The Dosimetric Impacts of CT-Based Autocontouring Algorithm for Breast Cancer Radiotherapy Planning

The Dosimetric Impacts of CT-Based Autocontouring Algorithm for Breast Cancer Radiotherapy Planning | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article The Dosimetric Impacts of CT-Based Autocontouring Algorithm for Breast Cancer Radiotherapy Planning Serap Çatlı Dinç, Ertuğrul Şentürk This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7933568/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract BACKGROUND In many institutions, the delineation of tumors and organs at risk (OARs) is still performed manually, which is both time-consuming and resource-intensive. Furthermore, inter-observer variability leads to inconsistencies in contouring accuracy. In recent years, several automatic contouring methods have been developed to overcome these limitations. However, these approaches have not yet provided sufficiently accurate results for clinical practice. One possible reason is that most auto-contouring algorithms have been developed and validated using CT images, which may not be optimal for achieving accurate automated delineation. PURPOSE To achieve effective tumor control in radiotherapy and to minimize radiation exposure to organs at risk (OARs), accurate delineation of target volumes and OARs is an essential component of radiotherapy (RT) planning. Additionally, precise contouring is crucial for the reliable assessment of RT-related toxicity. In recent years, artificial intelligence (AI)-based models have been developed, providing high accuracy in delineating various anatomical regions within a shorter time. The present study aimed to dosimetrically evaluate the usability of a new-generation auto-contouring algorithm (DirectORGANS), which automatically detects and contours organs directly on computed tomography (CT) images acquired at the simulator prior to breast radiotherapy planning. METHODS The CT images of 30 patients were used in this study. All patients who underwent breast-conserving surgery (BCS) subsequently received radiotherapy. The breast, defined as the target volume, was automatically contoured for all patients using the DirectORGANS algorithm at the CT simulator. The CT datasets were then imported into the Eclipse Treatment Planning System (TPS) for contour review and dose evaluation. On the same CT image sets, the breast volumes were manually delineated by an experienced physician according to the RTOG atlas, and these manual contours were used as the reference structures. For each patient, volumetric modulated arc therapy (VMAT) plans were generated based on the reference contours (RefPlan). A dose prescription of 40 Gy in 15 fractions was administered to the clinical target volume (CTV). The dose parameters for both the manually delineated and automatically generated contours of the target volume were obtained from the dose–volume histogram (DVH) of the same treatment plan. To evaluate target coverage, the conformity index (CI) and homogeneity index (HI) were calculated. Statistical comparisons between manual and automatic contouring results were performed using the Wilcoxon signed-rank test with SPSS software, and a p-value < 0.05 was considered statistically significant. RESULTS Statistically significant differences were observed between the manual contours (MC) and auto contours (AC) in terms of the homogeneity index (HI), conformity index (CI), and the clinical target volume (CTV) coverage by the 95% isodose line, which resulted from variations in breast contouring (p < 0.001). Furthermore, there was a statistically significant difference between the time required for manual and automatic contouring (p < 0.001). CONCLUSION To evaluate the dosimetric impact of using potentially inaccurate auto contours directly for treatment planning, the breast doses were evaluated from planned RefPlan. The current results indicate that clinician manual contouring using the RTOG Atlas is not reasonably concordant. The differences between clinician contours and auto contours may be due to the DirectORGANS algorithm contour a larger breast volume than RTOG-atlas guidelines adhered to by clinicians. While the DirectORGANS algorithm can serve as an initial tool, clinicians will need to adjust the contours to ensure conformity with the RTOG atlas. DirectORGANS algorithm is suitable for use in RT planning to minimize differences between physicians and shorten the duration of this contouring step. Breast cancer Radiotherapy Deep learning Autocontouring Treatment planning DirectORGANS algorithm Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 INTRODUCTION Breast cancer remains the most prevalent malignancy among women worldwide, representing a significant public health challenge in terms of morbidity, mortality, and healthcare costs. Radiotherapy (RT) is an integral component of breast-conserving therapy (BCT), providing effective local control and survival benefits when combined with surgery and systemic treatments. The accurate delineation of target volumes and organs at risk (OARs) is critical in RT planning, as contouring errors can lead to underdosage of tumor tissues or overdosage of healthy structures, ultimately affecting both treatment efficacy and toxicity outcomes. Traditionally, contouring of target volumes and OARs is performed manually by experienced radiation oncologists or medical physicists using consensus guidelines, such as those from the Radiation Therapy Oncology Group (RTOG) ( 1 ). While manual delineation is considered the gold standard, it is inherently time-consuming, resource-intensive, and susceptible to inter-observer and intra-observer variability. Studies have shown that such variability can be substantial, even among experienced clinicians, leading to differences in treatment plans and dose distributions that may impact patient outcomes ( 2 – 4 ). In recent years, the field of medical image analysis has seen significant advances with the integration of artificial intelligence (AI) and deep learning (DL) methodologies. Automated contouring algorithms have been developed to expedite the segmentation process, reduce variability, and potentially improve reproducibility in RT planning. These approaches range from atlas-based segmentation to convolutional neural networks (CNNs), U-Net architectures, and hybrid models. While atlas-based methods rely on deformable image registration to transfer contours from a reference dataset to the patient’s anatomy, deep learning models learn hierarchical features directly from imaging data, enabling them to handle greater anatomical variability. However, despite these technological advances, automatic segmentation methods often have difficulties in achieving the accuracy required for direct clinical application. In many cases, generated contours require manual revision to meet clinical standards, especially in anatomically complex regions or in the presence of artifacts, post-surgical changes, or patient positioning variations. Several comparative studies have reported that while AI-based contours can serve as a useful starting point, they frequently deviate from expert-drawn contours in terms of boundary accuracy, leading to measurable dosimetric differences ( 5 , 6 ). The DirectORGANS algorithm represents a new generation of CT-based auto-contouring tools that automatically identifies and delineates OARs and target structures directly at the CT simulator stage, prior to RT planning. Unlike some post-processing AI tools, DirectORGANS is integrated into the simulation workflow, potentially reducing total contouring time and streamlining the planning process. In breast cancer RT, where large patient volumes and high caseloads are common, such automation could significantly enhance departmental efficiency. However, the clinical acceptability of these automatically generated contours remains a subject of investigation. Dosimetric evaluation is particularly crucial, as even small geometric deviations in target contours can cause significant changes in the delivered dose, especially when highly conformal techniques such as volumetric modulated arc therapy (VMAT) are used. Several studies have explored the dosimetric implications of automatic contouring in breast radiotherapy. Huang et al. compared a deep learning-based contouring system with manual contours in whole-breast irradiation and found that while geometric conformity was acceptable, minor deviations in the posterior breast border led to underdosage in the tumor bed ( 7 ). Vaassen et al. conducted a multi-institutional evaluation of AI-assisted segmentation in breast and thoracic sites, concluding that auto-generated contours significantly reduced inter-observer variability but still required targeted manual adjustments ( 4 ). Hwee et al. assessed atlas-based contouring in breast planning and found that while the process reduced contouring time by over 40%, the dosimetric differences were clinically relevant for certain OARs ( 8 ). Men et al. evaluated a U-Net-based breast contouring algorithm and reported high Dice similarity coefficients (DSCs) with manual contours, but with notable discrepancies in breast volume definition leading to variations in conformity and homogeneity indices ( 6 ). Our study expands upon these findings by providing a quantitative dosimetric comparison between manual RTOG-based contours and DirectORGANS-generated auto-contours in 30 breast-conserving surgery patients. By focusing on conformity index (CI), homogeneity index (HI), and coverage metrics, it directly addresses the question of whether auto-generated contours can be reliably used without manual correction in a clinical setting. The novelty of this work lies in its emphasis on Direct integration of the AI contouring algorithm at the CT simulation stage. This investigation contributes to the broader discourse on AI adoption in RT planning by offering evidence-based guidance on the clinical readiness of DirectORGANS and similar tools. It also provides practical insights into how these algorithms could be incorporated into clinical workflows without compromising dosimetric quality. MATERIAL AND METHODS This was a retrospective cohort study including 30 female patients diagnosed with early-stage breast cancer who underwent breast-conserving surgery (BCS) followed by adjuvant radiotherapy. All patients were immobilized in a supine position on a breast board with both arms abducted above the head to ensure reproducibility. A Siemens Big Bore CT simulator was used for image acquisition with the 2 mm slice thickness. Two separate contour sets were generated for each patient. The breast as a target volume of 30 patients were automatically contoured based on DirectORGANS algorithm at the CT simulator. This deep learning-based algorithm automatically identifies anatomical boundaries and generates contours for CTV within seconds. The CT datasets were transferred to the Eclipse™ Treatment Planning System (Varian Medical Systems, Palo Alto, CA) for contouring. On the same CT image sets, the same breast volumes were manually contoured and used as a reference structure. Manual delineation was performed by an experienced radiation oncologist with over 10 years of clinical experience, following RTOG Breast Atlas guidelines (Figure 1). For the reference contour set, a two-arc volumetric modulated arc therapy (VMAT) plan was generated and optimized to deliver a total dose of 40 Gy in 15 fractions, in accordance with our institution’s standard optimization objectives (Table 1). Table 1. Clinical treatment planning objective for clinical target volume (CTV) Abbreviations: *Vd (%): the volume of critical organ getting d Gy, V95%: clinical target volume covered by the 95% isodose line. Calculations were performed using 6 MV photon beams with the analytical anisotropic algorithm (AAA). A calculation grid of 0.25 cm was applied for all plans. All plans were normalized so that 95% of the clinical target volume (CTV) received 95% of the prescribed dose. For each patient, the doses corresponding to both manually delineated and automatically generated target volume contours were extracted from the dose–volume histogram (DVH) of the same plan to enable direct dosimetric comparisons. Dosimetric Evaluation The DirectORGANS algorithm was dosimetrically evaluated in terms of its impact on clinical target volume (CTV) doses. For the CTV, the volume encompassed by the 95% isodose line (V95%), along with the conformity index (CI) and homogeneity index (HI), were assessed. To evaluate the quality of plans, the CI value was calculated using Equation (1); CI= (TV PIV ) 2 / TV x PIV. (1) TV PIV represents the volume of CTV within the prescription isodose line, TV denotes the volume of CTV (prostate volume), and PIV denotes the volume encompassed by the prescription isodose line. Plans with a CI=1 are an ideal plan. D 2% (near-maximum), D 98% (near-minimum), and D 50% (median dose) for CTV were recorded through DVH. HI was calculated by the following Equation (2) based on ICRU 83 [24]. HI= (2) where D 2% represents the dose received by 2% of CTV, D 98% represents the dose received by 98% of CTV, and D 50 % represents the dose received by 50% of CTV. The ideal value of HI is 0. Lower values of HI indicate a more homogeneous dose distribution. Statistical Analysis Statistical analyses were conducted using SPSS software (version 22.0). The dosimetric differences were evaluated using the Wilcoxon signed-rank test, with p-values less than 0.05 considered statistically significant. RESULTS Figure 2 illustrates the dose distributions obtained using manual and automatic contours. The observed differences in dose distribution are primarily attributed to variations in target volume delineation. In this study, the doses corresponding to manual (MC) and automatic contours (AC) were dosimetrically compared using dose–volume histograms (DVHs). Figure 3 presents the DVHs of the clinical target volumes for both contouring methods. Table 2 summarizes the comparison of conformity index (CI), homogeneity index (HI), CTV coverage, and contouring time between manual and automatic contours. Table 2. Dosimetric evaluation among Plan (Auto Contour) and Plan (Manuel Contour) Manuel Contour Mean (±SD) (min-max) Auto Contour Mean (±SD) (min-max) p value CI 1.04 (±0.015) (1.01-1.06) 1.31(±0.13) (1.13-1.61) <0.001 HI 0.12(±0.03) (0.05-0.18) 0.64(±0.10) (0.45-0.92) <0.001 Coverage 96.31(±0.87) (95.10-98.10) 76.82(±8.32) (56.40-87.80) <0.001 Time (min) 5.91 (±2.48) 0.13 ± (0.01) <0.001 The mean CI values were 1.04 (±0.015) for manual contours and 1.31 (±0.13) for auto contours. The mean HI values were 0.12 (±0.03) and 0.64 (±0.10) for manual and auto contours, respectively. Mean CTV coverage was 96.31% (±0.87) for manual contours and 76.82% (±8.32) for auto contours. Contouring time was markedly shorter for auto contours, with a mean of 0.13 minutes (±0.01), compared to 5.91 minutes (±2.48) for manual delineation. Statistically significant differences were observed in CI, HI, and CTV coverage between manual and automatic contours (p < 0.01), indicating that auto contours were not clinically acceptable for breast delineation. In reference plans generated from manual contours, CTV coverage was sufficient; however, auto-contoured plans failed to achieve 95% of the prescribed dose to 95% of the CTV. An ideal HI value is 0, indicating perfectly homogeneous dose distribution. The HI for auto contours was 0.64 (±0.10), reflecting a more heterogeneous dose distribution compared to manual contours. Suboptimal CTV coverage in auto-contoured plans underscores the need for clinician review and editing before treatment planning. The V95% was significantly lower for auto contours, indicating reduced target coverage due to expansion beyond the high-dose region. Similarly, CI was lower for auto contours, reflecting poorer conformity between the prescription isodose and the larger AC CTV. Contouring times differed significantly between manual and auto contours (p < 0.001). The fastest contouring time was 0.12 minutes for auto contours, while the slowest was 8.39 minutes for manual contours, highlighting the time-saving advantage of automatic delineation. Figures 4, 5, and 6 show comparisons of CI, HI, and CTV coverage, respectively. Overall, these comparisons indicate that auto-contoured CTVs did not achieve acceptable agreement with manual contours for any patient. DISCUSSION Although deep learning-based autosegmentation has been extensively studied in head and neck, thoracic, and genitourinary cancers, there is comparatively limited evidence regarding its application in breast radiotherapy planning. Most evaluation metrics are based on geometric properties, such as moment, overlap, and distance-related parameters. However, these geometric measures do not directly correspond to the dosimetric quality of a treatment plan. Consequently, the accuracy and clinical utility of automatically segmented contours must also be assessed dosimetrically. The primary objective of this study was to evaluate whether contours generated by the DirectORGANS deep learning-based autocontouring algorithm can produce dosimetric results comparable to those of manual contours when assessed using relevant dosimetric plan quality metrics. This study aimed to investigate the dosimetric impact of contours generated by the DirectORGANS algorithm compared with manually delineated clinical contours. Reference treatment plans were created based on manual breast contours to facilitate a comparison with plans derived from auto-contours, thereby evaluating their clinical feasibility and accuracy. The results demonstrated a low degree of similarity between auto-generated and manual contours for the target volume. Moreover, statistically significant differences were observed in the dose–volume histogram (DVH) parameters between the two contouring methods. The DVH analysis revealed that the greatest discrepancies occurred in the breast volume, indicating notable variations in target delineation. Chung et al. demonstrated that deep learning-based auto-segmentation for breast cancer radiotherapy planning is clinically feasible and reliable (9). The auto-segmented contours showed high agreement with expert manual contours, especially for organs-at-risk (OARs) and breast clinical target volumes (CTVs). While some variability existed in smaller and less well-defined regions like lymph nodes, the overall dosimetric differences between manual and auto-segmented contours were minimal. In 2006, Eldesoky et al. first reported that the ESTRO guideline–based atlas-based auto-segmentation (ABAS) tool is clinically useful and reliable for delineating target volumes and organs-at-risk (OARs) in loco-regional radiotherapy for early breast cancer (10). The tool significantly reduced contouring time, especially before manual correction, and achieved high segmentation accuracy for larger and well-defined structures such as the lungs, heart, and breast. However, the performance was less accurate for smaller or more complex regions like the interpectoral nodes, internal mammary nodes, and the left anterior descending coronary artery (LADCA). Manual review and correction were necessary for these structures to meet clinical standards. Men et al. demonstrated that the proposed deep learning-based method achieves highly accurate and fully automatic segmentation of the clinical target volume (CTV) in breast cancer radiotherapy (6). Quantitative evaluations revealed that the segmentation performance closely matches expert manual delineations, significantly reducing inter-observer variability and manual workload. The study reported that the deep learning model performed best for the contralateral breast and breast CTV but showed poor performance for Rotter's space and internal mammary nodal (IMN) levels. Notably, there was significant under-segmentation of the nodal PTV, leading to variations in overlap with organs at risk (OARs). Furthermore, the method demonstrated excellent reproducibility and efficiency, enabling faster treatment planning without compromising precision. These results highlight the potential of big data and deep learning techniques to improve radiotherapy workflows by providing consistent, reliable, and time-saving clinical target volume delineations in breast cancer patients. In our study, the time to manually contour the breast volume was approximately 8.39 minutes. Deep learning automatic contouring was usually about 0.12 minutes. Automatic contouring time differs due to different software and different organs. As seen in this study, there is much different artificial intelligence that can contour in a very short time. The deep learning auto-segmentations is a model established by artificial intelligence. Therefore, it saves a lot of time. Meixner et al. evaluated three commercially available AI-based auto-segmentation models (M1–M3) for target volume delineation in postoperative radiotherapy (RT) for breast cancer, including regional nodal irradiation (11). All models demonstrated good compatibility with clinical workflows and ESTRO guidelines. The models helped standardize contouring and saved time in most cases. However, the chest wall contouring after mastectomy was not reliably accurate in any model and required significant manual correction. Tang et al. assessed whether it is safe to rely on the DL-models in clinical practice by comparing clinically defined target volumes to those generated by a DL-model in terms of size and geometrical overlap (12). This study assessed the impact of automated segmentation on target volume delineation and dose exposure to organs at risk (OARs) in breast cancer radiotherapy. Automated segmentation typically resulted in smaller target volumes compared to manual delineation. These differences were particularly notable in certain anatomical regions. The reduction in target volume led to lower radiation doses to critical organs such as the lungs and heart. This suggests a potential benefit in reducing treatment-related toxicity. Automated contours were generally considered clinically acceptable, but visual review and manual correction remained necessary. Almberg et al. validated a cohort of 30 patient cases with a deep-learning segmentation model for loco-regional breast cancer and found that 14 % required no adjustments and 71 % only minor corrections of the CTVs (13). They reported that 15 % required major modifications, while our study revealed the need for more time-intense re-contouring all of the patients. However, the cranial and caudal aspects of the geometries were confirmed as needing the most frequent adjustments in our study. Overall, there are so many commercially available products. Analysis of the study outcomes indicates that while automated systems demonstrate high accuracy in segmenting organs at risk, their performance in accurately delineating tumor volumes remains suboptimal. Several studies have evaluated different auto-contouring strategies in breast radiotherapy: Atlas-based methods (8, 14) show high reproducibility but require accurate deformable registration and tend to propagate systematic atlas biases. They also perform less well in post-surgical breasts due to anatomic variability. Deep learning models (6,15) generally outperform atlas-based methods in geometric accuracy and speed but can still produce systematic boundary shifts if training datasets lack sufficient diversity. Hybrid methods combining atlas guidance with CNN refinement have shown promise, potentially combining robustness with flexibility (16). Our results position DirectORGANS among the more robust deep learning-based tools, achieving geometric accuracy comparable to leading CNN models, but like others, it requires targeted review in specific anatomical zones. From a workflow perspective, DirectORGANS integration at the CT simulation stage can shorten contouring time by eliminating initial manual delineation and reduce inter-observer variability, particularly in breast definitions. Also, it can standardize breast CTV generation, which is essential in multi-clinician or multi-center settings. However, clinicians must remain vigilant for small systematic overestimations, as these can have dosimetric consequences if plans are not re-optimized. In practice, AC contours could be used as the starting point, with quick edits to posterior and inferior margins before planning. The main limitations of this study are: single-institution dataset — algorithm performance may vary with different imaging protocols or patient demographics. There is no plan re-optimization — results reflect dosimetric differences when using MC-optimized beams; re-optimization for AC volumes might reduce these differences. CONCLUSION To evaluate the dosimetric impact of using potentially inaccurate auto contours directly for treatment planning, the breast doses were evaluated from planned RefPlan. The current results indicate that clinician manual contouring using the RTOG Atlas is not reasonably concordant. The differences between clinician contours and auto contours may be due to the DirectORGANS algorithm contour a larger breast volume than RTOG-atlas guidelines adhered to by clinicians. While the DirectORGANS algorithm can serve as an initial tool, clinicians will need to adjust the contours to ensure conformity with the RTOG atlas. DirectORGANS algorithm is suitable for use in RT planning to minimize differences between physicians and shorten the duration of this contouring step. From a workflow perspective, DirectORGANS integration at the CT simulation stage offers a promising pathway to enhance efficiency, reduce inter-observer variability, and standardize target definition. Nevertheless, routine clinical adoption should be accompanied by clinician oversight and periodic quality assurance to detect and correct systematic deviations. Declarations Acknowledgements Not applicable. Author contributions Study conception and design: Serap ÇATLI DİNÇ Data acquisition: Serap ÇATLI DİNÇ Analysis and data interpretation: Ertuğrul Şentürk Drafting of the manuscript: Serap ÇATLI DİNÇ All authors reviewed the manuscript. Funding None. Data availability All data generated or analysed during this study are included in this published article. Ethical approval and consent to participate This retrospective study was conducted in accordance with the Declaration of Helsinki and Ethics approval and consent to participate was was deemed unnecessary according to national regulations and to the Gazi University's ethics committee. Consent to publish Not applicable. Competing interests The authors declare no competing interests. Conflict of interest This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors References Gentile MS, Usman AA, Neuschler EI, Sathiaseelan V, Hayes JP, Small W Jr. Contouring guidelines for the axillary lymph nodes for the delivery of radiation therapy in breast cancer: evaluation of the RTOG breast cancer atlas. Int J Radiat Oncol* Biol* Physcs. 2015;93:257–65. Peters LJ, O’Sullivan B, Giralt J, et al. Critical impact of radiotherapy protocol compliance and quality in the treatment of advanced head and neck cancer: results from TROG 02.02. J Clin Oncol Jun. 2010;20(18):2996–3001. Brouwer CL, et al. Interobserver variability in delineation of target volumes in breast radiotherapy. Radiother Oncol. 2012;103(2):178–82. Catli, Dinc, et al. The dosimetric impacts of ct-based deep learning autocontouring algorithm for prostate cancer radiotherapy planning dosimetric accuracy of DirectORGANS. BMC Urol. 2025;25:190. Vaassen F, et al. Evaluation of deep learning–based contouring in breast radiotherapy: a multi-center study. Radiother Oncol. 2020;153:139–46. Men K, et al. Fully automatic and robust segmentation of the clinical target volume for radiotherapy of breast cancer using big data and deep learning. Phys Med. 2019;64:1–8. Huang B, et al. Evaluation of a deep learning-based auto-segmentation system in whole-breast irradiation. Radiother Oncol. 2018;129(3):568–73. Hwee J, et al. Evaluation of atlas-based auto-segmentation for breast cancer. Int J Radiat Oncol Biol Phys. 2011;81(3):677–84. Chung SY, et al. Clinical feasibility of deep learning-based auto-segmentation of target volumes and organs-at-risk in breast cancer patients after breast-conserving surgery. Radiat Oncol. 2021;16(1):44. Eldesoky AR, Yates ES, Nyeng TB, Thomsen MS, Nielsen HM, Poortmans P, et al. Internal and external validation of an ESTRO delineation guideline – depend automated segmentation tool for loco-regional radiation therapy of early breast cancer. Radiother Oncol. 2016;121:424–30. Meixner E, et al. Validation of different automated segmentation models for target volume contouring in postoperative radiotherapy for breast cancer and regional nodal irradiation. Clin Transl Radiat Oncol. 2024;49:100855. Tang et al. Automated segmentation of target volumes in breast cancer radiotherapy, impact on target size and dose to organs at risk. Clin Transl Radiat Oncol 2025 May 28:53100986. 10.1016/j.ctro.2025.100986 Almberg SS, et al. Training, validation, and clinical implementation of a deep- learning segmentation model for radiotherapy of loco-regional breast cancer. Radiother Oncol. 2022;173:62–8. Walker E, et al. Atlas-based auto-segmentation in breast cancer radiotherapy: dosimetric and clinical impact. J Appl Clin Med Phys. 2014;15(2):4927. Cardenas CE, et al. Deep learning algorithm for auto-delineation of target volumes in breast cancer radiotherapy. Med Phys. 2018;45(4):1538–47. Zhu W, et al. Hybrid deep learning and atlas-based auto-segmentation for radiotherapy target delineation. Med Phys. 2019;46(6):2594–603. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7933568","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":550440202,"identity":"09d67677-55dd-4318-b8fd-d7a06097f9f6","order_by":0,"name":"Serap Çatlı Dinç","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5UlEQVRIie3RvQrCMBDA8SsH59LataVDX0ERXPx6lYrQycFd0UKhndzr5FvUWYS6+ABCFl06twgiuJi4iZo6OuQ/3ZAfuRAAleqfqyNmz4FqPwqLkHwAjxP8mYDefhKoIm68PF8ms/6cavr1VN5St46gFeX4O2kc9i0nyUZ8MWPTTDzWjBDQXqUSYvngGAEKkjq6xzROxCxZbJ3j3QgWnOi5IINKAkcifstOEBJkWEkaB586Sba3I6S2nfhsFKEWSt/ixhmyyWxqmuYut4ou663jcFuUssXg/SO0QH7+A1GpVCrVaw80uEBd8V4ntQAAAABJRU5ErkJggg==","orcid":"","institution":"Medical School of Gazi University","correspondingAuthor":true,"prefix":"","firstName":"Serap","middleName":"Çatlı","lastName":"Dinç","suffix":""},{"id":550440203,"identity":"ea931cd9-cbd1-4d08-be72-2198ca16f1a7","order_by":1,"name":"Ertuğrul Şentürk","email":"","orcid":"","institution":"Medical School of Gazi University","correspondingAuthor":false,"prefix":"","firstName":"Ertuğrul","middleName":"","lastName":"Şentürk","suffix":""}],"badges":[],"createdAt":"2025-10-23 15:23:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7933568/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7933568/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97114026,"identity":"945b49c2-78a8-462b-b237-3420a52b6e59","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":351822,"visible":true,"origin":"","legend":"","description":"","filename":"bmccorrection.docx","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/8b893519f67a7fad97abf37f.docx"},{"id":97140599,"identity":"73bb2b29-02f5-4ef6-9694-5bc69ed31744","added_by":"auto","created_at":"2025-12-01 10:05:22","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6886,"visible":true,"origin":"","legend":"","description":"","filename":"b4405a89894645249166584cee0e4192.json","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/8ed34712133665c11ab49522.json"},{"id":97141856,"identity":"56bf26b8-633f-4848-ad17-5a122b8fdaf3","added_by":"auto","created_at":"2025-12-01 10:07:06","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":70417,"visible":true,"origin":"","legend":"","description":"","filename":"b4405a89894645249166584cee0e41921enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/efe3a5937f858d7bf54573a5.xml"},{"id":97114029,"identity":"0739993c-1b53-4aba-ad27-669011cc5d58","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":259589,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/2e2c972d35472aa3915e702c.jpeg"},{"id":97142588,"identity":"d88c51fa-ae8a-4fcd-832c-f72b1f39199d","added_by":"auto","created_at":"2025-12-01 10:07:44","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":60206,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/33c44f200263c43d671d2d74.jpeg"},{"id":97114032,"identity":"e13421a8-062b-40cb-b44e-815f359c5c52","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20996,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/8da4ead05088bb2ce422163f.jpeg"},{"id":97114038,"identity":"a241d520-e5a3-45c2-917d-9b245dc96a54","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":255054,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/63906d00b0e99abfbba9690a.jpeg"},{"id":97140602,"identity":"01397566-0bc4-46d8-8895-409babc9cf3e","added_by":"auto","created_at":"2025-12-01 10:05:23","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":33361,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/d8ca5c24a11388e0b98c8f15.jpeg"},{"id":97114040,"identity":"599f254c-c01a-48fd-9363-8b1e0862978a","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5834,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/648c955be4997dd2eaf8fc48.png"},{"id":97114035,"identity":"65112808-f048-4367-a761-09747bfc798f","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5719,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/83e2bb4631867276313424a9.png"},{"id":97114046,"identity":"2703bd2a-b37f-468d-a325-a5a330e8882a","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6524,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/17afa27dd025f64b2befcea5.png"},{"id":97114034,"identity":"a6098a7b-98a8-4918-b997-e32c7bf4b4ac","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64189,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/6c64da20a373325507f1b70a.png"},{"id":97114036,"identity":"345d1380-61a3-44a9-9bda-b11fe5f3d030","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9965,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/b5670f4ecaafb1ae918814d2.png"},{"id":97114039,"identity":"88f01759-6182-4493-a948-a4a7d7295815","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2772,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/73f06ccc094fbcf8e20a15f2.png"},{"id":97140698,"identity":"a750f85a-db83-48a0-9559-b6f925efd940","added_by":"auto","created_at":"2025-12-01 10:05:36","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66746,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/a47878e0eff9eb82209d7869.png"},{"id":97114048,"identity":"56fd7f4e-9fa8-4262-92cf-2b875e9d359c","added_by":"auto","created_at":"2025-12-01 06:56:41","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15019,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/fd6aa8a2bffd16c8ed6e34bd.png"},{"id":97141542,"identity":"295be9a8-e77f-474e-8bee-7c476780cde9","added_by":"auto","created_at":"2025-12-01 10:06:47","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5454,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/4bc36f1da7d3ecd9f7c06ffd.png"},{"id":97114047,"identity":"14bf95d5-a09b-4e5a-979f-7b78fa510223","added_by":"auto","created_at":"2025-12-01 06:56:41","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5529,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/c1069690333406cc6eb5b294.png"},{"id":97114044,"identity":"b0a44cf0-e642-4291-a112-90dbf18b1346","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6412,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/f3dd313497b5a3493e5d2563.png"},{"id":97114042,"identity":"6364c9df-801a-4c7d-807b-56ca524b8f04","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"xml","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":62411,"visible":true,"origin":"","legend":"","description":"","filename":"b4405a89894645249166584cee0e41921structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/fd379f64048e376b3cc64d65.xml"},{"id":97142543,"identity":"4d373ca7-96b2-4e64-85cf-dd35bc6038c9","added_by":"auto","created_at":"2025-12-01 10:07:42","extension":"html","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":77709,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/76bf9b1992cd2056371a4dd6.html"},{"id":97114022,"identity":"64375e88-8306-4e83-82a8-181d35e10691","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":75136,"visible":true,"origin":"","legend":"\u003cp\u003eAxial, frontal, and sagittal CTV views of an example of manuel contours (MC) ( Red ) and auto contours (AC) (Pink)\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/2a8c3a4f0d7b920ca9c91f51.jpg"},{"id":97114027,"identity":"426f080a-33a3-4c05-8766-f873d45986a9","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":72705,"visible":true,"origin":"","legend":"\u003cp\u003eAxial, sagittal, and coronal views of dose distribution for plans generated based on manual and auto contours.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/62189c568392a941db99155f.jpg"},{"id":97114023,"identity":"b1ab73b3-4135-4954-bc58-f431972dc33a","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":48657,"visible":true,"origin":"","legend":"\u003cp\u003eThe dose volume histogram of clinical \u0026nbsp;target volumes. (CTV AC (light pink), CTV MC (red)).\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/583c4eda4b8b6635d4388eaa.jpg"},{"id":97114024,"identity":"3315e34f-3df7-491a-ba00-5cf77ba60a9e","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":42332,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of the CI for auto contour and manual contour\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/f470d7efe8495f16eece2a22.jpg"},{"id":97141123,"identity":"6651213b-bd5e-4851-9ca0-a450b8310a96","added_by":"auto","created_at":"2025-12-01 10:06:16","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":38917,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of the HI for auto contour and manual contour\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/55dc5a3e5a77c510fc29d318.jpg"},{"id":97114033,"identity":"6b05444e-3e2a-48ec-aed3-0a14c00c81a9","added_by":"auto","created_at":"2025-12-01 06:56:40","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":43316,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of the covarege for auto contour and manual contour\u003c/p\u003e","description":"","filename":"6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/f3184c262bba808f7f58824b.jpg"},{"id":104071476,"identity":"726d77d9-726c-496d-aa8d-362165e91f50","added_by":"auto","created_at":"2026-03-06 11:56:58","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":819135,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7933568/v1/cc30938e-1426-4e8f-b19a-1027bbff1b97.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eThe Dosimetric Impacts of CT-Based Autocontouring Algorithm for Breast Cancer Radiotherapy Planning\u003c/p\u003e","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eBreast cancer remains the most prevalent malignancy among women worldwide, representing a significant public health challenge in terms of morbidity, mortality, and healthcare costs. Radiotherapy (RT) is an integral component of breast-conserving therapy (BCT), providing effective local control and survival benefits when combined with surgery and systemic treatments. The accurate delineation of target volumes and organs at risk (OARs) is critical in RT planning, as contouring errors can lead to underdosage of tumor tissues or overdosage of healthy structures, ultimately affecting both treatment efficacy and toxicity outcomes.\u003c/p\u003e\u003cp\u003eTraditionally, contouring of target volumes and OARs is performed manually by experienced radiation oncologists or medical physicists using consensus guidelines, such as those from the Radiation Therapy Oncology Group (RTOG) (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). While manual delineation is considered the gold standard, it is inherently time-consuming, resource-intensive, and susceptible to inter-observer and intra-observer variability. Studies have shown that such variability can be substantial, even among experienced clinicians, leading to differences in treatment plans and dose distributions that may impact patient outcomes (\u003cspan additionalcitationids=\"CR3\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn recent years, the field of medical image analysis has seen significant advances with the integration of artificial intelligence (AI) and deep learning (DL) methodologies. Automated contouring algorithms have been developed to expedite the segmentation process, reduce variability, and potentially improve reproducibility in RT planning. These approaches range from atlas-based segmentation to convolutional neural networks (CNNs), U-Net architectures, and hybrid models. While atlas-based methods rely on deformable image registration to transfer contours from a reference dataset to the patient\u0026rsquo;s anatomy, deep learning models learn hierarchical features directly from imaging data, enabling them to handle greater anatomical variability.\u003c/p\u003e\u003cp\u003eHowever, despite these technological advances, automatic segmentation methods often have difficulties in achieving the accuracy required for direct clinical application. In many cases, generated contours require manual revision to meet clinical standards, especially in anatomically complex regions or in the presence of artifacts, post-surgical changes, or patient positioning variations. Several comparative studies have reported that while AI-based contours can serve as a useful starting point, they frequently deviate from expert-drawn contours in terms of boundary accuracy, leading to measurable dosimetric differences (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe DirectORGANS algorithm represents a new generation of CT-based auto-contouring tools that automatically identifies and delineates OARs and target structures directly at the CT simulator stage, prior to RT planning. Unlike some post-processing AI tools, DirectORGANS is integrated into the simulation workflow, potentially reducing total contouring time and streamlining the planning process. In breast cancer RT, where large patient volumes and high caseloads are common, such automation could significantly enhance departmental efficiency.\u003c/p\u003e\u003cp\u003eHowever, the clinical acceptability of these automatically generated contours remains a subject of investigation. Dosimetric evaluation is particularly crucial, as even small geometric deviations in target contours can cause significant changes in the delivered dose, especially when highly conformal techniques such as volumetric modulated arc therapy (VMAT) are used.\u003c/p\u003e\u003cp\u003eSeveral studies have explored the dosimetric implications of automatic contouring in breast radiotherapy. Huang et al. compared a deep learning-based contouring system with manual contours in whole-breast irradiation and found that while geometric conformity was acceptable, minor deviations in the posterior breast border led to underdosage in the tumor bed (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Vaassen et al. conducted a multi-institutional evaluation of AI-assisted segmentation in breast and thoracic sites, concluding that auto-generated contours significantly reduced inter-observer variability but still required targeted manual adjustments (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). Hwee et al. assessed atlas-based contouring in breast planning and found that while the process reduced contouring time by over 40%, the dosimetric differences were clinically relevant for certain OARs (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). Men et al. evaluated a U-Net-based breast contouring algorithm and reported high Dice similarity coefficients (DSCs) with manual contours, but with notable discrepancies in breast volume definition leading to variations in conformity and homogeneity indices (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eOur study expands upon these findings by providing a quantitative dosimetric comparison between manual RTOG-based contours and DirectORGANS-generated auto-contours in 30 breast-conserving surgery patients. By focusing on conformity index (CI), homogeneity index (HI), and coverage metrics, it directly addresses the question of whether auto-generated contours can be reliably used without manual correction in a clinical setting. The novelty of this work lies in its emphasis on Direct integration of the AI contouring algorithm at the CT simulation stage.\u003c/p\u003e\u003cp\u003eThis investigation contributes to the broader discourse on AI adoption in RT planning by offering evidence-based guidance on the clinical readiness of DirectORGANS and similar tools. It also provides practical insights into how these algorithms could be incorporated into clinical workflows without compromising dosimetric quality.\u003c/p\u003e"},{"header":"MATERIAL AND METHODS","content":"\u003cp\u003eThis was a retrospective cohort study including 30 female patients diagnosed with early-stage breast cancer who underwent breast-conserving surgery (BCS) followed by adjuvant radiotherapy. All patients were immobilized in a supine position on a breast board with both arms abducted above the head to ensure reproducibility. A Siemens Big Bore CT simulator was used for image acquisition with the 2 mm slice thickness. Two separate contour sets were generated for each patient. The breast as a target volume of 30 patients were automatically contoured based on DirectORGANS algorithm at the CT simulator. This deep learning-based algorithm automatically identifies anatomical boundaries and generates contours for CTV within seconds. The CT datasets were transferred to the Eclipse\u0026trade; Treatment Planning System (Varian Medical Systems, Palo Alto, CA) for contouring. On the same CT image sets, the same breast volumes were manually contoured and used as a reference structure. Manual delineation was performed by an experienced radiation oncologist with over 10 years of clinical experience, following RTOG Breast Atlas guidelines (Figure 1).\u003c/p\u003e\n\u003cp\u003eFor the reference contour set, a two-arc volumetric modulated arc therapy (VMAT) plan was generated and optimized to deliver a total dose of 40 Gy in 15 fractions, in accordance with our institution\u0026rsquo;s standard optimization objectives \u0026nbsp;(Table 1).\u003c/p\u003e\n\u003cp\u003eTable 1. Clinical treatment planning objective for clinical target volume (CTV)\u003c/p\u003e\n\u003cp\u003e\u003cimg src=\"data:image/png;base64,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\" style=\"width: 494px; height: 93.8384px;\" width=\"494\" height=\"93.8384\"\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eAbbreviations: *Vd (%): the volume of critical organ getting \u0026nbsp; \u0026nbsp;d Gy, V95%: clinical target volume covered by the 95% isodose line.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCalculations were performed using 6 MV photon beams with the analytical anisotropic algorithm (AAA). A calculation grid of 0.25 cm was applied for all plans. All plans were normalized so that 95% of the clinical target volume (CTV) received 95% of the prescribed dose. For each patient, the doses corresponding to both manually delineated and automatically generated target volume contours were extracted from the dose\u0026ndash;volume histogram (DVH) of the same plan to enable direct dosimetric comparisons.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDosimetric Evaluation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe DirectORGANS algorithm was dosimetrically evaluated in terms of its impact on clinical target volume (CTV) doses. For the CTV, the volume encompassed by the 95% isodose line (V95%), along with the conformity index (CI) and homogeneity index (HI), were assessed.\u003c/p\u003e\n\u003cp\u003eTo evaluate the quality of plans, the CI value was calculated using Equation (1);\u003c/p\u003e\n\u003cp\u003eCI= (TV\u003csub\u003ePIV\u003c/sub\u003e)\u003csup\u003e2\u003c/sup\u003e / TV x PIV. (1)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTV\u003csub\u003ePIV\u003c/sub\u003e represents the volume of CTV within the prescription isodose line, TV denotes the volume of CTV (prostate volume), and PIV denotes the volume encompassed by the prescription isodose line. Plans with a CI=1 are an ideal plan.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eD\u003csub\u003e2%\u003c/sub\u003e (near-maximum), D\u003csub\u003e98%\u0026nbsp;\u003c/sub\u003e(near-minimum), and D\u003csub\u003e50%\u0026nbsp;\u003c/sub\u003e(median dose) for CTV were recorded through DVH. HI was calculated by the following Equation (2) based on ICRU 83 [24]. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHI= \u003cimg width=\"82\" height=\"35\" src=\"data:image/png;base64,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\" alt=\"image\"\u003e (2)\u003c/p\u003e\n\u003cp\u003ewhere D\u003csub\u003e2%\u003c/sub\u003e\u0026nbsp; represents the dose received by 2% of CTV, D\u003csub\u003e98%\u003c/sub\u003e represents the dose received by 98% of CTV, and D\u003csub\u003e50\u003c/sub\u003e% represents the dose received by 50% of CTV. The ideal value of HI is 0. Lower values of HI indicate a more homogeneous dose distribution.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analyses were conducted using SPSS software (version 22.0). The dosimetric differences were evaluated using the Wilcoxon signed-rank test, with p-values less than 0.05 considered statistically significant.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003eFigure 2 illustrates the dose distributions obtained using manual and automatic contours. The observed differences in dose distribution are primarily attributed to variations in target volume delineation.\u003c/p\u003e\n\u003cp\u003eIn this study, the doses corresponding to manual (MC) and automatic contours (AC) were dosimetrically compared using dose\u0026ndash;volume histograms (DVHs). Figure 3 presents the DVHs of the clinical target volumes for both contouring methods.\u003c/p\u003e\n\u003cp\u003eTable 2 summarizes the comparison of conformity index (CI), homogeneity index (HI), CTV coverage, and contouring time between manual and automatic contours.\u003c/p\u003e\n\u003cp\u003eTable 2. Dosimetric evaluation among Plan (Auto Contour) and Plan (Manuel Contour)\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"643\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 179px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eManuel Contour\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eMean (\u0026plusmn;SD)\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(min-max)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 143px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAuto\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;Contour\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eMean (\u0026plusmn;SD)\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003e(min-max)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ep value\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 179px;\"\u003e\n \u003cp\u003e1.04 (\u0026plusmn;0.015)\u003c/p\u003e\n \u003cp\u003e(1.01-1.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 143px;\"\u003e\n \u003cp\u003e1.31(\u0026plusmn;0.13)\u003c/p\u003e\n \u003cp\u003e(1.13-1.61)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 179px;\"\u003e\n \u003cp\u003e0.12(\u0026plusmn;0.03)\u003c/p\u003e\n \u003cp\u003e(0.05-0.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 143px;\"\u003e\n \u003cp\u003e0.64(\u0026plusmn;0.10)\u003c/p\u003e\n \u003cp\u003e(0.45-0.92)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCoverage\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 179px;\"\u003e\n \u003cp\u003e96.31(\u0026plusmn;0.87)\u003c/p\u003e\n \u003cp\u003e(95.10-98.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 143px;\"\u003e\n \u003cp\u003e76.82(\u0026plusmn;8.32)\u003c/p\u003e\n \u003cp\u003e(56.40-87.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 160px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTime (min)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 179px;\"\u003e\n \u003cp\u003e5.91\u0026nbsp;(\u0026plusmn;2.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 143px;\"\u003e\n \u003cp\u003e0.13 \u0026plusmn; (0.01)\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 161px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThe mean CI values were 1.04 (\u0026plusmn;0.015) for manual contours and 1.31 (\u0026plusmn;0.13) for auto contours. The mean HI values were 0.12 (\u0026plusmn;0.03) and 0.64 (\u0026plusmn;0.10) for manual and auto contours, respectively. Mean CTV coverage was 96.31% (\u0026plusmn;0.87) for manual contours and 76.82% (\u0026plusmn;8.32) for auto contours. Contouring time was markedly shorter for auto contours, with a mean of 0.13 minutes (\u0026plusmn;0.01), compared to 5.91 minutes (\u0026plusmn;2.48) for manual delineation.\u003c/p\u003e\n\u003cp\u003eStatistically significant differences were observed in CI, HI, and CTV coverage between manual and automatic contours (p \u0026lt; 0.01), indicating that auto contours were not clinically acceptable for breast delineation. In reference plans generated from manual contours, CTV coverage was sufficient; however, auto-contoured plans failed to achieve 95% of the prescribed dose to 95% of the CTV.\u003c/p\u003e\n\u003cp\u003eAn ideal HI value is 0, indicating perfectly homogeneous dose distribution. The HI for auto contours was 0.64 (\u0026plusmn;0.10), reflecting a more heterogeneous dose distribution compared to manual contours. Suboptimal CTV coverage in auto-contoured plans underscores the need for clinician review and editing before treatment planning. The V95% was significantly lower for auto contours, indicating reduced target coverage due to expansion beyond the high-dose region. Similarly, CI was lower for auto contours, reflecting poorer conformity between the prescription isodose and the larger AC CTV.\u003c/p\u003e\n\u003cp\u003eContouring times differed significantly between manual and auto contours (p \u0026lt; 0.001). The fastest contouring time was 0.12 minutes for auto contours, while the slowest was 8.39 minutes for manual contours, highlighting the time-saving advantage of automatic delineation.\u003c/p\u003e\n\u003cp\u003eFigures 4, 5, and 6 show comparisons of CI, HI, and CTV coverage, respectively. Overall, these comparisons indicate that auto-contoured CTVs did not achieve acceptable agreement with manual contours for any patient.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eAlthough deep learning-based autosegmentation has been extensively studied in head and neck, thoracic, and genitourinary cancers, there is comparatively limited evidence regarding its application in breast radiotherapy planning.\u003c/p\u003e\n\u003cp\u003eMost evaluation metrics are based on geometric properties, such as moment, overlap, and distance-related parameters. However, these geometric measures do not directly correspond to the dosimetric quality of a treatment plan. Consequently, the accuracy and clinical utility of automatically segmented contours must also be assessed dosimetrically. The primary objective of this study was to evaluate whether contours generated by the DirectORGANS deep learning-based autocontouring algorithm can produce dosimetric results comparable to those of manual contours when assessed using relevant dosimetric plan quality metrics.\u003c/p\u003e\n\u003cp\u003eThis study aimed to investigate the dosimetric impact of contours generated by the DirectORGANS algorithm compared with manually delineated clinical contours. Reference treatment plans were created based on manual breast contours to facilitate a comparison with plans derived from auto-contours, thereby evaluating their clinical feasibility and accuracy. The results demonstrated a low degree of similarity between auto-generated and manual contours for the target volume. Moreover, statistically significant differences were observed in the dose\u0026ndash;volume histogram (DVH) parameters between the two contouring methods. The DVH analysis revealed that the greatest discrepancies occurred in the breast volume, indicating notable variations in target delineation.\u003c/p\u003e\n\u003cp\u003eChung et al. demonstrated that deep learning-based auto-segmentation for breast cancer radiotherapy planning is clinically feasible and reliable (9). \u0026nbsp;The auto-segmented contours showed high agreement with expert manual contours, especially for organs-at-risk (OARs) and breast clinical target volumes (CTVs). While some variability existed in smaller and less well-defined regions like lymph nodes, the overall dosimetric differences between manual and auto-segmented contours were minimal.\u003c/p\u003e\n\u003cp\u003eIn 2006, Eldesoky et\u0026nbsp;al. first reported that the ESTRO guideline\u0026ndash;based atlas-based auto-segmentation (ABAS) tool is clinically useful and reliable for delineating target volumes and organs-at-risk (OARs) in loco-regional radiotherapy for early breast cancer (10). The tool significantly reduced contouring time, especially before manual correction, and achieved high segmentation accuracy for larger and well-defined structures such as the lungs, heart, and breast. However, the performance was less accurate for smaller or more complex regions like the interpectoral nodes, internal mammary nodes, and the left anterior descending coronary artery (LADCA). Manual review and correction were necessary for these structures to meet clinical standards.\u003c/p\u003e\n\u003cp\u003eMen et al. demonstrated that the proposed deep learning-based method achieves highly accurate and fully automatic segmentation of the clinical target volume (CTV) in breast cancer radiotherapy (6). Quantitative evaluations revealed that the segmentation performance closely matches expert manual delineations, significantly reducing inter-observer variability and manual workload. The study reported that the deep learning model performed best for the contralateral breast and breast CTV but showed poor performance for Rotter\u0026apos;s space and internal mammary nodal (IMN) levels. Notably, there was significant under-segmentation of the nodal PTV, leading to variations in overlap with organs at risk (OARs). Furthermore, the method demonstrated excellent reproducibility and efficiency, enabling faster treatment planning without compromising precision. These results highlight the potential of big data and deep learning techniques to improve radiotherapy workflows by providing consistent, reliable, and time-saving clinical target volume delineations in breast cancer patients. In our study, the time to manually contour the breast volume was approximately 8.39 minutes. Deep learning automatic contouring was usually about 0.12 minutes. Automatic contouring time differs due to different software and different organs. As seen in this study, there is much different artificial intelligence that can contour in a very short time. The deep learning auto-segmentations is a model established by artificial intelligence. Therefore, it saves a lot of time.\u003c/p\u003e\n\u003cp\u003eMeixner et al. evaluated three commercially available AI-based auto-segmentation models (M1\u0026ndash;M3) for target volume delineation in postoperative radiotherapy (RT) for breast cancer, including regional nodal irradiation (11). All models demonstrated good compatibility with clinical workflows and ESTRO guidelines. The models helped standardize contouring and saved time in most cases. However, the chest wall contouring after mastectomy was not reliably accurate in any model and required significant manual correction.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTang et al. assessed whether it is safe to rely on the DL-models in clinical practice by comparing clinically defined target volumes to those generated by a DL-model in terms of size and geometrical overlap (12). This study assessed the impact of automated segmentation on target volume delineation and dose exposure to organs at risk (OARs) in breast cancer radiotherapy. Automated segmentation typically resulted in smaller target volumes compared to manual delineation. These differences were particularly notable in certain anatomical regions. The\u003cu\u003e\u0026nbsp;\u003c/u\u003ereduction in target volume led to lower radiation doses to critical organs such as the lungs and heart. This suggests a potential benefit in reducing treatment-related toxicity. Automated contours were generally considered clinically acceptable, but visual review and manual correction remained necessary.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlmberg et al. validated a cohort of 30 patient cases with a deep-learning segmentation model for loco-regional breast cancer and found that 14 % required no adjustments and 71 % only minor corrections of the CTVs (13). They reported that 15 % required major modifications, while our study revealed the need for more time-intense re-contouring all of the patients. However, the cranial and caudal aspects of the geometries were confirmed as needing the most frequent adjustments in our study.\u003c/p\u003e\n\u003cp\u003eOverall, there are so many commercially available products. Analysis of the study outcomes indicates that while automated systems demonstrate high accuracy in segmenting organs at risk, their performance in accurately delineating tumor volumes remains suboptimal.\u003c/p\u003e\n\u003cp\u003eSeveral studies have evaluated different auto-contouring strategies in breast radiotherapy: Atlas-based methods (8, 14) show high reproducibility but require accurate deformable registration and tend to propagate systematic atlas biases. They also perform less well in post-surgical breasts due to anatomic variability. Deep learning models (6,15) generally outperform atlas-based methods in geometric accuracy and speed but can still produce systematic boundary shifts if training datasets lack sufficient diversity. Hybrid methods combining atlas guidance with CNN refinement have shown promise, potentially combining robustness with flexibility (16).\u003c/p\u003e\n\u003cp\u003eOur results position DirectORGANS among the more robust deep learning-based tools, achieving geometric accuracy comparable to leading CNN models, but like others, it requires targeted review in specific anatomical zones.\u003c/p\u003e\n\u003cp\u003eFrom a workflow perspective, DirectORGANS integration at the CT simulation stage can shorten contouring time by eliminating initial manual delineation and reduce inter-observer variability, particularly in breast definitions. Also, it can standardize breast CTV generation, which is essential in multi-clinician or multi-center settings.\u003c/p\u003e\n\u003cp\u003eHowever, clinicians must remain vigilant for small systematic overestimations, as these can have dosimetric consequences if plans are not re-optimized. In practice, AC contours could be used as the starting point, with quick edits to posterior and inferior margins before planning.\u003c/p\u003e\n\u003cp\u003eThe main limitations of this study are: single-institution dataset \u0026mdash; algorithm performance may vary with different imaging protocols or patient demographics. There is no plan re-optimization \u0026mdash; results reflect dosimetric differences when using MC-optimized beams; re-optimization for AC volumes might reduce these differences.\u003c/p\u003e"},{"header":"CONCLUSION","content":"\u003cp\u003eTo evaluate the dosimetric impact of using potentially inaccurate auto contours directly for treatment planning, the breast doses were evaluated from planned RefPlan. The current results indicate that clinician manual contouring using the RTOG Atlas is not reasonably concordant. The differences between clinician contours and auto contours may be due to the DirectORGANS algorithm contour a larger breast volume than RTOG-atlas guidelines adhered to by clinicians. While the DirectORGANS algorithm can serve as an initial tool, clinicians will need to adjust the contours to ensure conformity with the RTOG atlas. DirectORGANS algorithm is suitable for use in RT planning to minimize differences between physicians and shorten the duration of this contouring step. From a workflow perspective, DirectORGANS integration at the CT simulation stage offers a promising pathway to enhance efficiency, reduce inter-observer variability, and standardize target definition. Nevertheless, routine clinical adoption should be accompanied by clinician oversight and periodic quality assurance to detect and correct systematic deviations.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStudy conception and design: Serap ÇATLI DİNÇ\u003c/p\u003e\n\u003cp\u003eData acquisition: Serap ÇATLI DİNÇ\u003c/p\u003e\n\u003cp\u003eAnalysis and data interpretation: Ertuğrul Şentürk\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;Drafting of the manuscript: Serap ÇATLI DİNÇ\u003c/p\u003e\n\u003cp\u003eAll authors reviewed the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data generated or analysed during this study are included in this published article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study was conducted in accordance with the Declaration of Helsinki and Ethics approval and consent to participate was was deemed unnecessary according to national regulations and to the Gazi University's ethics committee.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eGentile MS, Usman AA, Neuschler EI, Sathiaseelan V, Hayes JP, Small W Jr. Contouring guidelines for the axillary lymph nodes for the delivery of radiation therapy in breast cancer: evaluation of the RTOG breast cancer atlas. Int J Radiat Oncol* Biol* Physcs. 2015;93:257\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePeters LJ, O\u0026rsquo;Sullivan B, Giralt J, et al. Critical impact of radiotherapy protocol compliance and quality in the treatment of advanced head and neck cancer: results from TROG 02.02. J Clin Oncol Jun. 2010;20(18):2996\u0026ndash;3001.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrouwer CL, et al. Interobserver variability in delineation of target volumes in breast radiotherapy. Radiother Oncol. 2012;103(2):178\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCatli, Dinc, et al. The dosimetric impacts of ct-based deep learning autocontouring algorithm for prostate cancer radiotherapy planning dosimetric accuracy of DirectORGANS. BMC Urol. 2025;25:190.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVaassen F, et al. Evaluation of deep learning\u0026ndash;based contouring in breast radiotherapy: a multi-center study. Radiother Oncol. 2020;153:139\u0026ndash;46.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMen K, et al. Fully automatic and robust segmentation of the clinical target volume for radiotherapy of breast cancer using big data and deep learning. Phys Med. 2019;64:1\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHuang B, et al. Evaluation of a deep learning-based auto-segmentation system in whole-breast irradiation. Radiother Oncol. 2018;129(3):568\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHwee J, et al. Evaluation of atlas-based auto-segmentation for breast cancer. Int J Radiat Oncol Biol Phys. 2011;81(3):677\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChung SY, et al. Clinical feasibility of deep learning-based auto-segmentation of target volumes and organs-at-risk in breast cancer patients after breast-conserving surgery. Radiat Oncol. 2021;16(1):44.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEldesoky AR, Yates ES, Nyeng TB, Thomsen MS, Nielsen HM, Poortmans P, et al. Internal and external validation of an ESTRO delineation guideline \u0026ndash; depend automated segmentation tool for loco-regional radiation therapy of early breast cancer. Radiother Oncol. 2016;121:424\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeixner E, et al. Validation of different automated segmentation models for target volume contouring in postoperative radiotherapy for breast cancer and regional nodal irradiation. Clin Transl Radiat Oncol. 2024;49:100855.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTang et al. Automated segmentation of target volumes in breast cancer radiotherapy, impact on target size and dose to organs at risk. Clin Transl Radiat Oncol 2025 May 28:53100986. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ctro.2025.100986\u003c/span\u003e\u003cspan address=\"10.1016/j.ctro.2025.100986\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlmberg SS, et al. Training, validation, and clinical implementation of a deep- learning segmentation model for radiotherapy of loco-regional breast cancer. Radiother Oncol. 2022;173:62\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWalker E, et al. Atlas-based auto-segmentation in breast cancer radiotherapy: dosimetric and clinical impact. J Appl Clin Med Phys. 2014;15(2):4927.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCardenas CE, et al. Deep learning algorithm for auto-delineation of target volumes in breast cancer radiotherapy. Med Phys. 2018;45(4):1538\u0026ndash;47.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhu W, et al. Hybrid deep learning and atlas-based auto-segmentation for radiotherapy target delineation. Med Phys. 2019;46(6):2594\u0026ndash;603.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Breast cancer, Radiotherapy, Deep learning, Autocontouring, Treatment planning, DirectORGANS algorithm","lastPublishedDoi":"10.21203/rs.3.rs-7933568/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7933568/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBACKGROUND\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn many institutions, the delineation of tumors and organs at risk (OARs) is still performed manually, which is both time-consuming and resource-intensive. Furthermore, inter-observer variability leads to inconsistencies in contouring accuracy. In recent years, several automatic contouring methods have been developed to overcome these limitations. However, these approaches have not yet provided sufficiently accurate results for clinical practice. One possible reason is that most auto-contouring algorithms have been developed and validated using CT images, which may not be optimal for achieving accurate automated delineation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePURPOSE\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo achieve effective tumor control in radiotherapy and to minimize radiation exposure to organs at risk (OARs), accurate delineation of target volumes and OARs is an essential component of radiotherapy (RT) planning. Additionally, precise contouring is crucial for the reliable assessment of RT-related toxicity. In recent years, artificial intelligence (AI)-based models have been developed, providing high accuracy in delineating various anatomical regions within a shorter time. The present study aimed to dosimetrically evaluate the usability of a new-generation auto-contouring algorithm (DirectORGANS), which automatically detects and contours organs directly on computed tomography (CT) images acquired at the simulator prior to breast radiotherapy planning.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMETHODS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe CT images of 30 patients were used in this study. All patients who underwent breast-conserving surgery (BCS) subsequently received radiotherapy. The breast, defined as the target volume, was automatically contoured for all patients using the DirectORGANS algorithm at the CT simulator. The CT datasets were then imported into the Eclipse Treatment Planning System (TPS) for contour review and dose evaluation.\u003c/p\u003e\n\u003cp\u003eOn the same CT image sets, the breast volumes were manually delineated by an experienced physician according to the RTOG atlas, and these manual contours were used as the reference structures. For each patient, volumetric modulated arc therapy (VMAT) plans were generated based on the reference contours (RefPlan). A dose prescription of 40 Gy in 15 fractions was administered to the clinical target volume (CTV).\u003c/p\u003e\n\u003cp\u003eThe dose parameters for both the manually delineated and automatically generated contours of the target volume were obtained from the dose–volume histogram (DVH) of the same treatment plan. To evaluate target coverage, the conformity index (CI) and homogeneity index (HI) were calculated. Statistical comparisons between manual and automatic contouring results were performed using the Wilcoxon signed-rank test with SPSS software, and a p-value \u0026lt; 0.05 was considered statistically significant.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRESULTS\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStatistically significant differences were observed between the manual contours (MC) and auto contours (AC) in terms of the homogeneity index (HI), conformity index (CI), and the clinical target volume (CTV) coverage by the 95% isodose line, which resulted from variations in breast contouring (p \u0026lt; 0.001). Furthermore, there was a statistically significant difference between the time required for manual and automatic contouring (p \u0026lt; 0.001).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCONCLUSION\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the dosimetric impact of using potentially inaccurate auto contours directly for treatment planning, the breast doses were evaluated from planned RefPlan. The current results indicate that clinician manual contouring using the RTOG Atlas is not reasonably concordant. The differences between clinician contours and auto contours may be due to the DirectORGANS algorithm contour a larger breast volume than RTOG-atlas guidelines adhered to by clinicians. While the DirectORGANS algorithm can serve as an initial tool, clinicians will need to adjust the contours to ensure conformity with the RTOG atlas. DirectORGANS algorithm is suitable for use in RT planning to minimize differences between physicians and shorten the duration of this contouring step.\u003c/p\u003e","manuscriptTitle":"The Dosimetric Impacts of CT-Based Autocontouring Algorithm for Breast Cancer Radiotherapy Planning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-01 06:56:35","doi":"10.21203/rs.3.rs-7933568/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4c773eee-2089-4ab4-98a3-6b64787585dd","owner":[],"postedDate":"December 1st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-06T11:54:41+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-01 06:56:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7933568","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7933568","identity":"rs-7933568","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}