Physics in medicine and biology最新文献

{"title":"Mapping human brain networks using resting-state metabolic connectivity: concepts, methods, and foundational results.","authors":"Sharna Jamadar","doi":"10.1088/1361-6560/aeae9e","DOIUrl":"https://doi.org/10.1088/1361-6560/aeae9e","url":null,"abstract":"<p><p>The human brain is a complex network, whose functional properties are constrained by its anatomy and energetic constraints. Metabolic connectivity refers to coherent metabolic interactions between spatially distinct regions, and is most commonly measured using [18F]-fluorodeoxyglucose (FDG) positron emission tomography (PET). While metabolic connectivity as a field predates fMRI-derived functional connectivity measures, it was not until recently that it was possible to estimate a metabolic network at the individual person level. Here, we review the theoretical and historical contexts of metabolic connectivity, its physiological bases, major methodological approaches, and key findings in health and disease. The review highlights that metabolic connectivity yields unique insight into the functional interactions of the brain relative to fMRI, is predictive of cognition and individual phenotype, and displays unique patterns of time-variant state switching and gradient architecture. We conclude the review with a forward-looking perspective of the future of metabolic connectivity, and recommendations for extending the conceptual framework of the measure.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149112976","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Self-cascade latent Schrödinger bridge for CT field-of-view extension.","authors":"Xihao Wu, Jiacheng Liu, Haijun Yu, Andreas Maier, Yixing Huang","doi":"10.1088/1361-6560/aea5d0","DOIUrl":"10.1088/1361-6560/aea5d0","url":null,"abstract":"<p><p><i>Objective.</i>Computed tomography (CT) field-of-view (FOV) truncation leaves peripheral anatomy entirely absent, introducing systematic errors in radiotherapy dose calculation and body-composition analysis. Existing methods either rely on unavailable raw projection data or suffer from over-smoothing and hallucination under severe truncation.<i>Approach.</i>We propose the self-cascade latent Schrödinger bridge (SCL-SB), combining a latent image-to-image Schrödinger bridge (ISB) with a self-cascade mechanism in a single shared-weight U-Net. ISB constructs a diffusion bridge from the truncated-image distribution to the full-FOV distribution, enabling high-quality reconstruction in ten denoising steps (faster than a 100-step denoising diffusion implicit model baseline). The self-cascade mechanism runs two sequential rounds sharing identical weights: the first produces an initial extrapolation; the second refines it using the first-round output as a hint. A frequency-aware residual mechanism amplifies gradient contributions at high-frequency inter-round residuals to stabilise cascade training and recover anatomically faithful detail.The truncation radius is sampled from a stratified distribution spanning mild to severe truncation during training, enabling a single model to cover the full severity spectrum without radius-specific fine-tuning.<i>Main results.</i>On a head-to-chest CT dataset spanning three truncation severities, SCL-SB, trained as a single model without radius-specific tuning, achieves the highest structural similarity index measure (SSIM) and a high-frequency ratio (HF-Ratio) closest to ideal among all compared methods at every severity, running nearlyfaster than the latest vision transformer-based baseline. Ablation shows that cascade training alone already improves the first round's output without running the second round at inference, an effect we term cascade implicit enhancement. Relative to the same architecture trained without cascade (ISB), the model recovers more high-frequency anatomical detail (HF-Ratio) and achieves a higher SSIM by() at severe truncation, at no additional inference cost.<i>Significance.</i>SCL-SB addresses the posterior-mean regression that blurs high-frequency anatomical detail in ill-posed FOV extrapolation, produces spatially continuous Hounsfield unit reconstructions free of the patch-boundary discontinuities introduced by Vision-transformer-based architectures, and avoids the spectral hallucination characteristic of generative adversarial network-based methods. Although downstream clinical validation on larger, multi-centre cohorts remains necessary, these properties can potentially improve image guidance in clinical applications such as adaptive radiotherapy, body-composition assessment, and spine surgery.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148930395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Autocalibration signal-guided multi-frequency diffusion model for dynamic MRI reconstruction.","authors":"Yu Guan, Wenchao Deng, Qi Qi, Minghui Zhang, Haikun Qi, Xiang Ji, Qiegen Liu","doi":"10.1088/1361-6560/aea51a","DOIUrl":"10.1088/1361-6560/aea51a","url":null,"abstract":"<p><p><i>Objective.</i>The autocalibration-signal (ACS) region plays a pivotal role in dynamic magnetic resonance imaging (MRI) reconstruction, yet its full potential remains under-exploited in current generative frameworks, which rely mainly on data-driven priors while overlooking ACS-embedded physical information. This work aims to propose a novel multi-frequency aware diffusion model guided by structural priors derived from the<i>ACS</i>region to enhance the quality of dynamic MRI reconstruction.<i>Approach.</i>Specifically, the intrinsic phase consistency and spatial smoothness within the<i>ACS</i>region are leveraged to construct a robust prior that stabilizes the diffusion trajectory. This prior further guides the restoration of high-frequency components, thereby improving temporal coherence and structural fidelity. Furthermore, the non-ACS region is divided into two complementary frequency bands, each jointly modeled with the<i>ACS</i>region to establish frequency-conditioned diffusion trajectories across the frequency domain. To further refine reconstruction performance, we incorporate low-rank regularization across time frames and enforce data-consistency constraints, which effectively suppress motion artifacts while preserving fine anatomical details.<i>Main results.</i>Quantitative evaluations demonstrate that the proposed method achieves competitive reconstruction quality across various sampling patterns and acceleration factors. Under Radial sampling at an acceleration factor of<i>R</i>= 8, it attains PSNR of 32.53 ± 0.34 dB, NMSE of 0.011 ± 0.001 and SSIM of 0.8583 ± 0.0057, outperforming all competing methods. Notably, the proposed method achieves competitive performance comparable to supervised models without requiring any fully-sampled data or external training labels.<i>Significance.</i>The proposed method enables high-quality dynamic MRI reconstruction by exploiting intrinsic<i>ACS</i>information, offering a practical solution without fully-sampled training data.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148920432","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Monte Carlo evaluation of a novel hybrid-shielded gamma probe with integrated counting and emission direction-finding capabilities.","authors":"O B Kolcu, T Yetkin, A T Zengin, E Iren","doi":"10.1088/1361-6560/aea9e7","DOIUrl":"10.1088/1361-6560/aea9e7","url":null,"abstract":"<p><p><i>Objective.</i>Gamma probes for sentinel lymph node localization usually work in counting mode only, with no information on the angular origin of detected gammas. Active-shielded designs that add emission direction finding reach about 95%-97% shielding effectiveness, below the >99% level of passive-shielded commercial probes. We propose and evaluate a hybrid-shielded gamma probe combining both capabilities.<i>Approach.</i>The module integrates a central GAGG(Ce) scintillator for counting, an asymmetric lead or sintered tungsten shielding layer, and an outer ring of GAGG(Ce) scintillators that provide passive attenuation and direction finding functionality. The scintillators are read out by asilicon photomultiplier array. We performed GEANT4 Monte Carlo simulations for^{99 m}Tc source. Sensitivity, spatial and angular resolution, and shielding effectiveness were characterized according to the NEMA NU3-2004 protocol in air and water, and direction-finding accuracy was assessed over a 360azimuthal scan with a centre-of-gravity (COG) estimator and a von Mises maximum-likelihood fit.<i>Main results.</i>Shielding effectiveness reachedwith the lead collimator andwith the sintered tungsten one, both within the range of passive-shielded commercial probes. Sensitivity at 10 mm in air was around 4560 cps MBq, with spatial and angular resolutions comparable to commercial devices. Over the 360scan, the mean absolute deviation of the estimated direction was 4.05COG and 4.55(maximum-likelihood), with a maximum below 7.2in both cases.<i>Significance.</i>Keeping shielding and directional sensing in separate elements allows each to be optimized independently. To our knowledge, this is the first gamma probe concept combining emission direction finding with shielding above 99%, which has been the main limitation of active-shielded designs. GAGG(Ce), not used in any commercial probe yet, is non-hygroscopic and well matched to silicon photomultipliers. These results motivate experimental prototyping and clinical validation.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149001672","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development of a novel material specific imaging algorithm and dual-energy CBCT technology for the next generation imaging solution on O-ring linacs.","authors":"Bryanna Smith, Mike Sattarivand","doi":"10.1088/1361-6560/aea9e8","DOIUrl":"10.1088/1361-6560/aea9e8","url":null,"abstract":"<p><p><i>Objective.</i>Dual-energy (DE) imaging techniques using cone-beam computed tomography (CBCT) have applications for radiotherapy (RT) including material separation and quantification, improved contrast-to-noise ratios (CNR), and reduced metal artifacts. Traditional limitations of CBCT can be addressed by using HyperSight imaging on O-ring gantry RT units, however, DE imaging technologies have yet to be established for HyperSight. The objective of this study was to implement DE technology for HyperSight by developing a novel material specific imaging algorithm and adopting a virtual monoenergetic imaging (VMI) method.<i>Approach.</i>Images of a calibration phantom with various material inserts, including concentrations of iodine and calcium, were obtained at 80 and 140 kVp. A novel material separation algorithm was developed using DE Hounsfield unit values with statistical classification via the Mahalanobis distance and confidence filtering to differentiate six material groups (lung, soft tissue, fat tissue, iodine, calcium, bone). The algorithm includes concentration/density mapping via interpolation between calibrated values and was evaluated using an anthropomorphic phantom. VMIs were developed using an established method based on weighted addition of high- and low-energy images to quantify CNR enhancement and metal artifact reduction (MAR).<i>Main results.</i>Iodine and calcium were separable for concentrations above 47 mg I ml^-1and 303 mg Ca ml^-1, whereas iodine and bone were separable above 25 mg I ml^-1and 1.4 g cm^-3, respectively. Soft tissue, fat tissue, and lung demonstrated reasonable classification with accuracies of 94.6%, 91.7%, and 97.8%, respectively. At 75 keV, CNR improved by 52% for inner bone, 41% for liver, and 88% for exhale lung compared to the single-energy (SE) reference. At 70 keV, MAR improved by 11% compared to the SE reference.<i>Significance.</i>This study demonstrates the feasibility of DE CBCT technology on the HyperSight imaging system through a novel statistical material specific imaging algorithm and VMI-based CNR improvements and metal artifact reduction.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149002500","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Beyond NEMA: robustness, DOI encoding, and multi-mouse imaging performance of the Bruker PET/CT Si78.","authors":"Laura Moliner, Carlos Correcher, Adrian Beltrá, Cesar Molinos, Todd Sasser, Alejandro Lucero, Keshav Nagpal, Noriel Pavón, Michael Heidenreich, Jose María Benlloch","doi":"10.1088/1361-6560/aea8c7","DOIUrl":"10.1088/1361-6560/aea8c7","url":null,"abstract":"<p><p><i>Objective.</i>This work presents a comprehensive evaluation of the Bruker PET/CT Si78 preclinical imaging system, combining standardized performance assessment, inter-system reproducibility analysis, and investigation of depth-of-interaction (DOI) encoding effects on image quality.<i>Approach.</i>Ten independently assembled systems were characterized following the National Electrical Manufacturers Association NU 4-2008 protocol to assess spatial resolution, sensitivity, count rate performance, and image quality. Beyond standard evaluation, this study investigates the impact of DOI estimation using simulations, controlled experiments, and<i>in vivo</i>multi-mouse imaging.<i>Main results.</i>The results demonstrate consistent performance across the evaluated systems, with nearly uniform spatial resolution across the field of view (FOV), peak sensitivity around 10.5%, and low variability among systems, typically below 5%, indicating low inter-system variability across most evaluated metrics-. Continuous DOI estimation, enabled by monolithic detector technology, improves spatial resolution and its uniformity compared with dual-layer and non-DOI approaches, particularly at the periphery of the imaging volume where parallax effects are most pronounced. These improvements were consistently observed in both simulated and experimental datasets. The practical implications of DOI encoding were demonstrated in multi-animal imaging studies, where enhanced visualization of fine anatomical structures and improved quantitative accuracy were achieved. In particular, the absence of DOI information led to substantial spatially dependent quantification errors, while dual-layer DOI partially mitigated these effects. Continuous DOI estimation provided the smallest deviations and most spatially uniform results among the evaluated configurations, supporting reliable quantification in demanding applications such as brain imaging. Significance. Overall, the PET/CT Si78 system combines competitive performance with high reproducibility and advanced DOI capabilities, enabling uniform image quality across the FOV and supporting efficient multi-animal studies. These characteristics supports its use for reproducible and ethically optimized preclinical imaging aligned with the principles of reduction and refinement in animal research.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148983310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Assessment of data-driven gating for cardiac motion extraction and LVEF estimation from routine [¹⁸F]FDG LAFOV PET.","authors":"Nanna Overbeck, Ulrich Lindberg, Esben Andreas Carlsen, Anders Bertil Rodell, Paul Schleyer, Jorge Cabello, Philip Hasbak, Thomas Lund Andersen, Flemming Littrup Andersen","doi":"10.1088/1361-6560/aea2eb","DOIUrl":"10.1088/1361-6560/aea2eb","url":null,"abstract":"<p><p><i>Objective.</i>Chemotherapy-induced cardiotoxicity can lead to irreversible heart failure. Left ventricular ejection fraction (LVEF) is routinely monitored during treatment, but conventional assessment requires dedicated cardiac imaging and clinical resources. This study evaluated a deviceless data-driven gating (DDG) framework for extracting cardiac motion from routine [¹⁸F] FDG positron emission tomography (PET) emission data and, secondary, assessed its feasibility for LVEF estimation.<i>Approach.</i>The DDG framework is based on histo images and uses anatomical masking for frequency-domain filtering to obtain the cardiac signals of 169 [¹⁸F]FDG PET/CT examinations. The cardiac gating performance was evaluated, and LVEF estimates were compared with the patient's respective clinical reference method; echocardiography or multigated acquisition (MUGA).<i>Main results.</i>The DDG framework successfully extracted a cardiac gating signal in 138 of 169 examinations (81.7%). Sufficient myocardial [¹⁸F]FDG uptake for software-based LVEF estimation was present in 101 patients (60%), and LVEF was successfully estimated in all cases. Compared with the reference methods, the DDG-based LVEF estimates demonstrated a mean bias of 3.4% relative to echocardiography and -2.6% relative to MUGA. Agreement was strongest with echocardiography, although the limits of agreement exceeded the threshold required for interchangeable clinical use. The extracted pulse frequencies were physiologically plausible and showed good agreement with the corresponding reference examinations, supporting the validity of the DDG-derived cardiac signal.<i>Significance.</i>The deviceless DDG framework can successfully extract clinically relevant cardiac motion directly from routine [¹⁸F] FDG PET acquisitions without external hardware. The method achieved a high success rate with respect to gating the images and provided LVEF estimates that showed good agreement with established clinical reference methods, supporting the feasibility of deriving functional cardiac information from standard PET examinations. The proposed approach provides a promising foundation for retrospective functional cardiac assessment and has the potential to complement conventional LVEF evaluation while reducing additional patient burden and simplifying clinical workflows.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148892263","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Learning-enhanced hybrid planar-SPECT dosimetry for voxel-level Monte Carlo dose estimation in radionuclide therapy.","authors":"Zhengkun Dong, Xiao Jin, Xiangxi Meng, Wei Liu, Nan Li, Zhi Yang, Qiushi Ren, Jiangyuan Yu, Zhaoheng Xie","doi":"10.1088/1361-6560/aea39b","DOIUrl":"10.1088/1361-6560/aea39b","url":null,"abstract":"<p><p><i>Objective.</i>Quantitative imaging-based dosimetry is essential for optimizing [¹⁷⁷Lu]Lu-DOTATATE peptide receptor radionuclide therapy, yet full multi-time-point (MTP) SPECT/CT is difficult to implement in routine clinical practice. This study proposes a hybrid planar-SPECT framework for organ-level time-integrated activity (TIA) estimation and single-time-point (STP)-constrained voxel-level Monte Carlo (MC) dose analysis from a single post-therapy SPECT/CT acquisition.<i>Approach.</i>The proposed framework integrates STP SPECT/CT with serial planar imaging to derive organ-level TIAs. A physics-informed support vector regression (SVR) model is incorporated to refine planar-to-SPECT scaling factors and improve the robustness of planar-derived activity quantification. The resulting organ TIAs are used to scale the fixed within-organ STP SPECT pattern and generate voxelized TIA maps as source distributions for MC-based dose calculation. The method is evaluated against reference MTP SPECT/CT data in 11 patients and further applied to a separate cohort of 23 patients to explore vertebral dose heterogeneity and its association with hematologic toxicity.<i>Main results.</i>Hybrid dosimetry showed good agreement with MTP SPECT/CT, with organ-level errors of approximately 10%-15%. The SVR-based refinement modestly reduced scaling bias, particularly in anatomically challenging structures. Voxel-level vertebral dose heterogeneity metrics showed moderate but exploratory correlations with hematologic toxicity (<i>ρ</i>≈ 0.4-0.5) after false-discovery-rate correction, whereas conventional organ-averaged dose metrics showed no significant association.<i>Significance.</i>The proposed hybrid planar-SPECT framework supports clinically feasible organ-level dosimetry from a reduced acquisition protocol. The resulting voxelized maps provide STP-constrained approximate source distributions for MC dose calculation and exploratory characterization of spatial dose heterogeneity, rather than reconstructions of time-varying MTP voxel kinetics. The SVR improvement observed in the 11-patient internal validation remains preliminary, requiring larger cohorts for robustness and external validation for generalizability.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148907002","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Real-time optical respiratory guidance for upright particle therapy: a pilot human-in-the-loop study.","authors":"J Chen, G Marchesi, I Barbato, M Ciocca, E Orlandi, M R Fiore, R Bellazzi, G Baroni, A Pella","doi":"10.1088/1361-6560/aea736","DOIUrl":"10.1088/1361-6560/aea736","url":null,"abstract":"<p><p><i>Objective.</i>To develop and evaluate a real-time optical respiratory guidance framework for upright particle therapy, focusing on respiratory regularity, guided-task performance, posture-dependent effects, and participant-specific task fit.<i>Approach.</i>A real-time guidance platform was implemented using optical motion tracking as the primary analytic channel, integrated with a Unity-based visual guidance interface and auxiliary pressure-belt and Anzai reference signals. After phantom-based assessment of cross-system consistency, eleven healthy volunteers completed breathing tasks under guided feedback, guided no-feedback (GNF), and no-target conditions in upright and supine postures. Outcomes included target-following performance, respiratory regularity, and posture-related respiratory organization.<i>Main results.</i>Phantom-based validation showed low residual bias and sub-millimeter error between reference-aligned optical tracking systems. In the volunteer study, real-time feedback improved guided-task performance relative to GNF, with lower follow error and greater time spent within the acceptable green band. Guided breathing improved respiratory regularity relative to no-target condition breathing, with lower period variability and higher cycle similarity. Posture did not produce a simple global performance difference, but it reshaped thoracoabdominal organization, with greater abdominal dominance in supine than in upright posture.<i>Significance.</i>The proposed framework supports synchronized, human-in-the-loop respiratory guidance for upright particle-therapy workflows. By combining real-time motion-derived feedback with posture-aware task execution, this pilot study provides evidence that respiratory guidance can be designed not only as a monitoring tool, but also as an interactive support strategy for clinically relevant motion-management workflows. These findings motivate future work on external-internal motion mapping, gating efficiency, treatment-session robustness, and prospective personalization of both target definition and interface design.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148970135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Robustness of deep learning-based denoising 4DCBCT methods.","authors":"Samuele Papa, Efstratios Gavves, Jan-Jakob Sonke","doi":"10.1088/1361-6560/aea8c6","DOIUrl":"10.1088/1361-6560/aea8c6","url":null,"abstract":"<p><p><i>Objective.</i>4D cone-beam computed tomography (4DCBCT) is a technique used to address respiratory motion in radiotherapy but is limited by significant view-aliasing artifacts. Recently, deep learning methods have been proposed to reduce view-aliasing. This study investigates the robustness and performance of these methods when variations in patient breathing period affect the noise pattern in the scans used as training data, which potentially affects the ability of the models to reduce view-aliasing.<i>Approach.</i>We evaluated both a supervised and a self-supervised scalable deep learning method. Using a dataset of 328 patients, scans were partitioned into cohorts according to breathing period, and independent training with cross-cohort validation was performed to assess robustness. We also evaluated the impact of a modern UNet-based architecture and data augmentation.<i>Main results.</i>Training and testing on matched breathing-period cohorts did not consistently improve performance, indicating that both methods are robust to clinically observed variation in average breathing period. Instead, increasing the overall dataset size led to the most noticeable improvements. For the cohort with long breathing periods, the supervised model trained on the combined dataset achieved aPSNR improvement over the cohort-specific model and a PSG improvement of.<i>Significance.</i>These findings support the reliability of deep learning-based 4DCBCT denoising across the observed range of average breathing periods. They suggest that large, diverse datasets should be prioritized over breathing-cohort-specific optimization. Finally, PSG provides a quantitative indicator that is sensitive to view-aliasing and may support future development of more robust 4DCBCT enhancement techniques.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.1,"publicationDate":"2026-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148983393","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}