Jack Humphreys, Christopher White, Florian Mentzel, David Bolst, Jason Paino, Ah Chung Tsoi, Lorenzo Arsini, Franco Scarselli, Carlo Mancini-Terracciano, Anatoly Rosenfeld, Moeava Tehei, Stéphanie Corde, Michael Lerch, Susanna Guatelli, Markus Hagenbuchner
{"title":"Point cloud dosimetry framework for preclinical microbeam radiation therapy.","authors":"Jack Humphreys, Christopher White, Florian Mentzel, David Bolst, Jason Paino, Ah Chung Tsoi, Lorenzo Arsini, Franco Scarselli, Carlo Mancini-Terracciano, Anatoly Rosenfeld, Moeava Tehei, Stéphanie Corde, Michael Lerch, Susanna Guatelli, Markus Hagenbuchner","doi":"10.1016/j.ejmp.2025.105198","DOIUrl":"https://doi.org/10.1016/j.ejmp.2025.105198","url":null,"abstract":"<p><strong>Background: </strong>Microbeam Radiation Therapy (MRT) is an emerging radiotherapy technique which is currently at the research stage. In order to further progress from research, toward clinical deployment, it is essential to develop a reliable and accurate dose engine such as Monte Carlo (MC) simulations.</p><p><strong>Purpose: </strong>MC execution times are far too long for practical, clinical applications. In previous studies, we used a 3D U-Net, trained with Geant4 MC simulations, to calculate the dose in digital rat phantoms. This choice of model imposes significant scalability challenges in the case of larger geometries. Casting MRT dose prediction as a 3D point cloud regression problem is a flexible and extensible solution to overcoming many of these hurdles. This problem formulation requires the use of point-based models which are unproven for the task of MRT dose prediction.</p><p><strong>Methods: </strong>In order to assay the viability of this family of models on this task, the SphereFormer is trained to accurately replicate the gold standard MC dosimetry on uniform voxel grids.</p><p><strong>Results: </strong>Furthermore, the benefits of the scalability of this method are demonstrated by the utilisation of out-of-field information to significantly improve over the existing state-of-the-art results on valley dose prediction, being accurate to within 3% for at least 84.1% of voxels compared with 78.2% for the baseline.</p><p><strong>Conclusion: </strong>This paper serves as a proof-of-concept study for the application of 3D point cloud methods to MRT dose prediction and marks the first time such a method have been applied to dosimetry in general.</p>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"105198"},"PeriodicalIF":2.7,"publicationDate":"2025-10-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145350170","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}
Lei Wang , Qiyong Ai , Weitian Chen , Zongyou Cai , Ziqiang Yu , Jill Abrigo , Tiffany Y So
{"title":"Effect of gadolinium-based contrast media on intravoxel incoherent motion (IVIM) MRI parameters in brain imaging","authors":"Lei Wang , Qiyong Ai , Weitian Chen , Zongyou Cai , Ziqiang Yu , Jill Abrigo , Tiffany Y So","doi":"10.1016/j.ejmp.2025.105172","DOIUrl":"10.1016/j.ejmp.2025.105172","url":null,"abstract":"<div><h3>Purpose</h3><div>Current intravoxel incoherent motion (IVIM) MRI protocols typically involve acquiring IVIM images before the injection of intravenous contrast media. The effect of contrast on IVIM model parameters remains unclear. This study aimed to assess the effect of intravenous gadolinium-based contrast media on IVIM parameters (f, D*, and D) in brain MRI.</div></div><div><h3>Materials and methods</h3><div>Thirty-one participants in the experimental contrast group (age 38.8 ± 8.81, 22 females, 9 males) and 31 participants in the control group (age 34.7 ± 9.68, 17 females, 14 males) were prospectively recruited. Pre-contrast, immediate post-contrast, and delayed post-contrast IVIM with 14b values following administration of gadoteric acid at 0.1 ml per kg body mass were performed in a 3.0 T MRI scanner. The IVIM parameters, and the nominal signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated and compared between scans using repeated measures analysis of variance and paired t-tests.</div></div><div><h3>Results</h3><div>No baseline differences in IVIM parameters were observed in the control group across the three scans (p = 0.077 to 0.866). In the contrast group, there were reductions in regional f, D*, D, and ADC values between the pre-contrast to post-contrast scan, followed by a return towards baseline in the delayed phase (p < 0.050). SNR (p = 0.056 to 0.513) and CNR (p = 0.359 to 0.999) were not significantly different before and after contrast.</div></div><div><h3>Conclusion</h3><div>Potential alterations in f, D*, D, and ADC are observed following administration of gadolinium-based contrast media. These changes may require consideration when performing IVIM after other contrast sequences.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105172"},"PeriodicalIF":2.7,"publicationDate":"2025-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145322449","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}
Jia Ding Wong , Aik Hao Ng , Li Kuo Tan , Jeannie Hsiu Ding Wong
{"title":"Timing or criteria to trigger adaptive radiotherapy (ART) for nasopharyngeal carcinoma (NPC): A systematic review","authors":"Jia Ding Wong , Aik Hao Ng , Li Kuo Tan , Jeannie Hsiu Ding Wong","doi":"10.1016/j.ejmp.2025.105203","DOIUrl":"10.1016/j.ejmp.2025.105203","url":null,"abstract":"<div><h3>Introduction</h3><div>Nasopharyngeal carcinoma (NPC), prevalent in Southern China and Southeast Asia, is primarily treated with radiation therapy (RT) and concurrent chemoradiotherapy (CCRT). Due to its anatomical complexity and changes during treatment, adaptive radiotherapy (ART) is employed to adjust plans mid-course. However, ART is time-consuming and labor intensive. This systematic review aims to summarize and analyze the timepoints and criteria used to trigger replanning in NPC patients undergoing ART.</div></div><div><h3>Methods</h3><div>A systematic search was conducted in PubMed and Scopus to identify studies reporting on the timing or criteria for replanning in NPC patients treated with ART. Eligible studies were screened, and relevant data were extracted.</div></div><div><h3>Results</h3><div>51 studies met the inclusion criteria. Among these, 42 recommended a single replanning session, 3 suggested one or two sessions depending on individual factors, 4 proposed two sessions, and 2 recommended three. Week 5 was the most frequently reported timepoint for initiating ART, followed by weeks 4 and 3. Tumor volume reduction was a common criterion. Changes in parotid gland anatomy and dosimetric parameters were frequently cited due to their relevance as organs at risk. Body weight loss was a triggering factor in eleven studies, while four studies highlighted ill-fitted masks as indicators for ART initiation.</div></div><div><h3>Conclusion</h3><div>There is notable variability in the timing or criteria for initiating ART in NPC patients. Prospective studies are necessary to develop evidence-based ART guidelines and protocols.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105203"},"PeriodicalIF":2.7,"publicationDate":"2025-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145322450","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":"Compensated biological effective dose in extended radiotherapy course via a time-modified linear quadratic model for biological adaptive radiotherapy","authors":"Takuya Wada , Daisuke Kawahara , Akito S Koganezawa , Nobuki Imano , Ikuno Nishibuchi , Yuji Murakami","doi":"10.1016/j.ejmp.2025.105202","DOIUrl":"10.1016/j.ejmp.2025.105202","url":null,"abstract":"<div><h3>Purpose</h3><div>Interruptions during radiotherapy, such as holidays or equipment issues, can decrease the biological effective dose (BED) and compromise treatment outcomes. Although the decrease in BED has been previously investigated, a detailed evaluation of three-dimensional (3D) BED distribution and a direct compensation method integrated with treatment planning systems is missing. We aimed to address the decrease in BED due to unexpected interruptions during radiotherapy by developing a workflow to generate deliverable treatment plans based on a modified linear-quadratic (mLQ) model.</div></div><div><h3>Methods</h3><div>We analyzed 81 head-and-neck cases treated with volumetric-modulated arc therapy, including 25 single prescription (SRx) and 56 simultaneous integrated boost (SIB) cases. The decrease in BED was calculated using the mLQ model based on treatment duration (TD) from the record & verify system. A base plan was generated using compensating factors (CFs) derived from the mLQ model, and a composite plan was created using existing optimization settings. Compensated and original plans were compared using biological dose-volume histogram (BDVH) parameters and homogeneity index (HI).</div></div><div><h3>Results</h3><div>The relative BED was 95.3% <span><math><mo>±</mo></math></span> 2.9% and 93.9% <span><math><mo>±</mo></math></span> 2.2% of the OP with TD from the R&V system. The standard deviation (SD) of the difference in the BED covering 98%, 50%, and 2% of the PTV of OP and CCP was approximately 2%. The difference in the HI was 0.6% <span><math><mo>±</mo></math></span> 1.5%.</div></div><div><h3>Conclusion</h3><div>We developed a workflow to generate clinically deliverable plans that compensate for the decrease in BED due to treatment interruptions using the mLQ model. This approach may facilitate rapid compensation planning in clinical radiotherapy settings.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105202"},"PeriodicalIF":2.7,"publicationDate":"2025-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145318949","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}
Tarik El Ghalbzouri , Tarek El Bardouni , Jaafar El Bakkali , Mohamed Drissi El-Bouzaidi , Hicham Satti , Randa Yerrou , Soumaya Oulad-Belayachi
{"title":"Updating internal dose coefficients of radiation therapy alpha-emitting radionuclides, Ac-225, At-211, Bi-213, Th-227, and Ra-223 for new reference male and female phantoms using DoseCalcs","authors":"Tarik El Ghalbzouri , Tarek El Bardouni , Jaafar El Bakkali , Mohamed Drissi El-Bouzaidi , Hicham Satti , Randa Yerrou , Soumaya Oulad-Belayachi","doi":"10.1016/j.ejmp.2025.105205","DOIUrl":"10.1016/j.ejmp.2025.105205","url":null,"abstract":"<div><h3>Purpose:</h3><div>Radiopharmaceuticals, particularly alpha-emitting radionuclides, have garnered significant attention for their potential in targeted cancer therapy because they can deliver lethal doses of radiation over a short range. Therefore, accurately estimating the absorbed dose within the body is critical for optimizing treatment effectiveness and ensuring patient safety. While internal dosimetry coefficients of alpha-emitting radionuclides, including Ac-225, At-211, Bi-213, Th-227, and Ra-223, and their progeny, are already available, they have not yet been computed for the latest generations of phantoms.</div></div><div><h3>Methods:</h3><div>In this context, we performed Monte Carlo simulations using DoseCalcs code with ICRP Publication 145 new mesh-type phantoms to calculate the S-values of these radionuclides-emitted radiation from different source organs/tissues in adult male and female models.</div></div><div><h3>Results:</h3><div>The self-irradiation S-values for these radionuclides were significantly elevated, as were the cross-irradiation S-values for contents-to-wall cases. When comparing the S-values calculated in mesh-type phantoms with those derived from OpenDose in voxelized phantoms, the values were generally consistent for most source-to-target cases with ratios around 1. However, ratios between 3–5 were noted for UBW<span><math><mo>←</mo></math></span>UBCs, StW<span><math><mo>←</mo></math></span>Liver and StW<span><math><mo>←</mo></math></span>Brain, and ratios around 2 for UBW<span><math><mo>←</mo></math></span>Liver and Gonads<span><math><mo>←</mo></math></span>Brain. This was especially evident in cases involving walls as target and/or separated by large distances. These differences were attributed to refined wall modeling, smooth organ interfaces, and modified cortical bone densities, leading to improved microscopic energy-deposition estimates.</div></div><div><h3>Conclusion:</h3><div>Our study provides a comprehensive, updated dataset of S-values of alpha-emitting radionuclides in the mesh-type phantoms; hence, they could be used for comparison and adult-patient dose estimation.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105205"},"PeriodicalIF":2.7,"publicationDate":"2025-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145318983","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}
William Donahue, Shu Xing, Lauren Smith, Paola Godoy Scripes, Eric Aliotta, Jim Mechalakos, Lakshmi Santanam, Neelam Tyagi, Seng Boh Lim
{"title":"Characterizing the motion management system accuracy on a 1.5T MR-Linac","authors":"William Donahue, Shu Xing, Lauren Smith, Paola Godoy Scripes, Eric Aliotta, Jim Mechalakos, Lakshmi Santanam, Neelam Tyagi, Seng Boh Lim","doi":"10.1016/j.ejmp.2025.105189","DOIUrl":"10.1016/j.ejmp.2025.105189","url":null,"abstract":"<div><h3>Introduction</h3><div>Motion management techniques have enabled drastic reductions in treatment volume allowing for dose escalation. The superior contrast of MRI combined with the ability to perform cine imaging during treatment enable surrogate-less tracking for gating. The purpose of this work is to commission motion management system with gating capabilities on the 1.5T MR-Linac and to characterize system performance.</div></div><div><h3>Methods</h3><div>Linear accelerator stability during gating was measured for a wide range of duty cycles (16 % − 100 %). Motion tracking accuracy and gating latency were characterized for varying target sizes and breathing rates, and with and without adaptive motion prediction using a commercially available motion phantom. Clinical treatment plans were created for lung, prostate, pancreas, and liver sites, using varied gating strategies and tumor motion amplitudes. Film measurements were performed for each plan and compared to either the treatment planning system (TPS) or a reference film delivery. Finally, end-to-end testing was performed with the motion phantom and film to quantify the delivery accuracy of the system for exception and free-breathing gating strategies.</div></div><div><h3>Results</h3><div>Beam characteristics were stable for all duty cycles analyzed, with all measurements within 0.6 % of the ungated baseline. Tracking accuracy showed a strong dependence on breathing rate. The system had difficulties tracking a 1 cm target due to through plane motion caused by the helical path of the phantom. Film comparisons for the clinical plans demonstrated average dose differences within 1.8 % and gamma passing rates (3 %/2 mm) >75 % within the 50 % isodose line. The cases with the worst gamma passing rates corresponded to exhale gating strategy where a systematic shift in dose was observed due to tumor motion interplay with the gating envelope. End-to-end testing showed excellent dose agreement (<3 % average dose difference) and a localization accuracy of <1 mm for targets with non-periodic motion and <2 mm for moving targets.</div></div><div><h3>Conclusions</h3><div>The motion management system on the Elekta Unity provides advanced capabilities for treating moving targets. The results of this work support the ability of the system to deliver accurate radiotherapy plans with reasonable uncertainties. However, as demonstrated by exhale gating results, better understanding of the clinical impacts of these new functionalities may be warranted.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105189"},"PeriodicalIF":2.7,"publicationDate":"2025-10-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145304548","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":"Evaluation of three radiochromic film types in x-ray minibeam radiation therapy pre-clinical dosimetry","authors":"Marios Sotiropoulos , Alfredo Fernandez-Rodriguez , Yolanda Prezado","doi":"10.1016/j.ejmp.2025.105195","DOIUrl":"10.1016/j.ejmp.2025.105195","url":null,"abstract":"<div><h3>Background</h3><div>Minibeam radiation therapy (MBRT) is an unconventional experimental technique that uses a strong spatial modulation of the dose. The narrow beam widths employed (0.5 to 1.0 mm) require the use of high-spatial resolution detectors. One of the most frequently employed is radiochromic films.</div></div><div><h3>Purpose</h3><div>This work aims at providing a systemic and thorough assessment of the performance of different radiochromic films and their calibration procedure for X-ray MBRT.</div></div><div><h3>Methods</h3><div>Three commonly used radiochromic films having different dynamic range and composition (GAFchromic EBT3 and EBT-XD (Ashland, Bridgewater, New Jersey, USA), and OrthoChromic OC-1 (OrthoChrome Inc, Hillsborough, New Jersey, USA) were evaluated. Different calibration procedures commonly employed in the literature were followed up. Then, the three different film types were irradiated with X-ray MBRT, and the impact of the different film types and calibration procedures were assessed by measuring peak and valley doses.</div></div><div><h3>Results</h3><div>Significant dosimetric differences, especially in the peak-to-valley dose ratio determination, may result from the use of different film types calibrated following different procedures. Additionally, the OC-1 films deemed not to be able provide the required spatial resolution to resolve beam of less than 1 mm FWHM.</div></div><div><h3>Conclusions</h3><div>The unconventional nature of MBRT requires unconventional dosimetry methods and procedures, which must be carefully evaluated. The combination of film type and calibration procedure employed is critical in minibeam radiation therapy, may lead to large uncertainties and should be attentively chosen.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105195"},"PeriodicalIF":2.7,"publicationDate":"2025-10-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145294455","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":"Reduction of respiratory motion artifacts in free-breathing abdominal MRI using strategic averaging of reassembled k-space data with Self-Modeled respiratory state sorting","authors":"Feng-Mao Chiu , Jyh-Wen Chai , Yu-Ting Fang , Yu-Chun Lo , Yao-Wen Liang , Yi-Ying Wu , Nan-Kuei Chen , You-Yin Chen","doi":"10.1016/j.ejmp.2025.105185","DOIUrl":"10.1016/j.ejmp.2025.105185","url":null,"abstract":"<div><h3>Objective</h3><div>Motion during MRI acquisition leads to varying phase errors in k-space, resulting in motion artifacts that degrade image quality. This study aimed to develop a novel reconstruction method called Strategic Averaging of Reassembled k-space Data (STREAK), which utilizes self-modeled respiratory signals to reduce motion artifacts in free-breathing abdominal MRI.</div></div><div><h3>Approach</h3><div>We compared the proposed STREAK method with conventional signal averaging (NSA) and free-breathing image acquisition. Three image groups were evaluated: free-breathing, NSA with three signal averages (NSA 3), and STREAK. Image quality was assessed using <em>structural similarity (SSIM)</em>, <em>peak signal-to-noise ratio (PSNR)</em>, and <em>artifact power (AP)</em>, along with subjective grading performed by experienced radiologists. Statistical analysis was conducted using the Mann–Whitney U and Dunn’s tests, with p-values less than 0.05 considered statistically significant.</div></div><div><h3>Results</h3><div>The STREAK group showed significantly improved SSIM, PSNR, and AP metrics in the liver (<em>p</em> < 0.05). Compared with free-breathing and NSA 3 images, STREAK significantly enhanced image quality in all objective and subjective assessments (<em>p</em> < 0.001). STREAK showed superior motion artifact reduction and image clarity, demonstrating its potential for enhanced MRI imaging quality compared to the NSA method. Inter-reader agreement among radiologists was above moderate (≥ 0.55).</div></div><div><h3>Conclusions</h3><div>STREAK, combining Cartesian sampling, sensitivity encoding, respiratory signal modeling, and strategic k-space reconstruction, significantly reduced motion artifacts and surpassed the NSA method, showing clinical potential for improved imaging quality.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105185"},"PeriodicalIF":2.7,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145271146","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}
Zhong Chen , Wangyao Li , Xinglei Shen , Ronald C. Chen , Yuting Lin , Hao Gao
{"title":"A few-shot u-net learning framework for fast and accurate three‐dimensional dose prediction in radiotherapy","authors":"Zhong Chen , Wangyao Li , Xinglei Shen , Ronald C. Chen , Yuting Lin , Hao Gao","doi":"10.1016/j.ejmp.2025.105184","DOIUrl":"10.1016/j.ejmp.2025.105184","url":null,"abstract":"<div><h3>Background</h3><div>Accurate and fast patient-specific dose prediction is crucial to improve the quality and efficiency of radiation treatment planning. Recently, deep learning-based approaches have achieved remarkable results in dose predictions across various cancer types. However, they typically require a large set of training examples to ensure robust generalization.</div></div><div><h3>Methods</h3><div>To the end, our proposed FS-UNet leverages the U-Net framework and enhances dose prediction performance from limited training data using the discriminative power of few-shot learning. To achieve this, our FS-UNet is decomposed into three seamless components: (1) a 3D U-Net is a convolutional neural network specifically designed for dose prediction; (2) a gradient-based method MAML with N-way K-shot settings is embedded in the dose regression problem, which is a meta-learning algorithm by updating model’s parameters based on a limited set of training samples; and (3) a Prototypical Network is utilized to perform an <em>l</em>-nearest centroid regression model to establish a feature metric generalizable for FS-UNet in dose prediction.</div></div><div><h3>Results</h3><div>The FS-UNet model was validated for prostate proton radiotherapy using data from 102 patients, demonstrating superior performance against four existing deep learning-based approaches with both large (81 prostate cases) and small (11 prostate cases) training datasets. Additionally, experiments on 29 liver cases and 30 pancreas cases showed that dose prediction performance improved with mixed-site dose prediction compared to single-site predictions, even in few-shot learning scenarios.</div></div><div><h3>Conclusion</h3><div>The FS-UNet demonstrated more accurate DVH and less MSE in dose prediction for prostate, liver, pancreas, and mixed treatment sites in radiotherapy.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105184"},"PeriodicalIF":2.7,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145271175","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}
Tanuj Puri , Tiziana Rancati , Petra Seibold , Adam Webb , Eliana Vasquez osorio , David Azria , Marie-Pierre Farcy-Jacquet , Jenny Chang-Claude , Alison Dunning , Maarten Lambrecht , Barbara Avuzzi , Dirk de Ruysscher , Elena Sperk , Ana Vega , Liv Veldeman , Barry Rosenstein , Sarah Kerns , Christopher Talbot , Alan McWilliam , Peter Hoskin , Marcel van Herk
{"title":"Sensitivity analysis of dose-response model parameters for the bladder in prostate cancer radiotherapy","authors":"Tanuj Puri , Tiziana Rancati , Petra Seibold , Adam Webb , Eliana Vasquez osorio , David Azria , Marie-Pierre Farcy-Jacquet , Jenny Chang-Claude , Alison Dunning , Maarten Lambrecht , Barbara Avuzzi , Dirk de Ruysscher , Elena Sperk , Ana Vega , Liv Veldeman , Barry Rosenstein , Sarah Kerns , Christopher Talbot , Alan McWilliam , Peter Hoskin , Marcel van Herk","doi":"10.1016/j.ejmp.2025.105178","DOIUrl":"10.1016/j.ejmp.2025.105178","url":null,"abstract":"<div><h3>Introduction</h3><div>This study evaluates how model parameter values affect dose–response maps (DRMs) in identifying high-risk bladder subregions associated with late urinary toxicities in prostate cancer patients post-radiotherapy.</div></div><div><h3>Methods</h3><div>Data from 1808 patients were analyzed for five late bladder toxicities. Baseline scores were subtracted from maximum toxicity at 12 and 24 months and dichotomized into grades ≥ 1 and ≥ 2. Bladders were segmented on computed tomography scans, and dose-surface maps (DSMs) were created on 91 × 90 voxel grids using spherical and cylindrical coordinates. Voxel doses were converted to equivalent dose in 2 Gy fractions (EQD2, α/β 1–3 Gy). Welch’s t and Mann–Whitney U equations were applied at each voxel location. Multiple comparisons were corrected via permutation testing (10–10000 iterations), and statistically significant voxels were identified using the 90th and 95th percentiles of Tmax/Umax. Sensitivity of parameters was assessed by varying one parameter at a time, with changes > 400 voxels (∼5% of 8190) classified as large and ≤ 400 as small.</div></div><div><h3>Results</h3><div>Urinary tract obstruction was the only toxicity significantly associated with bladder DSMs, focusing results on this outcome. After baseline adjustment and dichotomization, event/nonevent counts were 62/701 (grade≥1) and 21/742 (grade≥2; N = 763). DRM results showed large effects of toxicity grade threshold, coordinate system, statistical test equation, and Tmax/Umax thresholding. EQD2 α/β showed variable effects, large for cylindrical and small for spherical coordinates, while the number of permutations had only a small effect.</div></div><div><h3>Conclusions</h3><div>Parameter selection significantly influences high-risk subregion identification in DRMs, emphasizing the need for standardized parameter reporting for meaningful external comparisons.</div></div>","PeriodicalId":56092,"journal":{"name":"Physica Medica-European Journal of Medical Physics","volume":"139 ","pages":"Article 105178"},"PeriodicalIF":2.7,"publicationDate":"2025-10-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145271145","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}