Physics and Imaging in Radiation Oncology最新文献

{"title":"Advancing offline magnetic resonance-guided prostate radiotherapy through dedicated imaging and deep learning-based automatic contouring of targets and neurovascular structures","authors":"Philipp Schubert ,&nbsp;Matthias May ,&nbsp;Daniel Höfler ,&nbsp;Hans-Peter Fautz ,&nbsp;Jana Hutter ,&nbsp;Ricarda Merten ,&nbsp;Sina Mansoorian ,&nbsp;Thomas Weissmann ,&nbsp;Lisa Deloch ,&nbsp;Miriam Schonath ,&nbsp;Nathalia Belmas ,&nbsp;Felix Grabenbauer ,&nbsp;Benjamin Frey ,&nbsp;Udo Gaipl ,&nbsp;Bernd-Niklas Axer ,&nbsp;Juliane Szkitsak ,&nbsp;Michael Uder ,&nbsp;Christoph Bert ,&nbsp;Rainer Fietkau ,&nbsp;Florian Putz","doi":"10.1016/j.phro.2025.100825","DOIUrl":"10.1016/j.phro.2025.100825","url":null,"abstract":"<div><h3>Background and Purpose</h3><div>Erectile dysfunction (ED) affects quality of life following radiotherapy for prostate cancer. Magnetic resonance imaging (MRI) planning provides superior visualization of potency-related anatomical structures compared to computed tomography (CT), enabling improved sparing. However, contouring these structures in clinical practice is time-intensive and requires expertise. Deep learning (DL) auto-contouring with MRI simulation could make neurovascular-sparing radiotherapy more accessible.</div></div><div><h3>Material and Methods</h3><div>High-resolution 3D T2-weighted SPACE MRI sequences (&lt;1 mm³ resolution) were obtained for 50 patients in treatment position. An expert uro-radiation oncologist contoured erectile function-related anatomy (neurovascular bundles [NVB], pudendal arteries [IPA], penile bulb [PB], corpora cavernosa [CC]) and target structures (prostate [PR], seminal vesicles [SV]). Forty datasets trained and ten tested a 3D nnU-net model. DL-generated contours were geometrically evaluated (surface Dice Score [sDSC], mean surface distance [MSD]) and validated by blinded expert review.</div></div><div><h3>Results</h3><div>DL auto-segmentation achieved an average sDSC of 0.82 (IPA: 0.93, NVB: 0.71, PB: 0.84, CC: 0.90, PR: 0.74; SV: 0.79) and average MSD of 0.74 mm (IPA: 0.61 mm; NVB: 0.88 mm; PB: 0.63 mm; CC: 0.47 mm; PR 0.83 mm; SV: 1.01 mm). Blinded ratings showed no significant differences between DL and expert contours, except for pudendal arteries (Mean DL vs. expert; NVB: 82 vs. 85; PB: 86 vs. 88; CC: 83 vs. 88; PR 81 vs. 83; SV 78 vs. 81 all p &gt; 0.05; IPA: 82 vs. 89; p = 0.028).</div></div><div><h3>Conclusion</h3><div>Combining high-resolution MRI simulation with DL postprocessing enables accurate auto-contouring for MR-guided SBRT planning, potentially advancing neurovascular-sparing radiotherapy beyond current standards.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100825"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144889731","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 proof-of-concept study of direct magnetic resonance imaging-based proton dose calculation for brain tumors via neural networks with Monte Carlo-comparable accuracy","authors":"Muheng Li ,&nbsp;Carla Winterhalter ,&nbsp;Xia Li ,&nbsp;Sairos Safai ,&nbsp;Antony Lomax ,&nbsp;Ye Zhang","doi":"10.1016/j.phro.2025.100806","DOIUrl":"10.1016/j.phro.2025.100806","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Proton therapy currently relies on computed tomography (CT) imaging despite magnetic resonance imaging’s (MRI) superior soft-tissue contrast. While synthetic CTs can be generated from magnetic resonance (MR) images, this introduces additional complexity. We present a deep learning-based dose engine enabling direct proton dose calculation from MR images to streamline workflows while maintaining Monte Carlo (MC)-level accuracy.</div></div><div><h3>Materials and methods</h3><div>Using paired MR-CT scans from 39 brain tumor patients (29/3/7 for training/validation/testing), we developed a deep learning framework using various sequence models for individual proton pencil beam dose prediction. The framework processes beam-eye-view patches from 2000 random beam configurations per patient, varying in angles and energy, with corresponding MC dose distributions pre-calculated on CT. Models using CT images were trained for comparison.</div></div><div><h3>Results</h3><div>The xLSTM architecture performed best for both MR and CT-based scenarios among the evaluated sequence models. For full treatment plans, our model achieved gamma pass rates with median 99.8 % (range: 98.6 %–99.9 %, 1 mm/1%), and median percentage dose errors of 0.2 % (range: 0.1 %–0.4 %) within patient bodies and 1.3 % (range: 0.8 %–3.7 %) in high-dose regions (&gt;90 % prescription dose). The model required only 3 ms per beam prediction compared to 2 s for MC simulation.</div></div><div><h3>Conclusion</h3><div>This study demonstrated the feasibility of MC-quality proton dose calculations directly from MR images for brain tumor patients, achieving comparable accuracy with faster computation and simplified implementation.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100806"},"PeriodicalIF":3.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144596154","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":"Gaussian primitives for deformable image registration","authors":"Jihe Li ,&nbsp;Xiang Liu ,&nbsp;Fabian Zhang ,&nbsp;Xixin Cao ,&nbsp;Joachim M. Buhmann ,&nbsp;Ye Zhang ,&nbsp;Xia Li","doi":"10.1016/j.phro.2025.100821","DOIUrl":"10.1016/j.phro.2025.100821","url":null,"abstract":"<div><h3>Background and Purpose:</h3><div>Deformable image registration (DIR) plays a critical role in radiotherapy by compensating for anatomical deformations. However, existing iterative and data-driven methods are often hindered by computational inefficiency or limited generalization. In response, our objective was to develop a novel optimization-based DIR method that reduces computational overhead and preserves the robust generalization of iterative methods while enhancing interpretability.</div></div><div><h3>Materials and Methods:</h3><div>We proposed GaussianDIR, a novel DIR framework that explicitly represents the deformation field using a sparse set of adaptive Gaussian primitives. Each primitive is characterized by its centre, covariance, and associated local rigid deformation. Voxel-wise displacements are derived via blending the local rigid deformations of neighbouring primitives, enabling flexible yet efficient motion modelling.</div></div><div><h3>Results:</h3><div>On DIRLab lung dataset, GaussianDIR achieved a target registration error (TRE) of <span><math><mrow><mn>1</mn><mo>.</mo><mn>00</mn><mo>±</mo><mn>1</mn><mo>.</mo><mn>11</mn></mrow></math></span> millimeters in about 2.5 s, offering an effective trade-off between speed and precision for high-resolution images. On OASIS brain and ACDC cardiac datasets, the Dice similarity coefficient (DSC) improved from 80.6% to 81.3% and from 81.0% to 81.2% over previous state-of-the-art methods, respectively. Moreover, we compared the generalization performance of GaussianDIR and a data-driven method on IXI dataset, and found that GaussianDIR outperformed the data-driven method by 6.3% in DSC.</div></div><div><h3>Conclusion:</h3><div>GaussianDIR combines high registration accuracy with computational efficiency, interpretability, and strong generalization performance. It challenged the conventional notion that iterative methods were inherently slow and overcomed the generalization limitations of data-driven methods, with potential for real-time clinical applications in radiotherapy.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100821"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144827134","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":"First in vitro and in vivo experiments with ultra high-dose rate oxygen ion radiotherapy","authors":"Celine Karle ,&nbsp;Domenico I. Filosa ,&nbsp;Mahdi Akbarpour ,&nbsp;Nora Schuhmacher ,&nbsp;Stephan Brons ,&nbsp;Rainer Cee ,&nbsp;Christian Schömers ,&nbsp;Stefan Scheloske ,&nbsp;Kristoffer Petersson ,&nbsp;Thomas Haberer ,&nbsp;Amir Abdollahi ,&nbsp;Jürgen Debus ,&nbsp;Thomas Tessonnier ,&nbsp;Mahmoud Moustafa ,&nbsp;Andrea Mairani ,&nbsp;Ivana Dokic","doi":"10.1016/j.phro.2025.100803","DOIUrl":"10.1016/j.phro.2025.100803","url":null,"abstract":"<div><div>Within this study, we demonstrated the feasibility of ultra-high dose rate (UHDR) oxygen ion irradiation at three different levels of biological complexity. The difference in oxygen consumption between UHDR and standard dose rates (SDR) was negligible in a protein-enriched saline solution. For the studied conditions of dose, dose rate and linear energy transfer (LET), UHDR irradiation showed comparable efficacy to SDR in pancreatic cancer cell killing in vitro, along with inducing a similar tumor growth delay in vivo. These findings emphasize the potential of high-LET UHDR irradiation and support further investigation of oxygen ions at UHDR.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100803"},"PeriodicalIF":3.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144604696","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":"Evaluating the technical feasibility of biology-guided dose painting in proton therapy","authors":"Caterina Brighi ,&nbsp;Giovanni Parrella ,&nbsp;Letizia Morelli ,&nbsp;Silvia Molinelli ,&nbsp;Giuseppe Magro ,&nbsp;Sara Lillo ,&nbsp;Alberto Iannalfi ,&nbsp;Mario Ciocca ,&nbsp;Sara Imparato ,&nbsp;David E.J. Waddington ,&nbsp;Paul Keall ,&nbsp;Chiara Paganelli ,&nbsp;Ester Orlandi ,&nbsp;Guido Baroni","doi":"10.1016/j.phro.2025.100832","DOIUrl":"10.1016/j.phro.2025.100832","url":null,"abstract":"<div><div>Biology-guided voxel-level inverse prescription mapping for dose painting (DP) using diffusion-weighted magnetic resonance imaging was evaluated for technical feasibility in proton therapy for 10 skull-base chordoma patients. Patient-specific DP prescriptions were generated from tumour cellularity and implemented in a clinical treatment planning system. Compared with uniform plans, DP achieved lower conformity (although &gt;97 %), improved target dose metrics, reduced doses to most organs at risk, and increased tumour control probability without exceeding clinical constraints. DP proton therapy is technically feasible and may enhance treatment effectiveness.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100832"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144920152","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":"Feasibility of Monte Carlo-based patient-specific quality assurance in 1.5 Tesla magnetic resonance-guided online adaptive radiotherapy: a multi-institutional study","authors":"Ruiqi Li ,&nbsp;Yao Zhao ,&nbsp;Jingying Lin ,&nbsp;Tsuicheng Chiu ,&nbsp;Weiguo Lu ,&nbsp;Jinzhong Yang ,&nbsp;Mu-Han Lin","doi":"10.1016/j.phro.2025.100800","DOIUrl":"10.1016/j.phro.2025.100800","url":null,"abstract":"<div><h3>Introduction</h3><div>To evaluate the feasibility of Monte Carlo (MC)-based patient-specific quality assurance (PSQA) for MR-guided online adaptive radiotherapy and to explore the potential to eliminate the post-delivery measurement-based PSQA.</div></div><div><h3>Material and methods</h3><div>A total of 113 cases from two institutions, treated on MR-Linac machines, were included in the study. A customized GPU-accelerated, Monte Carlo-based secondary dose verification software (ART2Dose) was developed and integrated into the QA workflow, accounting for a 1.5 Tesla magnetic field. PSQA included ArcCheck (AC) delivery QA and online MC calculation-based QA. Reference plans underwent offline validation with AC and MC, while adapt-to-shape (ATS) plans were processed through MC and post-delivery QA. Gamma pass rates (GPR) with 3 %/2mm criteria were compared statistically across methods. Radcalc was applied to compare point dose difference with MC.</div></div><div><h3>Results</h3><div>MC QA achieved GPRs of 97.5 % ± 2.0 % and 97.1 % ± 2.9 % for reference and ATS plans, comparable to AC QA (97.6 % ± 2.0 % and 96.9 % ± 3.0 %). Wilcoxon signed-rank test showed statistically significant differences between reference and ATS plan QA (p &lt; 0.05), but a Pearson correlation coefficient of 0.76 confirmed a linear relationship for MC GPR. Lung cases exhibited lower GPRs with MC compared to AC QA. MC QA demonstrated supaireerior point dose agreement with TPS (1.7 % ± 1.2 %) compared to RadCalc (4.1 % ± 1.7 %). No significant differences were observed between institutions.</div></div><div><h3>Conclusion</h3><div>MC-based QA is a robust tool for adaptive QA workflows in 1.5-T MR-Linac systems. It enhances efficiency and potentially supports the elimination of post-delivery measurement-based QA for adaptive plans.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100800"},"PeriodicalIF":3.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144557536","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":"Dose-response mapping of bladder and rectum in prostate cancer patients undergoing radiotherapy with and without baseline toxicity correction","authors":"Tanuj Puri ,&nbsp;Tiziana Rancati ,&nbsp;Petra Seibold ,&nbsp;Adam Webb ,&nbsp;Eliana Vasquez Osorio ,&nbsp;Andrew Green ,&nbsp;Eliana Gioscio ,&nbsp;David Azria ,&nbsp;Marie-Pierre Farcy-Jacquet ,&nbsp;Jenny Chang-Claude ,&nbsp;Alison Dunning ,&nbsp;Maarten Lambrecht ,&nbsp;Barbara Avuzzi ,&nbsp;Dirk de Ruysscher ,&nbsp;Elena Sperk ,&nbsp;Ana Vega ,&nbsp;Liv Veldeman ,&nbsp;Barry Rosenstein ,&nbsp;Jane Shortall ,&nbsp;Sarah Kerns ,&nbsp;Marcel van Herk","doi":"10.1016/j.phro.2025.100805","DOIUrl":"10.1016/j.phro.2025.100805","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Radiotherapy dose–response maps (DRM) combine dose-surface maps (DSM) and toxicity outcomes to identify high-risk subregions in organ-at-risk. This study assesses the impact of baseline toxicity correction on the identification of high-risk subregions in dose–response modeling for prostate cancer patients undergoing radiotherapy.</div></div><div><h3>Materials and methods</h3><div>The analysis included 1808 datasets, with 589 exclusions before toxicity-specific data removal. Bladder/rectum were automatically segmented on planning computed tomography scans, DSMs unwrapped into 91x90 voxel grids, and converted to equivalent doses in 2 Gy fractions (EQD2; α/β = 1 Gy). Seventeen late toxicities were assessed with two methods: (i) baseline toxicity subtracted from the maximum of 12- and 24-months toxicity scores, dichotomized at grade 1, and (ii) maximum of 12- and 24-months toxicity scores dichotomized at grade 1. DSMs were split accordingly, and voxel-wise t-values computed using Welch’s t-equation. Statistically significant voxels were identified via the 95th percentile of maximum of t-value (Tmax) distribution.</div></div><div><h3>Results</h3><div>Event counts with baseline correction were 82/82/286/226 for urinary tract obstruction/retention/urgency/incontinence, respectively; without baseline correction, they were 93/104/465/361. For bladder DSMs, urinary incontinence, obstruction, retention, and urgency had 1143/186, 1768/1848, 516/0, and 33/0 significant voxels without/with baseline correction. For rectum DSMs, urinary incontinence and tract obstruction had 604/0 and 1980/889 significant voxels without/with baseline correction. However, no significant associations between rectal DSMs and rectum-related toxicities were found.</div></div><div><h3>Conclusions</h3><div>DRM without baseline correction appears more sensitive to high-risk subregions due to higher event counts. Non-linear toxicity grading and multivariable analysis may enhance DRM reliability.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100805"},"PeriodicalIF":3.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144570063","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":"In-vitro effects of modern radiotherapy regimes on cardiac implanted electrical devices","authors":"Steven Manley ,&nbsp;Amy Fenwick ,&nbsp;Megan Fraser ,&nbsp;Mark Lowrey ,&nbsp;Anna Walaszczyk ,&nbsp;Nick West","doi":"10.1016/j.phro.2025.100804","DOIUrl":"10.1016/j.phro.2025.100804","url":null,"abstract":"<div><h3>Background and purpose</h3><div>External Beam Radiotherapy (EBRT) of patients with Cardiac Implanted Electrical Devices (CIEDs) have guidelines developed over many decades, during which both technologies have advanced. Consequently, guidelines may not reflect modern device interactions. Asynchronous modes, with pace sensing and shocks turned off whilst regulating pacing output, is used routinely for MRI scanning and could reduce risks for pace sensing errors in radiotherapy. Evidence is limited on modern radiotherapy using high dose, high doserate beams, with CIEDs in asynchronous mode. We present the effects of irradiating modern CIEDs using contemporary radiotherapy regimes.</div></div><div><h3>Materials and methods</h3><div>One hundred and sixty explanted CIEDs were irradiated, to corroborate historical findings for modern devices in normal operational modes, and explore effects when in asynchronous mode. Regimes knowingly detrimental to CIEDs; 48 Gy single fraction, neutron producing, alongside clinically relevant regimes of 60 Gy in 5 fractions using 10 MV flattening filter free [FFF] beams at clinical, and maximal dose rates.</div></div><div><h3>Results</h3><div>No significant changes occurred to pacing voltages from 60 Gy in 5 fractions 10 MV FFF deliveries in asynchronous mode.</div><div>No evidence supported restricting 6 MV flattened beams for CIED patients, including defibrillation capable devices.</div></div><div><h3>Conclusions</h3><div>This study demonstrates asynchronous mode can reduce the frequency of CIED malfunctions during EBRT. However, clinical context, risks and benefit must be evaluated per patient. While some current guidelines potentially compromise plan quality to reduce dose to the CIED, the use of asynchronous mode may provide planning options, which more closely align to non-CIED cases.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100804"},"PeriodicalIF":3.4,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534824","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 prognostic value of baseline 18FDG – Positron Emission Tomography – Computed Tomography in cervical cancer patients treated with definitive chemoradiotherapy – External multicentre validation model","authors":"Emilia Staniewska ,&nbsp;Magdalena Stankiewicz ,&nbsp;Ewa Burchardt ,&nbsp;Karolina Grudzien ,&nbsp;Katarzyna Raczek-Zwierzycka ,&nbsp;Justyna Rembak-Szynkiewicz ,&nbsp;Matylda Sobczak ,&nbsp;Andrea d’Amico ,&nbsp;Damian Borys ,&nbsp;Izabela Gorczewska ,&nbsp;Kamil Krysiak ,&nbsp;Martin Rydzinski ,&nbsp;Rafal Stando ,&nbsp;Mateusz Spalek ,&nbsp;Zuzanna Nowicka ,&nbsp;Rafal Tarnawski ,&nbsp;Marcin Miszczyk","doi":"10.1016/j.phro.2025.100829","DOIUrl":"10.1016/j.phro.2025.100829","url":null,"abstract":"<div><h3>Background and purpose</h3><div>18F-Fluorodeoxyglucose-Positron Emission Tomography – Computed Tomography (18FDG-PET-CT) is commonly used for baseline clinical staging in cervical cancer. In this study, we assessed the prognostic value of standardised PET-CT parameters for the overall survival (OS) of patients treated with definitive chemoradiotherapy (CRT), or radiotherapy (RT) with subsequent brachytherapy (BT).</div></div><div><h3>Material and methods</h3><div>This study included consecutive cervical cancer patients treated with definitive CRT or RT and BT, between 2011 and 2017, at a single tertiary institution. Each patient had a 18-FDG-PET-CT scan before treatment. Patients from three institutions with equal inclusion criteria were included in external validation group. The metabolic parameters of the primary tumour: standardized uptake value (SUV) derivatives, metabolic tumour volume (MTV) and total lesion glycolysis (TLG), were evaluated using the semi-automated method. Statistical analysis was conducted using the Kaplan–Meier method, log-rank tests, Cox regression models, and the Akaike Information Criterion (AIC).</div></div><div><h3>Results</h3><div>The study group included 198 patients treated with RT (100 %) concurrent with Cisplatin-based CT (91.4 %), and subsequent BT (99 %). The majority of patients were diagnosed with squamous cell carcinoma (96.5 %) and International Federation of Gynaecology and Obstetrics (FIGO) stage III disease (84.3 %). The OS was significantly higher in patients with TLG30 below the median value (78.8 % vs. 58.9 %; p &lt; 0.01). TLG30 remained as the only independent prognostic factor (hazard ratio 1.32,95 % confidence interval:1.12–1.56, p &lt; 0.01). In the external validation model neither of analysed PET parameters were significant.</div></div><div><h3>Conclusions</h3><div>A high TLG30 was associated with worse OS in primary cohort. However, external validation model did not confirm the clinical utility of the cervical tumour’s metabolic parameters.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100829"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144987834","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":"Hybrid deep learning enables multi-institutional delineation of active bone marrow for gynecologic radiotherapy","authors":"Zhe Zhang ,&nbsp;Xiao Lu ,&nbsp;Sicheng He ,&nbsp;Tao Huang ,&nbsp;Shaobin Wang ,&nbsp;Mingjun Lu ,&nbsp;Xiaomin Zhang ,&nbsp;Zhibo Tan ,&nbsp;John Moraros ,&nbsp;Lei Zhang ,&nbsp;Xin Li ,&nbsp;Zhan Li ,&nbsp;Zihao Deng ,&nbsp;Yimeng Zhang ,&nbsp;Mengjie Dong ,&nbsp;Shuihua Wang ,&nbsp;Yajie Liu","doi":"10.1016/j.phro.2025.100823","DOIUrl":"10.1016/j.phro.2025.100823","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Pelvic radiotherapy for gynecologic cancer inevitably irradiates sensitive areas like iliac bones, lumbar vertebrae, and sacrum. Using ¹⁸F-FDG PET/CT as a reference, we developed a deep learning method to detect hematopoietic active bone marrow (ABM) on CT in gynecologic cancer patients and assess clinical benefits.</div></div><div><h3>Materials and methods</h3><div>We analyzed 319 patients from five institutions retrospectively. ABM was divided into three 18F-FDG PET/CT-defined subregions: active iliac bone marrow (A_IBM), active sacral bone marrow (A_SBM), and active lumbar vertebrae bone marrow (A_LVBM), defined as areas with standardized uptake values exceeding subregional means. Six deep learning models were trained: hybrid nnU-Net, U-Net, V-Net, ResU-Net, nnU-Net, and UNETR. The hybrid nnU-Net approach integrated nnU-Net predictions with anatomical bone structures via Boolean operations, providing a post-processing strategy. The dataset was split into 290 cases for training and 29 for independent testing. Performance was evaluated using Dice similarity coefficients (DSCs) and 95th percentile Hausdorff distance (HD95). Two clinical cases were prospectively evaluated for ABM-sparing radiotherapy with hematologic monitoring.</div></div><div><h3>Results</h3><div>The hybrid nnU-Net achieved the highest DSCs for A_IBM (0.74 ± 0.06), A_LVBM (0.79 ± 0.07), and A_SBM (0.75 ± 0.06), with significant improvements over most models (p &lt; 0.001), except nnU-Net. Despite ResU-Net’s lower HD95 in two subregions, hybrid nnU-Net showed superior accuracy. No grade ≥2 hematologic toxicity occurred in prospective cases.</div></div><div><h3>Conclusion</h3><div>This multi-institutional study confirms that the hybrid nnU-Net accurately segments ABM from CT images, showing potential for ABM-sparing radiotherapy.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"35 ","pages":"Article 100823"},"PeriodicalIF":3.3,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144826505","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}