Physics and Imaging in Radiation Oncology最新文献

{"title":"Conditional organs-of-interest segmentation with plausible inter-fraction variation simulation for pancreatic magnetic resonance-guided radiotherapy in limited-data settings","authors":"Mehdi Shojaei , Björn Eiben , Jamie R. McClelland , Simeon Nill , Alex Dunlop , Robert W. Chuter , Arabella Hunt , Brian Ng-Cheng-Hin , Uwe Oelfke","doi":"10.1016/j.phro.2026.100973","DOIUrl":"10.1016/j.phro.2026.100973","url":null,"abstract":"<div><h3>Background and Purpose:</h3><div>Manual contouring of organs of interest (OoIs) is a major bottleneck in pancreatic magnetic resonance-guided online adaptive radiotherapy (oART). We developed C-SegDeform, a data-efficient conditional segmentation framework that used structure-guided deformation-based augmentations to simulate plausible inter-fraction anatomical variation and leveraged organ-specific conditioning as an alternative to registration-based contour propagation (Prop-ROIs) in limited-data settings.</div></div><div><h3>Materials and Methods:</h3><div>Forty balanced 3DVane images from 12 patients were manually contoured and pre-processed, including duodenum, both kidneys, liver, large and small bowel, spinal canal, spleen, and stomach. The training dataset (26 images) was augmented by simulating plausible session images via structure-guided deformations. Data was arranged for conditional segmentation and leave-one-out cross-validation using the nnU-Net framework. Analysis included geometric (DSC, average surface distance (ASD), and 95th percentile of Hausdorff distance (HD<sub>95</sub>); against Prop-ROIs and TotalSegmentator MRI), dose (via 8 treatment plans, measuring <span><math><msub><mrow><mi>D</mi></mrow><mrow><mn>0</mn><mo>.</mo><mn>1</mn><mspace></mspace><mi>c</mi><msup><mrow><mi>m</mi></mrow><mrow><mn>3</mn></mrow></msup></mrow></msub></math></span> and <span><math><msub><mrow><mi>D</mi></mrow><mrow><mn>50</mn><mtext>%</mtext></mrow></msub></math></span> discrepancies relative to prescribed dose), and clinical (Likert scale and post-auto-contour editing times by two oncology consultants) assessments.</div></div><div><h3>Results:</h3><div>C-SegDeform (DSC: <span><math><mrow><mn>0</mn><mo>.</mo><mn>88</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>0</mn><mo>.</mo><mn>10</mn></mrow></math></span>, ASD: <span><math><mrow><mn>2</mn><mo>.</mo><mn>9</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>2</mn><mo>.</mo><mn>4</mn><mspace></mspace><mi>mm</mi></mrow></math></span>, HD<sub>95</sub>: <span><math><mrow><mn>10</mn><mo>.</mo><mn>5</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>8</mn><mo>.</mo><mn>9</mn><mspace></mspace><mi>mm</mi></mrow></math></span>) outperformed Prop-ROIs (<span><math><mrow><mn>0</mn><mo>.</mo><mn>70</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>0</mn><mo>.</mo><mn>17</mn></mrow></math></span>, <span><math><mrow><mn>8</mn><mo>.</mo><mn>5</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>9</mn><mo>.</mo><mn>9</mn><mspace></mspace><mi>mm</mi></mrow></math></span>, <span><math><mrow><mn>16</mn><mo>.</mo><mn>7</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>9</mn><mo>.</mo><mn>8</mn><mspace></mspace><mi>mm</mi></mrow></math></span>) and TotalSegmentator (<span><math><mrow><mn>0</mn><mo>.</mo><mn>75</mn><mspace></mspace><mo>±</mo><mspace></mspace><mn>0</mn><mo>.</mo><mn>16</mn></mrow></math></span>, <span><math><mrow><mn>5</mn><mo>.</mo><mn>1</mn><mspace></mspace><mo>±</mo><mspace></ms","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"39 ","pages":"Article 100973"},"PeriodicalIF":3.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147952766","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":"A multicentre audit of the geometric accuracy and water equivalence of 3D printed thermoplastic objects for use in radiotherapy","authors":"Matthew R. Jones ,&nbsp;Peter D. Woolliams ,&nbsp;Matthew A. Bolt ,&nbsp;Owen McLaughlin ,&nbsp;Maria Boutros ,&nbsp;Tristan Wright ,&nbsp;Rhys Jenkins ,&nbsp;Anna Tonino ,&nbsp;Fiona Milliken ,&nbsp;Neil Bentley ,&nbsp;Gordon D. Sands ,&nbsp;James C.L. Burnley ,&nbsp;Paul J. Doolan ,&nbsp;Dualta McQuaid ,&nbsp;Geoff J. Budgell ,&nbsp;Elizabeth J. Adams ,&nbsp;Conor K. McGarry ,&nbsp;Catharine H. Clark","doi":"10.1016/j.phro.2026.101006","DOIUrl":"10.1016/j.phro.2026.101006","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Plastic materials are widely used as water substitutes in radiotherapy; however, the dosimetric properties of thermoplastic polymers used in 3D printing can vary. A multicentre audit was conducted to quantify variations in geometric accuracy, density and water-mimicking properties of 3D-printed objects.</div></div><div><h3>Materials and methods</h3><div>Ten centres printed three polylactic acid (PLA) blocks at varying infills. Block dimensions, including protruding and recessed discs, were measured and the blocks weighed to determine geometric accuracy and mass density. Computed tomography (CT) scans were used to derive the printed infills and mass densities corresponding to water equivalence. Tissue phantom ratios (TPRs) were measured for 6<!--> <!-->MV photon beams and compared to vendor-provided reference data for water.</div></div><div><h3>Results</h3><div>A mean error of (−0.2 ± 0.3)<!--> <!-->mm (mean ± standard deviation, SD) was found for all printed dimensions and disc diameters. Measured CT numbers and mass densities varied by up to 200 Hounsfield units<!--> <!-->(HU) and 0.20 <!--> <!-->g/cm³ between centres, respectively. Printed infill and mass densities producing water-equivalent CT density were calculated as (94.5 ± 3.5)% and (1.10 ± 0.03) g/cm³, respectively. Calculated water-equivalent thicknesses formed using combinations of the blocks varied by up to 3.0 <!--> <!-->mm between centres, with measured TPR data varying by up to 1.2%. Measured and calculated data were normally distributed across centres and fell within ± 2 SD of their respective means. TPR data closely emulated reference values for water.</div></div><div><h3>Conclusions</h3><div>A multicentre audit was completed to develop understanding of geometric and dosimetric errors associated with 3D printing in radiotherapy.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"39 ","pages":"Article 101006"},"PeriodicalIF":3.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148185135","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":"Assessing the technical capability of a room- and a gantry-mounted kV imaging device for intra-fractional fluoroscopy during stereotactic lung radiotherapy at a C-arm linear accelerator","authors":"Hannah Jungreuthmayer ,&nbsp;Barbara Knäusl ,&nbsp;Julius Arnold ,&nbsp;Martin Buschmann ,&nbsp;Andreas Renner ,&nbsp;Maximilian Schmid ,&nbsp;Dietmar Georg ,&nbsp;Wolfgang Lechner","doi":"10.1016/j.phro.2026.100988","DOIUrl":"10.1016/j.phro.2026.100988","url":null,"abstract":"<div><h3>Background and Purpose:</h3><div>Monitoring lung tumors during stereotactic radiotherapy is important for managing breathing motion, yet the high-performance imaging modules on C-arm linear accelerators are mainly used for patient positioning. This study benchmarked a room-mounted against a gantry-mounted kilovolt (kV) imaging device to assess their potential for lung tumor fluoroscopy.</div></div><div><h3>Materials and Methods:</h3><div>The investigated systems were the ExacTrac Dynamic (ETD, Brainlab, Germany) and the X-ray Volumetric Imaging (XVI, Elekta, Sweden). Contrast-to-Noise Ratio (CNR) and entrance air kerma were measured for different imaging settings — tube voltage (70<!--> <!-->kV to 130<!--> <!-->kV), current (10<!--> <!-->mA to 320<!--> <!-->mA), exposure time (100<!--> <!-->ms and 40<!--> <!-->ms) — using two anthropomorphic thorax phantoms housing three tumors. Additionally, CNR change as a function of time, achievable fluoroscopy imaging duration based on X-ray generator and anode heat and X-ray tube cool-down were evaluated.</div></div><div><h3>Results:</h3><div>CNRs agreed within 15% for the two systems and changed less than 2% over the investigated fluoroscopy durations. Using default imaging settings and considering the average clinical duration of stereotactic lung treatments, entrance air kermas were 44<!--> <!-->mGy (room-mounted) and 41<!--> <!-->mGy (gantry-mounted). Achievable fluoroscopy durations with default settings were 87<!--> <!-->s and 1350<!--> <!-->s, respectively. Anode cool-down times were 125<!--> <!-->min (room-mounted) and 215<!--> <!-->min (gantry-mounted).</div></div><div><h3>Conclusion:</h3><div>The performance metrics of the room-mounted imaging system aligned with those of the gantry-mounted device with regard to CNR and air kerma. The room-mounted system showed a preferable cool-down behavior. However, adaptation of imaging settings is necessary to increase the fluoroscopy duration and avoid overheating.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"39 ","pages":"Article 100988"},"PeriodicalIF":3.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147860609","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":"Estimation of accumulated dose to organs at risk in head and neck cancer patients treated with scheduled replanning and dose painting in the ARTFORCE trial","authors":"Anna Liza M.P. de Leeuw ,&nbsp;Simon R. van Kranen ,&nbsp;Jordi Giralt ,&nbsp;Yungan Tao ,&nbsp;Sergi Benavente ,&nbsp;Thanh-Vân F. Nguyen ,&nbsp;Frank J.P. Hoebers ,&nbsp;Ann Hoeben ,&nbsp;Chris H.J. Terhaard ,&nbsp;Lip Wai Lee ,&nbsp;Signe Friesland ,&nbsp;Roel J.H.M. Steenbakkers ,&nbsp;Harry Bartelink ,&nbsp;Coen R.N. Rasch ,&nbsp;Jan-Jakob Sonke ,&nbsp;Olga Hamming-Vrieze","doi":"10.1016/j.phro.2026.100957","DOIUrl":"10.1016/j.phro.2026.100957","url":null,"abstract":"<div><h3>Background and purpose</h3><div>This study investigated whether scheduled adaptive radiotherapy (ART) improved delivered dose to organs at risk (OAR) in patients with locally advanced head and neck cancer treated with dose painting (DP).</div></div><div><h3>Materials and methods</h3><div>Delivered doses were estimated for 81 patients who were prospectively treated with DP + ART using deformable image registration between CBCT and the planning CT’s. Three simulations were conducted, evaluating (1) a scenario without replanning (Sim_noART), (2) a scenario with ART at fraction 12 (Sim_ART), and (3) the clinical practice with ad hoc replanning (Sim_delivered). It was further evaluated whether selecting patients for ART using accumulated dose in the first 10 fractions (Df10) to the parotid glands and larynx would improve ART efficacy.</div></div><div><h3>Results</h3><div>In Sim_noART, 41% of patients had delivered dose deviations ≥ 3 Gy compared to the planned dose in any OAR, primarily in the parotid glands and larynx. No significant differences were seen between Sim_noART, Sim_ART and Sim_delivered (P ≥ 0.10). Df10 predicted relevant changes upon completing treatment with AUC ≥ 0.95. By selecting patients for ART using Df10, delivered dose significantly improved for the larynx (P ≤ 0.01).</div></div><div><h3>Conclusions</h3><div>Although relevant dose differences between planned and delivered doses were seen in almost half of the patients without ART, minimal improvements in delivered doses were seen introducing ART. Nonetheless, Df10 was prognostic for relevant changes upon completing treatment and selecting patients for ART significantly improved larynx dose. Incorporating accumulated dose into patient selection for ART could help avoid unforeseen increases in delivered dose.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"39 ","pages":"Article 100957"},"PeriodicalIF":3.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147860612","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":"Assessing the quality of deformable image registration in magnetic resonance image guided carbon ion liver radiotherapy with an anthropomorphic phantom","authors":"Rita Pestana ,&nbsp;Anahita Bakhtiari Moghaddam ,&nbsp;Friderike K. Longarino ,&nbsp;Cedric Beyer ,&nbsp;Abdallah Qubala ,&nbsp;Katharina Seidensaal ,&nbsp;Jürgen Debus ,&nbsp;Sebastian Klüter ,&nbsp;Armin Runz ,&nbsp;Oliver Jäkel ,&nbsp;Julia Bauer","doi":"10.1016/j.phro.2026.100992","DOIUrl":"10.1016/j.phro.2026.100992","url":null,"abstract":"<div><h3>Background and Purpose</h3><div>Carbon ion radiotherapy (CIRT) for liver patients is usually administrated in four high-dose fractions, requiring high geometrical precision. Work is underway to implement online magnetic resonance (MR)-guided adaptive workflows for this indication. For the generation of a daily computed tomography (dCT), deformable image registration (DIR) can be used. Here, a study on a deformable anthropomorphic abdomen phantom was conducted to assess the accuracy of MR-to-CT DIR.</div></div><div><h3>Materials and Methods</h3><div>We applied eight different static compressions on the phantom to induce deformations, acquired CT and MR images and performed DIR to generate dCT images for each MR. We assessed the geometric uncertainties of DIR for T1- and T2-weighted MRI sequences by comparing the deformed structures against a manually segmented ground-truth. Furthermore, we assessed the DIR impact on the dose distribution on a single-beam CIRT plan, comparing the beam’s range (R80%) on the dCT against those of the CT images.</div></div><div><h3>Results</h3><div>From the geometric analysis, we obtained mean Dice similarity coefficients (DSC) for the liver of 0.94 ± 0.02 for the T1- and 0.95 ± 0.10 for the T2-weighted MR sequences. These values were above the mean deformation induced on the liver (DSC of 0.87 ± 0.08). As for the range uncertainty induced by DIR, we observed mean R80% uncertainties up to 1.35 mm when comparing treatment doses on the dCT and CT images.</div></div><div><h3>Conclusions</h3><div>Geometric and beam range uncertainties have been assessed systematically and were found to be small. These results support further implementation of an MR-guided DIR-based online adaptive workflow for liver CIRT.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"39 ","pages":"Article 100992"},"PeriodicalIF":3.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147952763","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":"Comparative analysis of artificial intelligence-based contouring of cardiac substructures on computed tomography scans for radiation therapy","authors":"Alexandra Moignier ,&nbsp;Tanguy Perennec ,&nbsp;Elise Prangères ,&nbsp;Bastien Bernard ,&nbsp;Angela Botticella ,&nbsp;Xinru Chen ,&nbsp;Robert Finnegan ,&nbsp;Sandrine Huger ,&nbsp;Anna Karlhede ,&nbsp;Thomas Lacornerie ,&nbsp;Fredrik Löfman ,&nbsp;Jérémy Palisson ,&nbsp;Charlotte Robert ,&nbsp;Killian Sambourg ,&nbsp;Jonas Söderberg ,&nbsp;Remus Stoica ,&nbsp;Grégory Delpon ,&nbsp;Elvire Martin-Mervoyer ,&nbsp;François Thillays ,&nbsp;Loïg Vaugier","doi":"10.1016/j.phro.2026.100935","DOIUrl":"10.1016/j.phro.2026.100935","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Artificial intelligence-based contouring tools enable assessment of radiation doses to cardiac substructures beyond mean heart dose. This study examined inter-solution variations in raw contours and the impact of non-contrast enhancement on contours for each solution.</div></div><div><h3>Materials and methods</h3><div>Contrast-enhanced (CE) and non-contrast-enhanced (NCE) breath-hold thoracic computed tomography (CT) scans, sequentially acquired during the same imaging session for twenty lung cancer patients, were used. Seven commercial, three open-source, and one in-house AI solutions were evaluated. On CE-CTs, solutions were compared using Dice Similarity Coefficient (DSC) and 95th percentile of Hausdorff distance (HD95) across each pair of solutions. Then, the effect of non-contrast enhancement on contours was assessed using volume ratios between NCE-CT and CE-CT for each solution.</div></div><div><h3>Results</h3><div>Typically, ten cardiac substructures were contoured by most of the solutions. For the whole heart, cardiac chambers and great vessels, the average median DSC was above 0.8 for 55 of the 123 structure-solution pairs (45%), and the average median HD95 was below 10 mm for 47 of the 123 structure-solution pairs (38%). For the coronary arteries, the average median DSC ranged between 0.03 and 0.50 and the average median HD95 ranged between 19 mm and 70 mm. Non-contrast enhancement influenced results variably; volume differences were below 10% for 84 of the 123 structure-solution pairs (68%).</div></div><div><h3>Conclusions</h3><div>Automatic contouring solutions exhibited inter-solution variability for cardiac substructures that may have clinical impact. Greater transparency and standardisation of models, ideally through international consensus and shared datasets, are essential.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"38 ","pages":"Article 100935"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147356941","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":"Magnetic resonance imaging thoracic organ-at-risk atlas for radiation oncology","authors":"Hannah Bainbridge ,&nbsp;Michael J. Dubec ,&nbsp;Andreas Wetscherek ,&nbsp;Rob H.N. Tijssen ,&nbsp;Jose Belderbos ,&nbsp;Corine Van Es ,&nbsp;David Cobben ,&nbsp;Guido H.W. van Bogerijen ,&nbsp;David Woolf ,&nbsp;Marcel van Herk ,&nbsp;Ferry Lalezari ,&nbsp;Dow-Mu Koh ,&nbsp;Fiona McDonald ,&nbsp;Corrine Faivre-Finn","doi":"10.1016/j.phro.2026.100952","DOIUrl":"10.1016/j.phro.2026.100952","url":null,"abstract":"<div><h3>Background and purpose</h3><div>The use of Magnetic Resonance imaging (MRI) for radiotherapy (RT) planning for locally advanced non-small cell lung cancer (LA NSCLC) could improve RT precision due to its superior soft tissue definition compared to computed tomography (CT). However, thoracic oncologists have limited experience of identifying thoracic structures on MRI. The aim of this study was to provide recommendations for MR sequences for thoracic organ at risk (OAR) contouring and present an atlas and descriptive instructions for delineation of thoracic OARs in the setting of MRI-guided radiation treatment planning and guidance.</div></div><div><h3>Materials and methods</h3><div>MRI scans were acquired in nine patients with early-stage lung cancer on a diagnostic 1.5 Tesla system. MRI sequences included T₁-weighted and T₂-weighted imaging techniques, each optimised for visualisation of particular OARs. OAR delineations were carried out and reviewed by an international panel of thoracic oncologists and MR radiologists.</div></div><div><h3>Results</h3><div>Thoracic MRI OAR contouring recommendations and atlas were developed by multi-institutional collaboration of six radiation oncologists and two MR radiologists. The atlas and contouring recommendations are described alongside high-resolution contoured MR images.</div></div><div><h3>Conclusions</h3><div>This consensus MRI contouring atlas has a variety of potential applications, from integration of MRI within the standard CT-based workflow in order to improve the contouring accuracy of challenging structures such as the brachial plexus, to forming the foundation of an MRI-only workflow for use in MRI-guided treatment machines. This guideline should provide a useful reference for education and will facilitate uniformity in MRI-based contouring of OARs.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"38 ","pages":"Article 100952"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147541345","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":"AID-RT: Standardising Artificial Intelligence Documentation in RadioTherapy with a domain-specific model card","authors":"Ana M. Barragán-Montero ,&nbsp;Margerie Huet-Dastarac ,&nbsp;Silvia M. Herranz-Hernández ,&nbsp;Benjamin Tengler ,&nbsp;Emma Skarsø Buhl ,&nbsp;Arthur Galapon ,&nbsp;Carlos E. Cárdenas ,&nbsp;Marco Fusella ,&nbsp;Geoffroy Herbin ,&nbsp;Yvonne de Hond ,&nbsp;Franziska Knuth ,&nbsp;Ciaran Malone ,&nbsp;Peter van Ooijen ,&nbsp;Charlotte Robert ,&nbsp;Michele Zeverino ,&nbsp;Coen Hurkmans ,&nbsp;Tomas Janssen ,&nbsp;Stine Sofia Korreman ,&nbsp;Charlotte L. Brouwer","doi":"10.1016/j.phro.2026.100940","DOIUrl":"10.1016/j.phro.2026.100940","url":null,"abstract":"<div><h3>Background and Purpose</h3><div>Insufficient documentation of artificial intelligence (AI) models remains a widespread issue, which hampers reproducibility in research environments and safe integration in clinical departments. Our goal was to develop a standardised, structured, and domain-specific reporting framework tailored to AI models in radiotherapy (RT), enhancing transparency and accountability.</div></div><div><h3>Methods</h3><div>A working group was formed after the ESTRO Physics Workshop 2023, <em>“AI for the Fully Automated Radiotherapy Treatment Chain”,</em> comprising 16 experts from 13 institutions<em>.</em> We reviewed existing initiatives for AI model and data reporting and drafted an initial template, which was sent for review to all participants. Three popular RT applications were selected to define task-specific fields: synthetic CT, segmentation, and dose prediction. Five review rounds were performed, where suggested changes were voted in a shared online document. Unclear fields and conflicting votes were discussed at online meetings, and consensus was reached by majority voting.</div></div><div><h3>Results</h3><div>The final template included 6 sections: 0) Card metadata, 1) Model basic information; 2) Model technical specifications (i.e. architecture, software and hardware); 3) Training data, methodology, and information; 4) Evaluation data, methodology, and results (a.k.a commissioning for clinical models); and 5) Other considerations, including ethical use, risk analysis, and monitoring. It is publicly available as a downloadable document template and as an interactive web-based form to facilitate information entry.</div></div><div><h3>Conclusions</h3><div>We proposed a practical, consensus-driven template tailored to the unique requirements of AI models in RT, with applicability in both research and clinical environments, conveying the key information required for informed use.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"38 ","pages":"Article 100940"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147450448","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":"Automatic segmentation of the urethra and prostate zones with deep learning on T2-weighted magnetic resonance imaging","authors":"William Holmlund ,&nbsp;Attila Simkó ,&nbsp;Karin Söderkvist ,&nbsp;Péter Palásti ,&nbsp;Szilvia Tótin ,&nbsp;Kamilla Kalmár ,&nbsp;Zsófia Domoki ,&nbsp;Zsuzsanna Fejes ,&nbsp;Tamás Z. Kincses ,&nbsp;Patrik Brynolfsson ,&nbsp;Tufve Nyholm","doi":"10.1016/j.phro.2026.100964","DOIUrl":"10.1016/j.phro.2026.100964","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Accurate segmentation of the urethra is crucial for safe focal dose escalated radiotherapy, while prostate zone identification is important for prostate cancer diagnosis. Manual delineations on magnetic resonance imaging (MRI) are labour-intensive and variable, and while deep learning offers promise in automating this process, no available solution currently exists. This study aimed to develop and evaluate a deep learning model for automatic segmentation of the urethra, prostate and all prostate zones and benchmark its performance against inter-reader variability and assess generalisability to external data from a different MRI vendor.</div></div><div><h3>Materials and methods</h3><div>The public datasets ProstateZones and PROSTATEx included 200 magnetic resonance images with manual delineations, with 160 used for training/validation and 40 with independent duplicate segmentations used as a test set. A nnU-Net deep learning model was evaluated on the unseen test set and externally validated on a dataset with 55 samples. Performance was assessed using Dice Similarity Coefficient (DSC), Surface DSC, percentile Symmetric Surface Distance, and Center Line Distance (CLD) metrics.</div></div><div><h3>Results</h3><div>The model outperformed the inter-reader variability on multiple structures, and notably on all metrics for the urethra, with median CLD values of 2.8 and 2.9 mm compared to 3.6 mm for inter-reader variability. External validation showed robust generalisability to a dataset collected from a different vendor.</div></div><div><h3>Conclusions</h3><div>This study demonstrated that a deep learning model can achieve expert-level performance in automated segmentation of the urethra, prostate, and prostate zones. Robust performance on external data highlighted potential as a decision support solution.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"38 ","pages":"Article 100964"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147656633","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":"Fuzzy masks: boosting radiomic reliability in head and neck tumors amid delineation uncertainty","authors":"Jin Cao ,&nbsp;Jiang Zhang ,&nbsp;Xinzhi Teng ,&nbsp;Xinyu Zhang ,&nbsp;Saikit Lam ,&nbsp;Ta Zhou ,&nbsp;Yuanpeng Zhang ,&nbsp;Jing Cai","doi":"10.1016/j.phro.2026.100947","DOIUrl":"10.1016/j.phro.2026.100947","url":null,"abstract":"<div><h3>Background and purpose</h3><div>The clinical utility of radiomics in head-and-neck (H&amp;N) cancer is hindered by poor reliability caused by delineation uncertainties from the use of binary mask (BinMask). This study introduced a fuzzy mask (FuzzMask) approach to enhance the reliability of computed tomography (CT)-based radiomics for precision prognosis.</div></div><div><h3>Materials and methods</h3><div>This retrospective study included 2,539 H&amp;N cancer patients (855 laryngeal cancer (LC), 1,336 oropharyngeal cancer (OPC), 348 nasopharyngeal carcinoma (NPC)). Delineation uncertainty was simulated via perturbation techniques. Radiomic features (RFs) were extracted using BinMask and FuzzMask, respectively. The evaluation focused on feature reliability and relevance via the intraclass correlation coefficient (ICC) and hierarchical clustering. In addition, the predictive performance and output reliability of penalized Cox’s proportional hazard models were assessed using the concordance index (C-index) and ICC, respectively.</div></div><div><h3>Results</h3><div>The FuzzMask improved feature reliability, yielding 21, 29, and 5 additional reliable features for LC, OPC, and NPC cancers, respectively, compared to BinMask. The FuzzMask also reduced feature redundancy, generating up to 70 more clusters in hierarchical clustering, particularly for smaller tumors in complex peritumoral environments although showed marginal improvements in predictive performance (C-index: +0.1% for OPC, +0.4% for NPC, <em>p</em> &gt; 0.05). However, model reliability was enhanced by FuzzMask, with ICC values increasing by 0.024, 0.022, and 0.007 for LC, OPC, and NPC, respectively, compared to BinMask (<em>p</em> ≥ 0.05).</div></div><div><h3>Conclusions</h3><div>The proposed FuzzMask technique significantly improved feature reliability and model robustness against delineation uncertainty, offering greater trustworthiness for clinical translation, although predictive accuracy remains unaffected.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"38 ","pages":"Article 100947"},"PeriodicalIF":3.3,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147541347","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}