Radiation and Plasma Medical Sciences Information for Authors","authors":"","doi":"10.1109/TRPMS.2026.3684589","DOIUrl":"https://doi.org/10.1109/TRPMS.2026.3684589","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"C3-C3"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11506014","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808148","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}

{"title":"IEEE DataPort","authors":"","doi":"10.1109/TRPMS.2026.3684595","DOIUrl":"https://doi.org/10.1109/TRPMS.2026.3684595","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"836-836"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11506010","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808102","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}

{"title":"Resource-Efficient and Truly FPGA-Only Multiple Voltage Threshold Sampling Strategy","authors":"Ziyi Weng;Fei Wang;Chao Cai;Qingguo Xie","doi":"10.1109/TRPMS.2025.3628222","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3628222","url":null,"abstract":"A positron emission tomography (PET) system often entails thousands of readout channels that sample fast scintillation pulses. Lowering the complexity per channel, especially in the hardware design, can effectively reduce power consumption, minimize mutual-channel interference, and overall cost. Despite multiple voltage threshold (MVT) sampling’s energy efficiency, its excessive resource consumption limits high-density PET applications. This article proposes a fully FPGA-only MVT architecture reducing resource usage by over 80%, employing a differential timestamping strategy. The key is to separate different timestamps with respect to each voltage reference in a sequential order and alter their corresponding triggering edges so that they can be identified. Moreover, a two-phase time-to-digital converter (TDC) is introduced. In the first phase, the logic level of an input signal is altered at programmable intervals and timestamps are imprinted on signal edges. This can shorten the time intervals to be interpolated in the second phase where conventional tapped delay line (TDL) is utilized. Furthermore, an integrated Sigma-Delta digital-to-analog converter (DAC) eliminates external hardware, enabling fully FPGA-based sampling. Experimental results demonstrate that the proposed method achieves 244 ps coincidence timing resolution (CTR), Full Width at Half Maximum and 14.6% energy resolution using less than 20% resources of the conventional method, enabling scalable high-density readout.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"743-752"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808139","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}

{"title":"Learning From Multimodal Privileged Information to Predict Radiation Dose for Glioblastoma Radiotherapy","authors":"Zhijie Zhang;Wenwen Zhang;Yantao Niu;Baohua Sun;Suqing Tian;Lihua Zhu;Wei Zhao","doi":"10.1109/TRPMS.2025.3627297","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3627297","url":null,"abstract":"Glioblastoma (GBM) is one of the most common and aggressive malignant brain tumors, with radiotherapy being the primary treatment modality. Designing an optimal radiotherapy plan, however, is often a labor-intensive process. To accelerate the planning process, several dose prediction methods have been developed with promising outcomes. Yet, existing dose prediction methods highly rely on the contouring of organs-at-risks (OARs) and tumors, which are typically cumbersome and time-consuming. In this study, we propose a mask-free dose prediction method leveraging multimodal privileged information within the framework of generalized knowledge distillation. The proposed network is a teacher-student model, and the anatomical masks are treated as privileged information, which is available for training the teacher model but not for the student model during inference. The pretrained teacher model transfers the learned anatomical information to the student model via knowledge distillation. To ensure the conformity of the predicted dose map, we design a high-dose contour loss. To the best of our knowledge, this is the first approach that learns from multimodal information to predict radiation dose distribution without requiring contour structures. Our experimental results demonstrate that incorporating multimodal information significantly assists the model in learning the tumor characteristics, even in the absence of anatomical contours.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"815-824"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808103","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}

{"title":"Experimental Evaluation of DE-SPECT: A Hyperspectral SPECT System for Region-Selective 3-D Gamma-Ray Spectroscopy","authors":"Yifei Jin;Elena Zannoni;Poopalasingam Sankar;Feng Zhang;Michael Streicher;Hao Yang;Zhong He;Scott D. Metzler;Chi Liu;Albert J. Sinusas;Lingjian Meng","doi":"10.1109/TRPMS.2025.3630554","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3630554","url":null,"abstract":"This study presents an experimental evaluation of the dynamic extremity SPECT (DE-SPECT) system, specifically engineered for precise, region-selective gamma-ray spectroscopy in the diagnosis of peripheral vascular disease (PVD) in lower extremities. The system incorporates cadmium zinc telluride (CZT) imaging spectrometers and dynamic dual-field-of-view (FOV) collimators to facilitate comprehensive, multifunctional molecular imaging. The CZT detectors, with depth-of-interaction (DOI) capabilities, deliver an exceptional energy performance across a wide energy range up to 600 keV. The novel dual-FOV aperture system allows selective imaging with two configurations: 1) a 28-cm diameter wide FOV suitable for dual-leg or scout imaging and 2) a 16-cm diameter high-resolution, and high-sensitivity (HR-HS) FOV designed for single-leg or focused imaging. Utilizing uniform phantoms, resolution phantoms, and multitracer phantoms, we experimentally assessed the system’s sensitivity, spatial resolution, and multitracer imaging capabilities. Spatial resolutions were approximately 6 mm in HR-HS-FOV mode and between 8 and 10 mm in wide-FOV mode. peak-to-valley ratios, indicative of image clarity, improved with enhanced DOI resolutions, rising from 1.03 to 1.22. The system’s ability to perform multitracer imaging, essential for deriving multifunctional molecular information, further highlights its potential to significantly enhance diagnostic accuracy for PVD.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"764-778"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808106","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}

{"title":"Head-to-Head Comparison of Brain Imaging Quality Between Brain-Dedicated PET and Whole-Body PET Based on Phantom and Clinical Data","authors":"Lei Fang;Yuan Xia;Qiongqiong Fang;Feihong Xu;Ran Hong;Bingxuan Li;Bo Zhang;Ang Li;Wanchun Zhang;Peng Xiao;Qingguo Xie","doi":"10.1109/TRPMS.2025.3629097","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3629097","url":null,"abstract":"A comprehensive comparison of imaging quality between the brain-dedicated all-digital Positron emission tomography (PET) system, DigitMI i30, and the whole-body DigitMI 930 system was performed using a series of phantom studies and clinical evaluations. Image quality was assessed with the point source and Derenzo phantom for spatial resolution, the National Electrical Manufacturers Association (NEMA) image quality phantom for contrast recovery coefficient (CRC) and background variability, and the Hoffman phantom for gray–white matter contrast, signal-to-noise ratio (SNR), and coefficient of variation (COV). Clinical data were further analyzed to evaluate gray–white matter contrast, SNR, COV, and image sharpness. The DigitMI i30 achieved finer central spatial resolution (2.0 mm versus 2.4 mm), although its resolution degraded faster than the DigitMI 930 with radial offset. In NEMA image quality tests, the DigitMI i30 exhibited higher CRC (77.2% versus 64.9%) and lower background variability (6.5% versus 7.2%) for the 10 mm sphere. Hoffman phantom results and clinical data further confirmed its advantages, revealing sharper anatomical detail and improved gray–white matter contrast. These findings demonstrate that the physical advantages of a brain-dedicated PET system translate into clinically meaningful improvements, supporting its value as a precise and reliable tool for neuroimaging.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"779-788"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808107","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}

{"title":"Microdosimetry Maps in Continuous and Pulsed Alpha Beams","authors":"Yuwei Zhu;Carla Riera-Llobet;Celeste Fleta;María Manna;Amélia Maia Leite;Gastón García López;Vicente García Távora;Diana Bachiller-Perea;Quentin Mouchard;Consuelo Guardiola","doi":"10.1109/TRPMS.2025.3620266","DOIUrl":"https://doi.org/10.1109/TRPMS.2025.3620266","url":null,"abstract":"The experimental assessment of the linear energy transfer (LET) at micrometric scale with high resolution is mandatory to achieve biologically optimized treatment plans in Hadron therapy. Therefore, it is crucial to create active radiation sensors able to quantify microdosimetry distributions in three-dimensions under clinical conditions. We have quantified the first microdosimetry 2-D-maps delivered by alpha beams with the highest resolution so far (<inline-formula> <tex-math>$400~mu $ </tex-math></inline-formula>m <inline-formula> <tex-math>$times 600~mu $ </tex-math></inline-formula>m) using a customized sensor. It consists of 3 multiarrays of <inline-formula> <tex-math>$50~3times 3$ </tex-math></inline-formula> microdetectors with a pitch of <inline-formula> <tex-math>$200~mu $ </tex-math></inline-formula>m that cover a total radiation-sensitive area of 9 cm <inline-formula> <tex-math>$times 0$ </tex-math></inline-formula>.4 mm. The microdetectors are based on a novel 3-D-cylindrical silicon architecture with 25-<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>m diameter and 20-<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>m thickness. The back-side of these sensors is etched to avoid back-scattering contributions in heavy ion fields. This system was calibrated with alpha beams from 5.75 to 20 MeV in the Accélérateur Linéaire et Tandem à Orsay (ALTO), France, and irradiated with pulsed beams (from <inline-formula> <tex-math>$1.4~mu $ </tex-math></inline-formula>s to <inline-formula> <tex-math>$3.3~mu $ </tex-math></inline-formula>s pulse widths) of 7.5 MeV alphas in the Centro de Micro-Análisis de Materiales (CMAM), Spain. We worked with clinical equivalent fluence rates in both continuous irradiation mode (<inline-formula> <tex-math>$sim ~10{^{{8}}}$ </tex-math></inline-formula> cm<inline-formula> <tex-math>${^{-}2 } cdot $ </tex-math></inline-formula>s<inline-formula> <tex-math>${}^{-}1 $ </tex-math></inline-formula>) and in pulsed regimes with higher fluence rates up to <inline-formula> <tex-math>$sim ~10{^{{10}~}}$ </tex-math></inline-formula>cm<inline-formula> <tex-math>$cdot {^{-}2 }$ </tex-math></inline-formula>s<inline-formula> <tex-math>${}^{-}1 $ </tex-math></inline-formula>. Measurements of pulse height spectra were collected with the sensor in both modes. The experimental results were crosschecked with Monte Carlo simulations using the GATE code. The simulated spectral shapes and peak positions showed good agreement with the experimental data, with relative differences of less than 3.5% for all the energies studied. Results show that the new microdosimetry array, specially manufactured for ions, is a reliable tool for LET characterization of ions in both continuous and pulsed modes.","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"799-814"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808166","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}

{"title":"IEEE Transactions on Radiation and Plasma Medical Sciences Publication Information","authors":"","doi":"10.1109/TRPMS.2026.3684591","DOIUrl":"https://doi.org/10.1109/TRPMS.2026.3684591","url":null,"abstract":"","PeriodicalId":46807,"journal":{"name":"IEEE Transactions on Radiation and Plasma Medical Sciences","volume":"10 5","pages":"C2-C2"},"PeriodicalIF":3.5,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11506013","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147808112","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}