Physics in medicine and biology最新文献

{"title":"Integrating AI-assisted image enhancement with physics-based synthesis of low-field MRI from high-field MRI.","authors":"Dang Bich Thuy Le, Muller De Matos Gomes, Anh Truong, Hoang Le, Dnyanada Kadiyal, Aleksandar Nacev, Ram Narayanan, Yuling Yan","doi":"10.1088/1361-6560/ae6017","DOIUrl":"10.1088/1361-6560/ae6017","url":null,"abstract":"<p><p><i>Objective.</i>Low-field magnetic resonance imaging (MRI) offers distinct advantages in terms of affordability, portability, and accessibility. However, its widespread adoption is limited by an inherently low signal-to-noise ratio (SNR) and reduced spatial resolution. This study proposes an AI-assisted framework to enhance low-field MRI image quality and overcome these limitations.<i>Approach.</i>We propose a two-stage framework to generate high-quality low-field MRI images. In the first stage, realistic low-field images are synthesized from high-field acquisitions using a physics-informed forward model that incorporates spiral<i>k</i>-space trajectories and accounts for nonlinear magnetic field gradients,<i>B</i>₀inhomogeneity,<i>k</i>-space undersampling, and image reconstruction characteristics of low-field systems. In the second stage, a 3D U-Net enhanced with a multi-head attention in a vision transformer (ViT) module is trained on paired synthetic low- and high-field images to serve as a post processing following conventional image reconstruction.<i>Main results.</i>On the synthetic test set, our framework demonstrates strong performance, achieving a peak SNR (PSNR) of 19.08 ± 2.85 dB for the baseline U-Net model and 21.00 ± 2.50 dB with the ViT block, demonstrating high reconstruction fidelity. The structural similarity index measure reaches 0.6456 ± 0.0779 (without ViT) and 0.6639 ± 0.0798 (with ViT), along with low normalized root mean squared error values of 0.3866 ± 0.0952 and 0.3084 ± 0.0695, respectively. These results highlight significant improvements in both image quality and reconstruction robustness. The trained network, applied as a post-processing step after conventional reconstruction, consistently enhances the contrast-to-noise ratio of the output images, supporting the qualitative observations of improved image contrast and clarity.<i>Significance.</i>The proposed framework addresses key limitations hindering the broader adoption of low-field MRI, including noise, artifacts, and resolution loss inherent to low-field acquisitions. By integrating deep learning with physics-based simulations, the approach achieves notable qualitative and quantitative enhancements in denoising, artifact removal, and overall image quality. These results highlight the framework's potential to improve the practical utility of low-field MRI substantially.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147691544","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":"Few shots transfer learning for universal SPECT denoising across diverse acquisition protocols.","authors":"Boyang Pan, Jianchen Pan, Kexin Gan, Yuting Shen, Xiaoxiao Chen, Langdi Zhong, Hang Yang, Jie Chen, Laiping Xie, Wei Guo, Hongmin Li, Nan-Jie Gong","doi":"10.1088/1361-6560/ae5eb8","DOIUrl":"10.1088/1361-6560/ae5eb8","url":null,"abstract":"<p><p><i>Objective.</i>Accelerated single photon emission computed tomography (SPECT) imaging, achieved by reducing either the number of projection angles or the acquisition time per angle, enhances clinical workflow efficiency but introduces elevated noise. This study aims to develop and validate a universal DL-based reconstruction framework that effectively generalizes across diverse, clinically-realistic SPECT acceleration protocols by overcoming the data scarcity challenge.<i>Approach.</i>SPECT bone scans from 103 patients were acquired under a standard scan (60 views, 12 s/view, 60v12s) followed by a fast scan using one of five acceleration protocols (60v6s, 60v3s, 30v12s, 30v6s, 16v12s). A<i>U</i>-Net-based reconstruction framework was implemented using three strategies: (1) single model, trained on individual acceleration protocol), (2) base model, trained on aggregated datasets from five acceleration protocols, and (3) transfer model, fine-tuned from the base model for protocol-specific optimization. Pixel-level accuracy and structural similarity were assessed using peak signal-to-noise ratio (PSNR), structural similarity index (SSIM) metrics, and maximum pixel value of lesions. Clinical evaluation of image quality, radionuclide detail, artifacts, and diagnostic confidence was conducted using a 5-point system.<i>Main results.</i>Quantitative evaluation showed the transfer model achieved better PSNR and SSIM across all protocols (highest 48.02 PSNR and 0.9918 SSIM in 30v6 s protocol). Qualitative analysis confirmed enhanced structural fidelity. Clinical evaluations rated the transfer model highest across metrics, with scores of 4.667 ± 0.508 (image quality, 30v12 s), 4.800 ± 0.250 (radionuclide detail, 60v6 s), 1.150 ± 0.173 (artifact reduction, 60v6 s), and 4.800 ± 0.250 (diagnostic confidence, 60v6 s), surpassing full-scan results in most cases.<i>Significance</i>. The proposed transfer learning (TL) framework effectively addressed data scarcity and improved reconstruction performance across diverse SPECT acceleration scenarios. The adoption of a TL strategy mitigates data scarcity by utilizing shared features and fine-tuning for specific protocols. The framework demonstrates potential for integration into fast SPECT workflows, facilitating reliable use across diverse imaging scenarios.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147675478","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":"Accelerated solution method for 3D phase-based cr-MREPT.","authors":"Mustafa Kaan Çan, Yusuf Ziya Ider","doi":"10.1088/1361-6560/ae64a7","DOIUrl":"https://doi.org/10.1088/1361-6560/ae64a7","url":null,"abstract":"<p><p>Objective Convection-Reaction equation based Magnetic Resonance Electrical Properties Imaging (cr-MREPT), especially in phase-based form, is one of the most commonly used MREPT methods since it can successfully overcome internal boundary artefacts, does not rely on Transceive Phase Approximation (TPA) and is relatively easy to implement. While the partial differential equation system solved in this method introduces regularization compared to Helmholtz MREPT, it also increases the solution time especially when the method is applied in large 3D volumes and reduces the practicality of the applications. Approach We divide the large 3D volume of interest into smaller regions (local ROIs) for which the solutions can be obtained faster, we then parallelize the solutions of the individual regions and finally combine the results to obtain the whole conductivity distribution. Sensitivities of the reconstructed conductivity at a certain voxel, to the B_1 phase data of nearby voxels and to the Dirichlet Boundary condition imposed at the boundary of the solution region are investigated to determine the optimum size for the small regions. Main Results Conductivity distributions for various phantoms are successfully reconstructed with significantly reduced computational times, and up to approximately 100 times acceleration of the solution is achieved using a 72-core server. This can be further increased with a high-performance computer, where theoretical limit is the solution time for a single local ROI. Significance Multislice reconstruction of large phantoms, such as a human head, is possible within a reasonable solution time with the proposed method. These findings enhance the potential for the practical applications of phase-based cr-MREPT.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147778524","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":"E2D-Unroll: efficient equivariant deformable unrolling networks for cardiac cine MRI reconstruction.","authors":"Yuliang Zhu, Zhaochi Wen, Shiying Ke, Yulin Wang, Zhuo-Xu Cui, Qingyong Zhu, Yuanyuan Liu, Jianfeng Ren, Jing Cheng, Chengbo Wang, Dong Liang","doi":"10.1088/1361-6560/ae5c53","DOIUrl":"10.1088/1361-6560/ae5c53","url":null,"abstract":"<p><p><i>Objective</i>. Deep unrolling methods have achieved notable success in cardiac cine MRI reconstruction. However, their effectiveness is often constrained by the limited receptive field of shallow subnetworks and the rigid sampling of standard convolutions on fixed grids. Existing solutions such as transformer-based or multi-scale architectures can enlarge the receptive field, but typically introduce substantial increases in computational cost and model size. This work aims to enlarge the receptive field and improve spatiotemporal feature modeling while keeping the reconstruction model computationally lightweight and parameter-efficient.<i>Approach</i>. We propose an efficient equivariant deformable unrolling network (E2D-Unroll) that integrates three key components: a spatiotemporal deformable module (STDM), a rotation equivariant module (REM), and a gated orientation module (GOM). Specifically, STDM expands the receptive field and adaptively adjusts sampling locations to better capture spatiotemporal features. Next, REM embeds deformable convolutions into a rotation-equivariant framework, allowing kernels to be shared across orientations and thereby improving parameter efficiency. Building upon REM, GOM selectively emphasizes informative orientations to improve the utilization of rotation-equivariant representations.<i>Main results</i>. Extensive experiments on an in-house cardiac cine MRI dataset and the public OCMR dataset demonstrate that E2D-Unroll consistently outperforms state-of-the-art methods in reconstruction accuracy.<i>Significance</i>. E2D-Unroll unifies spatiotemporal deformable convolutions with rotation-equivariant networks to suppress large-scale aliasing artifacts more effectively while maintaining high parameter efficiency and low computational cost, providing a practical solution for accelerated cardiac cine MRI in real-world settings.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147633994","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 and film dosimetry study of collimator effects on penumbra and out-of-field dose for very high-energy electrons.","authors":"Jade Fischer, Antonio Gilardi, Alexander Malyzhenkov, Pierre Korysko, Alexander Hart, Vilde Rieker, Joseph Bateman, Wilfrid Farabolini, Roberto Corsini, Manjit Dosanjh, Magdalena Bazalova-Carter","doi":"10.1088/1361-6560/ae6018","DOIUrl":"10.1088/1361-6560/ae6018","url":null,"abstract":"<p><p><i>Objective.</i>Very high-energy electrons (VHEEs) offer deep penetration, low scattering, and the potential for ultra-high dose rate delivery, making them promising candidates for future radiotherapy. However, the collimation of VHEE beams to achieve sharp beam penumbra remains poorly characterized. This study experimentally and computationally investigates how collimator material, thickness, and beam characteristics affect penumbra and out-of-field dose for VHEEs and establishes an initial foundation for the design of clinically feasible VHEE collimators.<i>Approach.</i>Tungsten, lead, and brass 5 mm diameter collimators were evaluated using film dosimetry with a 200 MeV electron beam delivered at the CERN Linear Electron Accelerator for Research and validated through Monte Carlo (MC) simulations. Experimental measurements of penumbra and out-of-field dose were compared with simulations that systematically varied material (tungsten, lead, brass), thickness (20-80 mm), and beam energy (150-250 MeV). Additional sensitivity tests quantified the impact of beam instability on field shaping.<i>Main results.</i>For measurements in air, penumbrae increased linearly with distance from the collimator and was smallest for tungsten. Out-of-field dose decreased with increasing thickness, falling below 0.5% for a 40 mm thick tungsten collimator. Brass exhibited the highest out-of-field dose (up to 4.8%) and broadest penumbra. MC models reproduced experimental trends within 5% for penumbrae but underestimated out-of-field dose, particularly for brass. The simulations indicated that VHEE beam divergence, beam size and collimator misalignment strongly influence beam penumbra and out-of-field dose.<i>Significance.</i>The presented work demonstrates that collimator material and geometry play a critical role in defining VHEE beam quality. Tungsten provided optimal attenuation and sharpness compared to brass and lead. These results establish quantitative benchmarks for VHEE collimator design and emphasize the importance of beam stability.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147691550","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":"MRDT-GAN: generative adversarial network with multi-scale residual dense transformer generator for low-dose CT denoising.","authors":"Shikai Guo, Jieyu Li, Yi Wu","doi":"10.1088/1361-6560/ae28b0","DOIUrl":"10.1088/1361-6560/ae28b0","url":null,"abstract":"<p><p><i>Objective.</i>Low-dose computed tomography (LDCT) reduces radiation exposure but introduces noise and artifacts that degrade diagnostic quality. Existing deep learning-based denoising methods still face challenges such as over-smoothing, loss of fine structures, and uneven contrast. This study aims to develop an LDCT denoising framework that enhances noise suppression while preserving anatomical details and structural fidelity.<i>Approach.</i>We propose a multi-scale residual dense transformer generative adversarial network (MRDT-GAN). In the generator, we adopt the MRDT Block as the core unit, which introduces multi-scale strategy into residual dense network to reduce over-smoothing and preserve fine details, and also patching transformer block to capture long-range dependencies, mitigating distortions caused by localized receptive fields in convolutional neural network-based approaches. A hybrid attention module is also introduced in the generator to process spatial, frequency, and contrast information, enabling the network to focus on critical regions for noise suppression, improve contrast uniformity, and maintain texture consistency. In the discriminator, we adversarially explore differences on global, pixel, and also sub-scale between denoised LDCT and normal dose CT to better capture structural variations, reduce local noise and distortions, and ensure more realistic texture reconstruction while minimizing artifacts.<i>Main results.</i>We validate MRDT-GAN on both the NIH-AAPM-Mayo Clinic LDCT dataset and a real-world dataset. Experimental results indicate that MRDT-GAN achieves superior denoising performance compared with existing methods, effectively preserves details, enhances visual quality, and achieves a better balance between noise suppression and structural integrity.<i>Significance.</i>MRDT-GAN provides an effective and generalizable LDCT denoising solution that balances noise reduction with fine-detail preservation. By integrating MRDT modeling, hybrid attention mechanisms, and multi-difference adversarial learning, the framework offers improved clinical applicability and supports high-quality image reconstruction for downstream diagnostic tasks.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145687831","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":"Systematic tissue oxygen variation shows the modulation of murine skin radiation toxicity at ultra-high dose rates.","authors":"David Ian Hunter, Jacob P Sunnerberg, Armin D Tavakkoli, Austin Sloop, Beverly Petusseau, Jiang Gui, Xu Cao, Rongxiao Zhang, Harold M Swartz, Lesley A Jarvis, P Jack Hoopes, David Gladstone, Brian W Pogue","doi":"10.1088/1361-6560/ae64a5","DOIUrl":"https://doi.org/10.1088/1361-6560/ae64a5","url":null,"abstract":"<p><strong>Objective: </strong>This study evaluated the hypothesis that baseline tissue oxygen (p_O2) would modulate FLASH toxicity sparing in murine skin, using a wide range of p_O2values, with ultra-high dose rate (UHDR) versus conventional dose rate (CDR) irradiation. &#xD;&#xD;Approach: Murine leg tissue p_O2was systematically varied and measured during irradiation from a FLASH Mobetron 9 MeV linac at 25 Gy, comparing UHDR (≈240 Gy/s) to CDR (≈0.16 Gy/s), for radiation induced skin toxicity outcomes. Baseline tissue p_O2was systematically modulated in 5 different treatment cohorts, using different ranges of inhaled gas (room air, 100% oxygen, or carbogen) and through varying limb vascular compression (partial or full). Radiolytic oxygen consumption, g_O2(mmHg/Gy), was quantified in vivo, and induced macroscopic skin toxicity was scored daily post treatment.&#xD;&#xD;Main Results: FLASH skin sparing was observed at a fixed dose of 25 Gy in groups with partial leg clamping (p_O2≈7±4mmHg), inhaled air (p_O2≈12±6mmHg) and 100% oxygen (p_O2≈16±4mmHg), while reduction in ulceration progression was significant only in the air inhalation group. No FLASH effect was observed under anoxic conditions, via complete blood flow occlusion (p_O2≈0±1mmHg), or when modulated by inhaled carbogen (p_O2≈21±7mmHg). In vivo measurements of radiolytic oxygen consumption, g_O2, correlated to initial p_O2under UHDR conditions (p_O2≈4-16mmHg), with ulceration predominantly occurring at p_O2values above 16mmHg. Inspired carbogen induced the highest p_O2at which point there was no FLASH sparing, for any dose groups between 25 to 15 Gy, despite having large changes in damage with dose. At the specific dose level of 25 Gy studied, the toxicity scores under anoxia for both UHDR and CDR were low (toxicity scores < 1) with no differences observed.&#xD;&#xD;Significance: These findings point to the fact that moderate and low tissue p_O2is associated with diminished oxygen-mediated damage at UHDR but not CDR, seen with inspired room air or 100% oxygen. Anoxic and hyperoxic murine skin are associated with minimal and maximal radiation damage respectively, but also exhibit no apparent FLASH toxicity sparing effect, with further investigation warranted into if the FLASH toxicity sparing effect persists at higher doses under anoxia. &#xD.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147778252","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":"Multi-resolution extended-volume model for iterative reconstruction in cone beam CT.","authors":"Razieh Azizi, Ville-Veikko Wettenhovi, Kati Niinimäki, Ville Kolehmainen","doi":"10.1088/1361-6560/ae64a4","DOIUrl":"https://doi.org/10.1088/1361-6560/ae64a4","url":null,"abstract":"<p><strong>Objective: </strong>Cone beam computed tomography (CBCT) often has a truncated acquired field of view (FOV) due to the limited detector size, leading to image reconstruction from truncated projection data. CBCT reconstructions using an image volume that just encloses the acquired FOV exhibit reconstruction artifacts due to attenuation in the tissues outside the image volume. On the other hand, extending the high-resolution voxel volume far enough beyond the FOV to fully enclose the imaged body often leads to a significant increase of the computational complexity in model based iterative reconstruction techniques. We propose a multi-resolution reconstruction model that eliminates the out-of-FOV reconstruction artifacts and enables accurate recovery of Hounsfield Unit (HU) values within &#xD;the FOV.&#xD;&#xD;Approach: We propose a multi-resolution extended reconstruction volume (MR-ERV) approach that extends the image volume beyond the FOV using separate extension volumes with coarser voxel representation, leading to appropriate modeling of the observed rays outside the FOV without significant increase of the computational complexity. Furthermore, we demonstrate that by augmenting the model with a simple projection extrapolation yields a further reduction of the out-of-FOV artifacts. In this study, the model is evaluated with model based iterative reconstruction minimization using high-resolution 3D CBCT data. The optimization problems considered are non-negativity constrained least-squares estimation, with and without regularization. The optimization is performed using a Primal-Dual Hybrid Gradient (PDHG) algorithm.&#xD;&#xD;Results: The proposed MR-ERV model effectively removes out-of-FOV reconstruction artifacts and it also achieves accurate HU values within the FOV when the volume extension fully encloses the imaged body in the transaxial direction. &#xD;&#xD;Significance: The MR-ERV model provides a platform for computationally efficient and accurate model based iterative reconstruction of CBCT data.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147778646","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":"Atmospheric pressure influence on the charge collection efficiency of air-vented ionization chambers in ultra-high dose per pulse electron beams for FLASH radiotherapy.","authors":"Miguel Angel Flores-Mancera, Andreas Schüller, José Paz-Martín, Araceli Gago-Arias, Wesley Culberson, Faustino Gómez","doi":"10.1088/1361-6560/ae556b","DOIUrl":"10.1088/1361-6560/ae556b","url":null,"abstract":"<p><p><i>Objective.</i>This work studies the dependency of the charge collection efficiency (CCE) on the atmospheric pressure of commercial ionization chambers (ICs) using an ultra-high dose-per-pulse (DPP) electron beam and proves a theoretical model of this phenomenon.<i>Approach.</i>A custom-made PMMA water phantom, water- and pressure-tight sealed, was connected to a vacuum/pressure pump to vary its inside pressure from 900 hPa to 1100 hPa. ICs were placed at different depths in water and irradiated with ultra-high DPP electron beams (20 MeV, 1.2<i>μ</i>s and 2.0<i>μ</i>s pulse duration). Chamber signals were air-density- and polarity-corrected as a function of DPP (0.1 Gy-6.5 Gy) at different pressures. The actual DPP was determined and monitored using a calibrated PTW flashDiamond and a current transformer, respectively. Experimental CCEs were compared with numerical calculations, describing the charge transport inside ICs, including free electrons and electric field distortion.<i>Main results.</i>The CCE decreased with increasing pressure (-6%/100 hPa and -8%/100 hPa for the Advanced Markus, and Roos IC, respectively) and followed 'logistic functions' with DPP. The CCE₀at a given pressure<i>P</i>₀can be obtained from a CCE₁at different pressure P₁(scaling rule) as: CCE₀(<i>P</i>₀,DPP) = CCE₁(<i>P</i>₁,DPP*(<i>P</i>₀/<i>P</i>₁)²). Our theoretical model predicted the relative variation of the CCE with pressure, with residuals <3%. The effect can be corrected using the scaling rule even at small changes in pressure (∼35 hPa), which can cause a 2% deviation on the CCE, without adding significant CCE uncertainty (∼0.1%) for a DPP up to 6.5 Gy per pulse.<i>Significance.</i>This work proposes a scaling rule to correct for recombination losses dependent on atmospheric pressure in ultra-high DPP electron beams. The proposed scaling rule provides a simple, low-uncertainty correction that can be applied to empirically predetermined CCE functions, improving the accuracy of commercial IC-based dosimetry in ultra-high DPP electron beams.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147491482","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":"Confounding factors in the generalisation of synthetic CT to diagnostic spine MRI.","authors":"Ryan Pollitt, Tom P C Schlösser, Marijn van Stralen, Edwin H G Oei, Joost P H J Rutges, L Wilbert Bartels, Peter R Seevinck","doi":"10.1088/1361-6560/ae5eb9","DOIUrl":"10.1088/1361-6560/ae5eb9","url":null,"abstract":"<p><p><i>Objective</i>. Generalisation of synthetic CT (sCT) generation to diagnostic MRI data in the spine faces many challenges, particularly when aiming at accurate visualisation of pathologies for orthopaedic treatment management. In this study we assessed the effect of potential confounding factors on the performance of sCT generation from diagnostic MR images.<i>Approach</i>. Paired spinal diagnostic MR and CT scans from two centres (51 patients) were collected retrospectively, spanning multiple spinal pathologies and regions. Each patient's dataset contained 3D T1-/T2-weighted and 2D T1- and T2-weighted scans. 3D U-Nets were trained per centre on the 3D MR data and CT using a mean absolute error (MAE) loss. The performance was assessed on sCTs derived from original 2D and 3D data and from simulated 2D data with varying orientation (axial/coronal/sagittal) and slice spacing (1.1-6.6 mm). The sCTs were compared to the CTs using MAE, peak signal-to-noise ratio, and dice similarity coefficient.<i>Main results</i>. We identified MR resolution, orientation and contrast as confounding factors for sCT generation from diagnostic MR. Additionally, CT noise and CT-to-MR registration were identified to influence the performance evaluation. Performance degraded with larger slice spacings and significant differences between orientations occurred more often at larger slice spacings. The effect of slice spacing and orientation was stronger inside the vertebrae than outside. While the networks performed better overall on contrasts more similar to the training contrasts, they demonstrated promising performance when trained on 3D T2w images and tested on 2D T1w images. Finally, higher CT noise levels resulted in worse sCT performance metrics, reflected in a significant correlation between noise and MAE.<i>Significance</i>. This work demonstrates that generalisation of sCT generation to diagnostic spine MRI data is hampered primarily by MRI acquisition related aspects including MR resolution, orientation and contrast, which provides guidance for future work on generalisable sCT from diagnostic MR.</p>","PeriodicalId":20185,"journal":{"name":"Physics in medicine and biology","volume":" ","pages":""},"PeriodicalIF":3.4,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147675922","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}