Medical physics最新文献

{"title":"A 3D deep learning model based on MRI for predicting lymphovascular invasion in rectal cancer.","authors":"Tangjuan Wang, Chuanyu Chen, Chang Liu, Shaopeng Li, Peng Wang, Dawei Yin, Ying Liu","doi":"10.1002/mp.17882","DOIUrl":"https://doi.org/10.1002/mp.17882","url":null,"abstract":"<p><strong>Background: </strong>The assessment of lymphovascular invasion (LVI) is crucial in the management of rectal cancer; However, accurately evaluating LVI preoperatively using imaging remains challenging. Recent advances in radiomics have created opportunities for developing more accurate diagnostic tools.</p><p><strong>Purpose: </strong>This study aimed to develop and validate a deep learning model for predicting LVI in rectal cancer patients using preoperative MR imaging.</p><p><strong>Methods: </strong>These cases were randomly divided into a training cohort (n = 233) and an validation cohort (n = 101) at a ratio of 7:3. Based on the pathological reports, the patients were classified into positive and negative groups according to their LVI status. Based on the preoperative MRI T2WI axial images, the regions of interest (ROI) were defined from the tumor itself and the edges of the tumor extending outward by 5 pixels, 10 pixels, 15 pixels, and 20 pixels. The 2D and 3D deep learning features were extracted using the DenseNet121 architecture, and the deep learning models were constructed, including a total of ten models: GTV (the tumor itself), GPTV5 (the tumor itself and the tumor extending outward by 5 pixels), GPTV10, GPTV15, and GPTV20. To assess model performance, we utilized the area under the curve (AUC) and conducted DeLong test to compare different models, aiming to identify the optimal model for predicting LVI in rectal cancer.</p><p><strong>Results: </strong>In the 2D deep learning model group, the 2D GPTV10 model demonstrated superior performance with an AUC of 0.891 (95% confidence interval [CI] 0.850-0.933) in the training cohort and an AUC of 0.841 (95% CI 0.767-0.915) in the validation cohort. The difference in AUC between this model and other 2D models was not statistically significant based on DeLong test (p > 0.05); In the group of 3D deep learning models, the 3D GPTV10 model had the highest AUC, with a training cohort AUC of 0.961 (95% CI 0.940-0.982) and a validation cohort AUC of 0.928 (95% CI 0.881-0.976). DeLong test demonstrated that the performance of the 3D GPTV10 model surpassed other 3D models as well as the 2D GPTV10 model (p < 0.05).</p><p><strong>Conclusion: </strong>The study developed a deep learning model, namely 3D GPTV10, utilizing preoperative MRI data to accurately predict the presence of LVI in rectal cancer patients. By training on the tumor itself and its surrounding margin 10 pixels as the region of interest, this model achieved superior performance compared to other deep learning models. These findings have significant implications for clinicians in formulating personalized treatment plans for rectal cancer patients.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144113176","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":"Fusing radiomics and deep learning features for automated classification of multi-type pulmonary nodule.","authors":"Lingyan Du, Guozhi Tang, Yue Che, Shihai Ling, Xin Chen, Xingliang Pan","doi":"10.1002/mp.17901","DOIUrl":"https://doi.org/10.1002/mp.17901","url":null,"abstract":"<p><strong>Background: </strong>The accurate classification of lung nodules is critical to achieving personalized lung cancer treatment and prognosis prediction. The treatment options for lung cancer and the prognosis of patients are closely related to the type of lung nodules, but there are many types of lung nodules, and the distinctions between certain types are subtle, making accurate classification based on traditional medical imaging technology and doctor experience challenging.</p><p><strong>Purpose: </strong>In this study, a novel method was used to analyze quantitative features in CT images using CT radiomics to reveal the characteristics of pulmonary nodules, and then feature fusion was used to integrate radiomics features and deep learning features to improve the accuracy of classification.</p><p><strong>Methods: </strong>This paper proposes a fusion feature pulmonary nodule classification method that fuses radiomics features with deep learning neural network features, aiming to automatically classify different types of pulmonary nodules (such as Malignancy, Calcification, Spiculation, Lobulation, Margin, and Texture). By introducing the Discriminant Correlation Analysis feature fusion algorithm, the method maximizes the complementarity between the two types of features and the differences between different classes. This ensures interaction between the information, effectively utilizing the complementary characteristics of the features. The LIDC-IDRI dataset is used for training, and the fusion feature model has been validated for its advantages and effectiveness in classifying multiple types of pulmonary nodules.</p><p><strong>Results: </strong>The experimental results show that the fusion feature model outperforms the single-feature model in all classification tasks. The AUCs for the tasks of classifying Calcification, Lobulation, Margin, Spiculation, Texture, and Malignancy reached 0.9663, 0.8113, 0.8815, 0.8140, 0.9010, and 0.9316, respectively. In tasks such as nodule calcification and texture classification, the fusion feature model significantly improved the recognition ability of minority classes.</p><p><strong>Conclusions: </strong>The fusion of radiomics features and deep learning neural network features can effectively enhance the overall performance of pulmonary nodule classification models while also improving the recognition of minority classes when there is a significant class imbalance.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144112116","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":"Thymoma habitat segmentation and risk prediction model using CT imaging and K-means clustering.","authors":"Zhu Liang, Jiamin Li, Shuyan He, Siyuan Li, Runzhi Cai, Chunyuan Chen, Yan Zhang, Biao Deng, Yanxia Wu","doi":"10.1002/mp.17892","DOIUrl":"https://doi.org/10.1002/mp.17892","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Thymomas, though rare, present a wide range of clinical behaviors, from indolent to aggressive forms, making accurate risk stratification crucial for treatment planning. Traditional methods such as histopathology and radiological assessments often lack the ability to capture tumor heterogeneity, which can impact prognosis. Radiomics, combined with machine learning, provides a method to extract and analyze quantitative imaging features, offering the potential to improve tumor classification and risk prediction. By segmenting tumors into distinct habitat zones, it becomes possible to assess intratumoral heterogeneity more effectively. This study employs radiomics and machine learning techniques to enhance thymoma risk prediction, aiming to improve diagnostic consistency and reduce variability in radiologists' assessments.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Objective: &lt;/strong&gt;This study aims to identify different habitat zones within thymomas through CT imaging feature analysis and to establish a predictive model to differentiate between high and low-risk thymomas. Additionally, the study explores how this model can assist radiologists.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;We obtained CT imaging data from 133 patients with thymoma who were treated at the Affiliated Hospital of Guangdong Medical University from 2015 to 2023. Images from the plain scan phase, venous phase, arterial phase, and their differential images (subtracted images) were used. Tumor regions were segmented into three habitat zones using K-Means clustering. Imaging features from each habitat zone were extracted using the PyRadiomics (van Griethuysen, 2017) library. The 28 most distinguishing features were selected through Mann-Whitney U tests (Mann, 1947) and Spearman's correlation analysis (Spearman, 1904). Five predictive models were built using the same machine learning algorithm (Support Vector Machine [SVM]): Habitat1, Habitat2, Habitat3 (trained on features from individual tumor habitat regions), Habitat All (trained on combined features from all regions), and Intra (trained on intratumoral features), and their performances were evaluated for comparison. The models' diagnostic outcomes were compared with the diagnoses of four radiologists (two junior and two experienced physicians).&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;The AUC (area under curve) for habitat zone 1 was 0.818, for habitat zone 2 was 0.732, and for habitat zone 3 was 0.763. The comprehensive model, which combined data from all habitat zones, achieved an AUC of 0.960, outperforming the model based on traditional radiomic features (AUC of 0.720). The model significantly improved the diagnostic accuracy of all four radiologists. The AUCs for junior radiologists 1 and 2 increased from 0.747 and 0.775 to 0.932 and 0.972, respectively, while for experienced radiologists 1 and 2, the AUCs increased from 0.932 and 0.859 to 0.977 and 0.972, respectively.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusion: &lt;/strong&gt;This study successfully identi","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096515","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":"Transformer model based on Sonazoid contrast-enhanced ultrasound for microvascular invasion prediction in hepatocellular carcinoma.","authors":"Qiong Qin, Jinshu Pang, Jingdan Li, Ruizhi Gao, Rong Wen, Yuquan Wu, Li Liang, Qiao Que, Changwen Liu, Jinbo Peng, Yun Lv, Yun He, Peng Lin, Hong Yang","doi":"10.1002/mp.17895","DOIUrl":"https://doi.org/10.1002/mp.17895","url":null,"abstract":"<p><strong>Background: </strong>Microvascular invasion (MVI) is strongly associated with the prognosis of patients with hepatocellular carcinoma (HCC).</p><p><strong>Purpose: </strong>To evaluate the value of Transformer models with Sonazoid contrast-enhanced ultrasound (CEUS) in the preoperative prediction of MVI.</p><p><strong>Methods: </strong>This retrospective study included 164 HCC patients. Deep learning features and radiomic features were extracted from arterial and Kupffer phase images, alongside the collection of clinicopathological parameters. Normality was assessed using the Shapiro-Wilk test. The Mann‒Whitney U-test and least absolute shrinkage and selection operator algorithm were applied to screen features. Transformer, radiomic, and clinical prediction models for MVI were constructed with logistic regression. Repeated random splits followed a 7:3 ratio, with model performance evaluated over 50 iterations. The area under the receiver operating characteristic curve (AUC, 95% confidence interval [CI]), sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), decision curve, and calibration curve were used to evaluate the performance of the models. The DeLong test was applied to compare performance between models. The Bonferroni method was used to control type I error rates arising from multiple comparisons. A two-sided p-value of < 0.05 was considered statistically significant.</p><p><strong>Results: </strong>In the training set, the diagnostic performance of the arterial-phase Transformer (AT) and Kupffer-phase Transformer (KT) models were better than that of the radiomic and clinical (Clin) models (p < 0.0001). In the validation set, both the AT and KT models outperformed the radiomic and Clin models in terms of diagnostic performance (p < 0.05). The AUC (95% CI) for the AT model was 0.821 (0.72-0.925) with an accuracy of 80.0%, and the KT model was 0.859 (0.766-0.977) with an accuracy of 70.0%. Logistic regression analysis indicated that tumor size (p = 0.016) and alpha-fetoprotein (AFP) (p = 0.046) were independent predictors of MVI.</p><p><strong>Conclusions: </strong>Transformer models using Sonazoid CEUS have potential for effectively identifying MVI-positive patients preoperatively.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096516","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":"Novel ¹H/¹⁹F double-tuned coil using an asymmetrical butterfly coil.","authors":"Suk-Min Hong, Chang-Hoon Choi, Jörg Felder, N Jon Shah","doi":"10.1002/mp.17890","DOIUrl":"https://doi.org/10.1002/mp.17890","url":null,"abstract":"<p><strong>Background: </strong>Fluorine-19 (¹⁹F) magnetic resonance imaging (MRI) is a non-invasive imaging tool for the targeted application of fluorinated agents, such as cell tracking, and for the demonstration of oximetry. However, as the SNR of ¹⁹F is significantly weaker than that of proton (¹H) imaging, the ¹⁹F coil must be combined with ¹H coils for anatomical co-registration and B₀ shimming. This is difficult due to the strong coupling between the coils when they are in proximity, and is problematic since the Larmor frequency of ¹⁹F is 94% that of ¹H, further increasing the potential for coupling between the ¹H and ¹⁹F elements.</p><p><strong>Purpose: </strong>Conventional double-tuned coil methods tend to generate loss compared to single-tuned reference coils. The asymmetrical butterfly coil has a split resonance peak, which can cover frequencies of ¹H and ¹⁹F without losses arising from lossy traps or switching circuits. In this study, the use of an asymmetrical butterfly coil was evaluated for ¹H/¹⁹F applications.</p><p><strong>Methods: </strong>To increase quadrature efficiency at both the ¹H and ¹⁹F frequencies, the left and right loops of the butterfly coil were tuned asymmetrically. The coil's tuning and performance were evaluated in simulations and MR measurements, and the results were compared to a dimension-matched single-tuned loop coil.</p><p><strong>Results: </strong>The split resonance peak of the asymmetrical butterfly coil successfully spanned the ¹⁹F to ¹H frequency. It operated with higher quadrature efficiency at both ¹H and ¹⁹F frequencies and demonstrated superior receive sensitivity and SNR compared to the dimension-matched single-tuned loop coil.</p><p><strong>Conclusions: </strong>The split resonance peak of the asymmetrical butterfly coil supported both ¹H and ¹⁹F frequencies, delivering a higher SNR than that of the single-tuned loop coil. Since the asymmetrical butterfly coil can cover ¹H and ¹⁹F frequencies without loss and provides higher efficiency than the reference single-tuned coil, it can be effectively utilized for ¹H/¹⁹F MRI applications.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096506","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":"The Cut-Sort-Group algorithm for efficient delivery of collimated step-and-shoot proton arc therapy.","authors":"Karsten K Wake, Laura C Bennett, Blake R Smith, Wesley S Culberson, Daniel E Hyer, Ryan T Flynn, Kaustubh A Patwardhan, Nicholas P Nelson, Patrick M Hill","doi":"10.1002/mp.17889","DOIUrl":"https://doi.org/10.1002/mp.17889","url":null,"abstract":"<p><strong>Background: </strong>Proton arc therapy is an emerging technology offering considerably more geometric flexibility than traditional multi-field treatments, thereby enhancing potential for more conformal proton treatments. The Dynamic Collimation System (DCS) offers energy-specific collimation during pencil beam scanning to further improve target conformity and reduce dose to normal tissues. Collimation with the DCS during arc delivery is referred to as dynamically collimated proton arc therapy (DC-PAT). The time required for energy switching, gantry movement during step-and-shoot arc delivery, and trimmer movement associated with dynamic collimation necessitates careful planning to create DC-PAT plans efficient enough to fit within a typical clinical workflow.</p><p><strong>Purpose: </strong>To demonstrate a post-processing algorithm to improve the delivery efficiency of DC-PAT plans while maintaining plan quality.</p><p><strong>Methods: </strong>A genetic optimizer was used to create baseline DC-PAT plans for three intracranial cases. These plans were then modified in the post-processing stage with the Cut-Sort-Group (CSG) algorithm. Specifically, each plan was modified through low-weight control point removal (\"Cut\"), a novel approach to energy layer sorting (\"Sort\"), and efficient DCS-trimmer reconfiguration (\"Group\"). The components of CSG were evaluated individually and in combination for changes in efficiency, plan quality, and robustness when compared to baseline plans.</p><p><strong>Results: </strong>After applying the CSG algorithm, the beam delivery time (BDT) for the three patients was reduced to between 10 and 14 min, more than 64% faster than the reference baseline plans. These efficiency gains were achieved with minimal impact on plan quality. The dose coverage to the PTV of the CSG-derived plans was comparable to the baseline plans for each patient, with the PTV D_2% remaining under 10% of the prescription and a Homogeneity Index (HI) ranging from 0.09 and 0.12. Dose to non-target structures and overall plan robustness was also minimally impacted by the implementation of the CSG algorithm.</p><p><strong>Conclusions: </strong>The CSG algorithm demonstrates a relatively simple approach to modifying step-and-shoot proton arc therapy plans to be more efficient in the post-processing stage regardless of the treatment planning system or optimization algorithm used to generate the initial plans and with minimal impact on plan quality. The overall BDT was reduced to just over 10 min, approaching plans produced using other advanced optimization algorithms in previous investigations, and fast enough for potential clinical implementation.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096513","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":"Distortion-free steady-state diffusion-weighted imaging with magnetic resonance fingerprinting.","authors":"Yiang Wang, Yingying Lin, Di Cui, Edward S K Hui, Elaine Y P Lee, Peng Cao","doi":"10.1002/mp.17894","DOIUrl":"https://doi.org/10.1002/mp.17894","url":null,"abstract":"<p><strong>Background: </strong>Magnetic resonance fingerprinting (MRF) could provide joint T1, T2, and proton density mapping. Measuring diffusion encoding using the MRF framework is promising, given its capacity to generate self-aligned quantitative maps and contrast-weighted images from a single scan. It could avoid potential errors that arise from the registration of multiple MRI images and reduce the total scan time. However, the application of a strong diffusion gradient on the MRF sequence results in phase inconsistency between acquisitions, which could corrupt the reconstructed images.</p><p><strong>Purpose: </strong>To propose a distortion-free diffusion-weighted imaging module for MRF (DWI-MRF) method using a self-navigated subspace reconstruction on k-space data obtained from a dual-density spiral trajectory.</p><p><strong>Methods: </strong>The proposed sequence consisted of two segments: inversion prepared steady-state free precession MRF for the first 800 time points and diffusion-weighted imaging (DWI) with two nominal b-values of 0 and 800 s/mm² for the following 200 time points. The temporal basis was acquired from the densely sampled central k-space during reconstruction. The subspace reconstruction was applied to generate aliasing-free and high-resolution images at each time point. The cardiac gating was retrospectively performed on the high-resolution and dynamic DWI images. Our T1, T2, and apparent diffusion coefficient (ADC) results were compared to conventional methods on a phantom and two healthy volunteers.</p><p><strong>Results: </strong>Our method's T1, T2, and ADC values agreed reasonably with the reference values, with a slope of 0.88, 0.94, and 1.04 for T1, T2, and ADC, and an R² value of 0.97, 0.97, and 0.71, respectively. The T1, T2, and ADC maps from DWI-MRF exhibited pixel-by-pixel correspondence on phantom and in vivo (T1 and ADC: R²= 0.75 on phantom and 0.84 in vivo; T2 and ADC: R²= 0.79 and 0.83, respectively). Our method achieved high acquisition efficiency, requiring less than 20 s per slice.</p><p><strong>Conclusions: </strong>The proposed method was free of artifacts from cardiac pulsation and generated pixel-wise correspondent T1, T2, and ADC maps on both phantom and in vivo images.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096497","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":"Non-orthogonal kV imaging guided patient position verification in non-coplanar radiation therapy with dataset-free implicit neural representation.","authors":"Siqi Ye, Yizheng Chen, Siqi Wang, Lei Xing, Yu Gao","doi":"10.1002/mp.17885","DOIUrl":"https://doi.org/10.1002/mp.17885","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;Cone-beam CT (CBCT) is crucial for patient alignment and target verification in radiation therapy (RT). However, for non-coplanar beams, potential collisions between the treatment couch and the on-board imaging system limit the range that the gantry can be rotated. Limited-angle measurements are often insufficient to generate high-quality volumetric images for image-domain registration, therefore limiting the use of CBCT for position verification. An alternative to image-domain registration is to use a few 2D projections acquired by the onboard kV imager to register with the 3D planning CT for patient position verification, which is referred to as 2D-3D registration.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;The 2D-3D registration involves converting the 3D volume into a set of digitally reconstructed radiographs (DRRs) expected to be comparable to the acquired 2D projections. The domain gap between the generated DRRs and the acquired projections can happen due to the inaccurate geometry modeling in DRR generation and artifacts in the actual acquisitions. We aim to improve the efficiency and accuracy of the challenging 2D-3D registration problem in non-coplanar RT with limited-angle CBCT scans.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Method: &lt;/strong&gt;We designed an accelerated, dataset-free, and patient-specific 2D-3D registration framework based on an implicit neural representation (INR) network and a composite similarity measure. The INR network consists of a lightweight three-layer multilayer perception followed by average pooling to calculate rigid motion parameters, which are used to transform the original 3D volume to the moving position. The Radon transform and imaging specifications at the moving position are used to generate DRRs with higher accuracy. We designed a composite similarity measure consisting of pixel-wise intensity difference and gradient differences between the generated DRRs and acquired projections to further reduce the impact of their domain gap on registration accuracy. We evaluated the proposed method on both simulation data and real phantom data acquired from a Varian TrueBeam machine. Comparisons with a conventional non-deep-learning registration approach and ablation studies on the composite similarity measure were conducted to demonstrate the efficacy of the proposed method.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;In the simulation data experiments, two X-ray projections of a head-and-neck image with &lt;math&gt; &lt;semantics&gt;&lt;msup&gt;&lt;mn&gt;45&lt;/mn&gt; &lt;mo&gt;∘&lt;/mo&gt;&lt;/msup&gt; &lt;annotation&gt;${45}^circ$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; discrepancy were used for the registration. The accuracy of the registration results was evaluated on experiments set up at four different moving positions with ground-truth moving parameters. The proposed method achieved sub-millimeter accuracy in translations and sub-degree accuracy in rotations. In the phantom experiments, a head-and-neck phantom was scanned at three different positions involving couch translations and","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096503","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":"Megavoltage photon FLASH for preclinical experiments.","authors":"Edward R J F Taylor, Iain D C Tullis, Borivoj Vojnovic, Kristoffer Petersson","doi":"10.1002/mp.17891","DOIUrl":"https://doi.org/10.1002/mp.17891","url":null,"abstract":"<p><strong>Background: </strong>FLASH radiotherapy using megavoltage (MV) photon beams should enable greater therapeutic efficacy, target deep seated tumors, and provide insights into mechanisms within FLASH.</p><p><strong>Purpose: </strong>In this study, we aim to show how to facilitate ultra-high dose rates (FLASH) with MV photons over a field size of 12-15 mm, using a 6 MeV (nominal) preclinical electron linear accelerator (linac). Our intention is to utilize this setup to deliver FLASH with MV photons in future preclinical experiments.   METHODS: An electron linear accelerator operating at a pulse repetition frequency of 300 Hz, a tungsten target, and a beam hardening filter were used, in conjunction with beam tuning and source-to-surface distance (SSD) reduction. Depth dose curves, beam profiles, and average dose rates were determined using EBT-XD Gafchromic film, and an Advanced Markus ionization chamber was used to measure the photon charge output.</p><p><strong>Results: </strong>A 0.55 mm thick tungsten target, in combination with a 6 mm thick copper hardening filter were found to produce photon FLASH dose rates, with minimal electron contamination, delivering dose rates > 40 Gy/s over fields of 12-15 mm. Beam flatness and symmetry were comparable in horizontal and vertical planes.</p><p><strong>Conclusion: </strong>Ultra-high average dose rate beams have been achieved with MV photons for preclinical irradiation fields, enabling future preclinical FLASH radiation experiments.</p>","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144096499","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":"Characterization of a Commercial Ionization Chamber Array With Scanned Proton Beams for Applications in MRI-Guided Proton Therapy.","authors":"Benjamin Gebauer, Sebastian Gantz, Daniela Kunath, Aswin Hoffmann, Jörg Pawelke, Felix Horst","doi":"10.1002/mp.17875","DOIUrl":"https://doi.org/10.1002/mp.17875","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;The integration of MRI-guidance and proton therapy is a current research topic. Proton therapy with the patient being placed inside an in-beam MR scanner would require the presence of its static magnetic ( &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;B&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;$B_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; ) field to be taken into account in dose calculation and treatment planning. Therefore, dosimetric tools are needed to characterize dose distributions in presence of the &lt;math&gt; &lt;semantics&gt;&lt;msub&gt;&lt;mi&gt;B&lt;/mi&gt; &lt;mn&gt;0&lt;/mn&gt;&lt;/msub&gt; &lt;annotation&gt;$B_0$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; field of the MR scanner. Furthermore, patient-specific quality assurance (QA) and treatment plan verification measurements should also be performed within the magnetic field.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;In this work, the PTW Octavius 1500 &lt;math&gt; &lt;semantics&gt;&lt;msup&gt;&lt;mrow&gt;&lt;/mrow&gt; &lt;mrow&gt;&lt;mi&gt;M&lt;/mi&gt; &lt;mi&gt;R&lt;/mi&gt;&lt;/mrow&gt; &lt;/msup&gt; &lt;annotation&gt;$^{MR}$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; ionization chamber array was characterized experimentally and tested for its suitability as a dosimetric tool for beam characterization and QA in MRI-guided proton therapy.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Methods: &lt;/strong&gt;The dose rate response, response homogeneity and effective measurement depth of the detector were determined in experiments with scanned proton beams delivered by a horizontal beamline at OncoRay, Dresden. A patient-specific QA test including gamma analysis was performed for a realistic proton patient treatment plan at two different distances from the beam nozzle. In addition, experiments were performed in a &lt;math&gt; &lt;semantics&gt;&lt;mrow&gt;&lt;mn&gt;0.32&lt;/mn&gt; &lt;mspace&gt;&lt;/mspace&gt; &lt;mi&gt;T&lt;/mi&gt;&lt;/mrow&gt; &lt;annotation&gt;$0.32 mathrm{T}$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; in-beam MR scanner. These included measurements of square reference scanning patterns at different proton energies as well as measurements of a two-field patient treatment plan at different water equivalent depths.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;The dose rate response was found to be linear up to &lt;math&gt; &lt;semantics&gt;&lt;mrow&gt;&lt;mn&gt;80&lt;/mn&gt; &lt;mspace&gt;&lt;/mspace&gt; &lt;mtext&gt;Gy/min&lt;/mtext&gt;&lt;/mrow&gt; &lt;annotation&gt;$80 text{Gy/min}$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; . The effective measurement depth was determined to be &lt;math&gt; &lt;semantics&gt;&lt;mrow&gt;&lt;mn&gt;8.1&lt;/mn&gt; &lt;mo&gt;±&lt;/mo&gt; &lt;mn&gt;0.2&lt;/mn&gt; &lt;mspace&gt;&lt;/mspace&gt; &lt;mi&gt;mm&lt;/mi&gt;&lt;/mrow&gt; &lt;annotation&gt;$8.1 pm 0.2 mathrm{mm}$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; . The response homogeneity was found to be suitable for the verification of proton treatment plans. The patient-specific QA test without magnetic field was satisfactory and also the measurements inside the &lt;math&gt; &lt;semantics&gt;&lt;mrow&gt;&lt;mn&gt;0.32&lt;/mn&gt; &lt;mspace&gt;&lt;/mspace&gt; &lt;mi&gt;T&lt;/mi&gt;&lt;/mrow&gt; &lt;annotation&gt;$0.32 mathrm{T}$&lt;/annotation&gt;&lt;/semantics&gt; &lt;/math&gt; in-beam MR scanner provided reasonable results. Their comparison allowed an assessment of the magnetic field effects on the dose distributions.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Conclusions: &lt;/strong&gt;Concluding from these tests, the Octavius 1500 &lt;math&gt; &lt;semant","PeriodicalId":94136,"journal":{"name":"Medical physics","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144016011","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}