Physics and Imaging in Radiation Oncology最新文献

{"title":"Ultra-fast, one-click radiotherapy treatment planning outside a treatment planning system","authors":"Gerd Heilemann ,&nbsp;Lukas Zimmermann ,&nbsp;Tufve Nyholm ,&nbsp;Attila Simkó ,&nbsp;Joachim Widder ,&nbsp;Gregor Goldner ,&nbsp;Dietmar Georg ,&nbsp;Peter Kuess","doi":"10.1016/j.phro.2025.100724","DOIUrl":"10.1016/j.phro.2025.100724","url":null,"abstract":"<div><div>We present an automated radiation oncology treatment planning pipeline that operates between segmentation and plan review, minimizing manual interaction and reliance on traditional planning systems. Two AI models work in sequence: the first generates a dose distribution, and the second creates a deliverable DICOM-RT plan. Trained and validated on 276 plans, and tested on 151 datasets, the system produced clinically deliverable plans—complete with all VMAT parameters—in about 38 s. These plans met target coverage and most organ-at-risk constraints. This proof-of-concept demonstrates the feasibility of generating high-quality, deliverable DICOM plans within seconds.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100724"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143387667","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Investigating the potential of diffusion tensor atlases to generate anisotropic clinical tumor volumes in glioblastoma patients","authors":"Kim Hochreuter ,&nbsp;Gregory Buti ,&nbsp;Ali Ajdari ,&nbsp;Christopher P. Bridge ,&nbsp;Gregory C. Sharp ,&nbsp;Sune Jespersen ,&nbsp;Slávka Lukacova ,&nbsp;Thomas Bortfeld ,&nbsp;Jesper F. Kallehauge","doi":"10.1016/j.phro.2024.100688","DOIUrl":"10.1016/j.phro.2024.100688","url":null,"abstract":"<div><h3>Background and purpose:</h3><div>Diffusion tensor imaging (DTI) has been proposed to guide the anisotropic expansion from gross tumor volume to clinical target volume (CTV), aiming to integrate known tumor spread patterns into the CTV. This study investigate the potential of using a DTI atlas as an alternative to patient-specific DTI for generating anisotropic CTVs.</div></div><div><h3>Materials and Methods:</h3><div>The dataset consisted of twenty-eight newly diagnosed glioblastoma patients from a Danish national DTI protocol with post-operative T1-contrast and DTI imaging. Three different DTI atlases, spatially aligned to the patient images using deformable image registration, were considered as alternatives. Anisotropic CTVs were constructed to match the volume of a 15 mm isotropic expansion by generating 3D distance maps using either patient- or atlas-DTI as input to the shortest path solver. The degree of CTV anisotropy was controlled by the migration ratio, modeling tumor cell migration along the dominant white matter fiber direction extracted from DTI. The similarity between patient- and atlas-DTI CTVs was analyzed using the Dice Similarity Coefficient (DSC), with significance testing according to a Wilcoxon test.</div></div><div><h3>Results:</h3><div>The median (range) DSC between anisotropic CTVs generated using patient-specific and atlas-based DTI was 0.96 (0.93–0.97), 0.96 (0.93–0.97), and 0.95 (0.93–0.97) for the three atlases, respectively (p <span><math><mo>&gt;</mo></math></span> 0.01), for a migration ratio of 10. The results remained consistent over the range of studied migration ratios (2 to 100).</div></div><div><h3>Conclusion:</h3><div>The high degree of similarity between all anisotropic CTVs indicates that atlas-DTI is a viable replacement for patient-specific DTI for incorporating fiber direction into the CTV.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100688"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11758580/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143047930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effect of end expiration breath hold on target volumes and organ at risk doses for oesophageal cancer radiotherapy","authors":"Christopher Mayhew ,&nbsp;Jeyaanth Venkatasai ,&nbsp;Marina Khan ,&nbsp;Victoria Butterworth ,&nbsp;Kasia Owczarczyk ,&nbsp;Georgios Ntentas","doi":"10.1016/j.phro.2025.100726","DOIUrl":"10.1016/j.phro.2025.100726","url":null,"abstract":"<div><h3>Background and Purpose</h3><div>The end expiration breath hold (EEBH) technique has the potential to reduce tumour motion during radiotherapy treatment of lower oesophageal cancer, and therefore, motion artefacts, target volumes and dose to surrounding organs at risk (OAR). EEBH is an emerging technique and clinical data on its use in oesophageal cancer is scarce.</div></div><div><h3>Methods and Materials</h3><div>A comparison of 20 lower oesophageal cancer patients was performed for radiotherapy treatment plans in both EEBH and free breathing (FB). EEBH and FB plans were evaluated and compared in terms of motion artefacts, target volumes and dose-volume metrics.</div></div><div><h3>Results</h3><div>EEBH was effective in reducing the observed motion artefacts seen in planning CTs compared to FB. EEBH also significantly reduced the average PTV size between EEBH and FB (ΔV = -48 ± 55 cm³; p &lt; 0.001). OAR-PTV overlap volumes were also effectively reduced in EEBH compared to FB, including for lung-PTV overlaps (ΔV = -13 ± 13 cm³; p &lt; 0.001) and for heart-PTV overlaps (ΔV = -8 ± 14 cm³; p = 0.02). Mean heart doses were lower on average by −1.2 ± 2.0 Gy with EEBH (p = 0.02), and mean lung doses by −1.0 ± 1.0 Gy (p &lt; 0.001). Mean liver doses were on average reduced with EEBH by −0.6 ± 1.5 Gy, whereas spinal D_2cm3 increased in EEBH compared to FB by 1.8 ± 6.3 Gy, but neither were statistically significant.</div></div><div><h3>Conclusion</h3><div>Use of EEBH for oesophageal cancer radiotherapy reduced motion artefacts and increased confidence in contouring volumes. Additionally, planning target volumes and doses to key OARs were reduced with EEBH compared to FB plans.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100726"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446139","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A systematic review of the role of artificial intelligence in automating computed tomography-based adaptive radiotherapy for head and neck cancer","authors":"Edoardo Mastella ,&nbsp;Francesca Calderoni ,&nbsp;Luigi Manco ,&nbsp;Martina Ferioli ,&nbsp;Serena Medoro ,&nbsp;Alessandro Turra ,&nbsp;Melchiore Giganti ,&nbsp;Antonio Stefanelli","doi":"10.1016/j.phro.2025.100731","DOIUrl":"10.1016/j.phro.2025.100731","url":null,"abstract":"<div><h3>Purpose</h3><div>Adaptive radiotherapy (ART) may improve treatment quality by monitoring variations in patient anatomy and incorporating them into the treatment plan. This systematic review investigated the role of artificial intelligence (AI) in computed tomography (CT)-based ART for head and neck (H&amp;N) cancer.</div></div><div><h3>Methods</h3><div>A comprehensive search of main electronic databases was conducted until April 2024. Titles and abstracts were reviewed to evaluate the compliance with inclusion criteria: CT-based imaging for photon ART of H&amp;N patients and AI applications. 17 original retrospective studies with samples sizes ranging from 37 to 239 patients were included. The quality of the studies was evaluated with the Quality Assessment of Diagnostic Accuracy Studies-2 and the Checklist for Artificial Intelligence in Medical Imaging (CLAIM) tools. Key metrics were examined to evaluate the performances of the proposed AI-methods.</div></div><div><h3>Results</h3><div>Overall, the risk of bias was low. The average CLAIM score was 70%. A major finding was that generated synthetic CTs improved similarity metrics with planning CT compared to original cone-beam CTs, with average mean absolute error up to 39 HU and maximum improvement of 80%. Auto-segmentation provided an efficient and accurate option for organ-at-risk delineation, with average Dice similarity coefficient ranging from 80 to 87%. Finally, AI models could be trained using clinical and radiomic features to predict the effectiveness of ART with accuracy above 80%.</div></div><div><h3>Conclusions</h3><div>Automation of processes in ART for H&amp;N cancer is very promising throughout the entire chain, from the generation of synthetic CTs and auto-segmentation to predict the effectiveness of ART.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100731"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143446137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Nontoxic generalized patient shielding devices for total skin electron therapy","authors":"Clinton Gibson,&nbsp;Joseph B. Schulz,&nbsp;Amy Yu,&nbsp;Piotr Dubrowski,&nbsp;Lawrie Skinner","doi":"10.1016/j.phro.2025.100697","DOIUrl":"10.1016/j.phro.2025.100697","url":null,"abstract":"<div><div>This study evaluates alternative shielding materials to lead for protecting the scalp and nails during total skin electron irradiation. We tested a silicone helmet, tungsten-doped silicone mittens, and planar aluminum and copper shields. The helmet and mittens were created using 3D modeling software and fused filament fabrication printing, while the planar shields were machined and assembled with printed hardware. Transmission measurements showed transmission rates of 4.5%–6.8% for the mittens, 5.8%–9.1% for the helmet, and 7.5% for the planar shields. The silicone-based devices improve comfort and usability, and slight design changes can enhance coverage and application.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100697"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143428785","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Synthetic Computed Tomography generation using deep-learning for female pelvic radiotherapy planning","authors":"Rachael Tulip ,&nbsp;Sebastian Andersson ,&nbsp;Robert Chuter ,&nbsp;Spyros Manolopoulos","doi":"10.1016/j.phro.2025.100719","DOIUrl":"10.1016/j.phro.2025.100719","url":null,"abstract":"<div><div>Synthetic Computed Tomography (sCT) is required to provide electron density information for MR-only radiotherapy. Deep-learning (DL) methods for sCT generation show improved dose congruence over other sCT generation methods (e.g. bulk density). Using 30 female pelvis datasets to train a cycleGAN-inspired DL model, this study found mean dose differences between a deformed planning CT (dCT) and sCT were 0.2 % (D98 %). Three Dimensional Gamma analysis showed a mean of 90.4 % at 1 %/1mm. This study showed accurate sCTs (dose) can be generated from routinely available T2 spin echo sequences without the need for additional specialist sequences.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100719"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143268019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimizing the dose-averaged linear energy transfer for the dominant intraprostatic lesions in high-risk localized prostate cancer patients","authors":"Bo Zhao ,&nbsp;Nobuyuki Kanematsu ,&nbsp;Shuri Aoki ,&nbsp;Shinichiro Mori ,&nbsp;Hideyuki Mizuno ,&nbsp;Takamitsu Masuda ,&nbsp;Hideyuki Takei ,&nbsp;Hitoshi Ishikawa","doi":"10.1016/j.phro.2025.100727","DOIUrl":"10.1016/j.phro.2025.100727","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Radiotherapy for localized prostate cancer often targets the entire prostate with a uniform dose despite the presence of high-risk dominant intraprostatic lesions (DILs). This study investigated the feasibility of focal dose-averaged linear energy transfer (LET_d) boost for prostate carbon-ion radiotherapy to deposit higher LET_d to DILs while ensuring desired relative biological effectiveness weighted dose coverage to targets and sparing organs at risk (OARs).</div></div><div><h3>Materials and methods</h3><div>A retrospective planning study was conducted on 15 localized prostate cancer cases. The DILs were identified on multiparametric MRI and used to define the boost target (PTV_boost). Two treatment plans were designed for each patient: 1) conventional plan optimized by the single-field uniform dose technique, and 2) boost plan optimized by the multifield optimization and LET painting technique, to achieve LET_d boost within the PTV_boost. Dose and LET_d metrics of the targets and OARs were compared between the two plans.</div></div><div><h3>Results</h3><div>Compared to the conventional plans, the boost plans delivered clinically acceptable dose coverage (D_90% and D_50%) to the target (PTV2) within 1% differences while significantly increasing the minimum LET_d by 16 ∼ 24 keV/μm for the PTV_boost (63.9 ± 2.8 vs. 44.0 ± 1.3 keV/μm, p &lt; 0.001). Furthermore, these improvements were consistent across all cases, irrespective of their anatomical features, including the boost volume’s size, location, and shape.</div></div><div><h3>Conclusion</h3><div>Focal LET_d boost was a feasible strategy for prostate carbon-ion radiotherapy. This investigation demonstrated its superiority in delivering LET_d boost without depending on tumor location and volume across different cases.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100727"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143379452","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Isolating the impact of tissue heterogeneities in high dose rate brachytherapy treatment of the breast","authors":"Jules Faucher ,&nbsp;Vincent Turgeon ,&nbsp;Boris Bahoric ,&nbsp;Shirin A. Enger ,&nbsp;Peter G.F. Watson","doi":"10.1016/j.phro.2025.100737","DOIUrl":"10.1016/j.phro.2025.100737","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Clinical brachytherapy treatment planning is performed assuming the patient is composed entirely of water and infinite in size. In this work, the effects of this assumption on calculated dose were investigated by comparing dose to water in water (D_w,w) in an unbound phantom mimicking TG-43 conditions, and dose to medium in medium (D_m,m) for breast cancer patients treated with high dose rate brachytherapy.</div></div><div><h3>Materials and methods</h3><div>Treatment plans for 123 breast cancer patients were recalculated with a Monte Carlo-based treatment planning software. The dwell times and dwell positions were imported from the clinical treatment planning system. The dose was computed and reported as D_w,w and D_m,m. Dose-volume histogram (DVH) metrics were evaluated for target volumes and organs at risk.</div></div><div><h3>Results</h3><div>D_w,w overestimated the dose for most studied DVH metrics. The largest median overestimations between D_m,m and D_w,w were seen for the planning target volume (PTV) V_200% (5.8%), lung D_{0.1 cm}³ (6.0%) and skin D_{0.1 cm}³ (4.2%). The differences between D_m,m and D_w,w were statistically significant for all investigated DVH metrics_. The PTV V_90% had the smallest deviation (0.7%).</div></div><div><h3>Conclusion</h3><div>There was a significant difference in the DVH metrics studied when tissue heterogeneities and patient-specific scattering are accounted for in high dose rate breast brachytherapy. However, for the studied patient cohort, the clinical coverage goal (PTV V_90%), had the smallest deviation.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100737"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143508690","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparative treatment planning of very high-energy electrons and photon volumetric modulated arc therapy: Optimising energy and beam parameters","authors":"Fabio S. D’Andrea ,&nbsp;Robert Chuter ,&nbsp;Adam H. Aitkenhead ,&nbsp;Ranald I. MacKay ,&nbsp;Roger M. Jones","doi":"10.1016/j.phro.2025.100732","DOIUrl":"10.1016/j.phro.2025.100732","url":null,"abstract":"<div><h3>Background</h3><div>Very High-Energy Electron (VHEE) beams offer potential advantages over current clinical radiotherapy modalities due to their precise dose targeting and minimal peripheral dose spread, which is ideal for treating deep-seated tumours. To aid the development of clinical VHEE machines, this study adressed the need to identify optimum VHEE beam characteristics for tumours across various anatomical sites.</div></div><div><h3>Materials and methods</h3><div>VHEE treatment planning employed matRad, an open-source treatment planning system, by adapting its proton pencil beam scanning implementation. VHEE beam characteristics were generated using TOPAS Monte Carlo simulations. A total of 820 plans were retrospectively created and analysed across 10 pelvic and 12 thoracic cases and compared against clinical photon VMAT plans to identify the most optimal VHEE beam configuration and energy requirement.</div></div><div><h3>Results</h3><div>VHEE plans outperformed photon VMAT in sparing organs-at-risk (OARs) while maintaining or improving target coverage. While 150 MeV served as the threshold for effectively treating deep-seated sites, 200 MeV was identified as a more optimal energy in the pelvis for achieving the best balance of penetration and sparing abutting OARs. Lower energies (70–110 MeV) also benefitted mid-to-superficial disease in the lung cohort. Typically, VHEE plans required 3–5 fields, and resulted in notable dose reductions to OARs across treatment sites, including: 22.5% reduction in rectal D_mean; 13.8% decrease in bladder D_mean; 8.2% reduction in heart D_mean; and a 24.4% decrease in lung V_20Gy.</div></div><div><h3>Conclusion</h3><div>The study reinforces VHEE’s potential in clinical settings, emphasising the need for varied energy ranges to enhance treatment flexibility and effectiveness.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100732"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143547972","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The prognostic value of pathologic lymph node imaging using deep learning-based outcome prediction in oropharyngeal cancer patients","authors":"Baoqiang Ma ,&nbsp;Alessia De Biase ,&nbsp;Jiapan Guo ,&nbsp;Lisanne V. van Dijk ,&nbsp;Johannes A. Langendijk ,&nbsp;Stefan Both ,&nbsp;Peter M.A. van Ooijen ,&nbsp;Nanna M. Sijtsema","doi":"10.1016/j.phro.2025.100733","DOIUrl":"10.1016/j.phro.2025.100733","url":null,"abstract":"<div><h3>Background and purpose</h3><div>Deep learning (DL) models can extract prognostic image features from pre-treatment PET/CT scans. The study objective was to explore the potential benefits of incorporating pathologic lymph node (PL) spatial information in addition to that of the primary tumor (PT) in DL-based models for predicting local control (LC), regional control (RC), distant-metastasis-free survival (DMFS), and overall survival (OS) in oropharyngeal cancer (OPC) patients.</div></div><div><h3>Materials and methods</h3><div>The study included 409 OPC patients treated with definitive (chemo)radiotherapy between 2010 and 2022. Patient data, including PET/CT scans, manually contoured PT (GTVp) and PL (GTVln) structures, clinical variables, and endpoints, were collected. Firstly, a DL-based method was employed to segment tumours in PET/CT, resulting in predicted probability maps for PT (TPMp) and PL (TPMln). Secondly, different combinations of CT, PET, manual contours and probability maps from 300 patients were used to train DL-based outcome prediction models for each endpoint through 5-fold cross validation. Model performance, assessed by concordance index (C-index), was evaluated using a test set of 100 patients.</div></div><div><h3>Results</h3><div>Including PL improved the C-index results for all endpoints except LC. For LC, comparable C-indices (around 0.66) were observed between models trained using only PT and those incorporating PL as additional structure. Models trained using PT and PL combined into a single structure achieved the highest C-index of 0.65 and 0.80 for RC and DMFS prediction, respectively. Models trained using these target structures as separate entities achieved the highest C-index of 0.70 for OS.</div></div><div><h3>Conclusion</h3><div>Incorporating lymph node spatial information improved the prediction performance for RC, DMFS, and OS.</div></div>","PeriodicalId":36850,"journal":{"name":"Physics and Imaging in Radiation Oncology","volume":"33 ","pages":"Article 100733"},"PeriodicalIF":3.4,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143437653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}