Factors Influencing the Accuracy and Coverage of CT-Based Lymph Node Delineation in Uterine Cervical Carcinoma: A Deep Learning Approach

Factors Influencing the Accuracy and Coverage of CT-Based Lymph Node Delineation in Uterine Cervical Carcinoma: A Deep Learning Approach | 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 Factors Influencing the Accuracy and Coverage of CT-Based Lymph Node Delineation in Uterine Cervical Carcinoma: A Deep Learning Approach Yi han Feng, Chun han Pan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7292070/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Purpose To analyze the factors influencing the accuracy and errors in the delineation of draining lymph nodes for uterine cervical carcinoma using a CT imaging diagnostic system based on two convolutional neural networks—GoogLeNet and Faster R-CNN. Material and methods A total of 679 lymph nodes in 56 patients with Uterine Cervical Carcinoma (UCC) below the renal hilar level, around the main lumen vessels, and in the pelvic cavity were delineated by the image diagnostic system. Then, two associate chief physicians in the imaging department evaluated the outlined lymph nodes to check for any missed lymph nodes. The lymph nodes were categorized into the following groups based on their basic characteristics: 1. Size. Based on previous research regarding the capability of convolutional neural networks (CNN) to recognize nodular nodes and the likelihood of lymph nodes appearing in images, they were classified into three groups: those with a length and diameter of less than 0.5 cm, between 0.5 cm and 2 cm, and greater than 2 cm; 2. Location. Considering the probability of metastatic lymph nodes in cervical cancer, they were divided into bilateral pelvic wall, common iliac vessels, retroperitoneum (common location group), and pelvic lymph nodes (uncommon location group); 3. Presence or absence of necrosis. We analyzed the coverage rate of lymph nodes identified by the imaging diagnostic system and its consistency with manual diagnoses. We analyzed the coverage rate of lymph nodes identified by the imaging diagnostic system and its consistency with manual diagnoses. Results A total of 679 lymph nodes were identified in this study, of which 281 were smaller than 0.5 cm, and 265 were outlined by the imaging diagnostic system. There were 312 lymph nodes measuring between 0.5 cm and 2 cm, of which 298 were outlined by the imaging diagnostic system. A total of 86 lymph nodes were larger than 2 cm, with 80 outlined. The diagnostic coverage rates for the three groups were 94.31%, 95.51%, and 93.02%, respectively, showing no statistically significant difference. Among the lymph nodes, 587 were located in common areas, with 575 outlined, resulting in a coverage rate of 97.96%. There were 92 unusual lymph nodes, of which 68 were outlined, yielding a coverage rate of 73.91% (p ≤ 0.01), indicating a statistically significant difference. There were 32 necrotic lymph nodes, with 23 outlined, resulting in a coverage rate of 71.88%. Additionally, there were 647 lymph nodes without necrosis, and 620 were outlined, yielding a coverage rate of 95.83% (p ≤ 0.01), indicating a statistically significant difference. Conclusion The imaging diagnostic system for (UCC) based on convolutional neural networks demonstrates a high degree of consistency with manual diagnoses. The coverage rate of lymph node delineation is significantly influenced by the location of lymph nodes and the presence or absence of necrosis, whereas the size of lymph nodes does not have a significant effect on the diagnostic system’s coverage rate. Therefore, the recognition capability of the diagnostic system for lymph nodes with necrosis and those in unusual locations should be enhanced to improve the overall coverage rate of diagnosis. Uterine cervical carcinoma (UCC) Convolutional neural networks (CNN) Image diagnosis system Computed tomography Figures Figure 1 Figure 2 Figure 3 Introduction Uterine cervical carcinoma (UCC) is a common malignancy among women, with its incidence ranking second only to breast cancer 1 . The primary treatment for UCC in early-stage patients is surgical resection. For patients who are not candidates for resection, radiotherapy is administered 2 . Both treatment strategies necessitate an accurate assessment of lymph node status prior to intervention. CT scans can effectively visualize the lymph nodes in the drainage area of UCC, serving as a routine examination method for patients with this condition 3 and providing the basis for delineating the target area for radiotherapy. With advancements in artificial intelligence, intelligent image analysis has emerged as a research hotspot 4 . Several mature artificial intelligence applications, such as those for lung nodule detection and rib fracture recognition, have been integrated into clinical practice, greatly facilitating clinicians’ work. This paper analyzes the accuracy and coverage rate of the UCC imaging diagnostic system based on convolutional neural networks in relation to lymph node delineation in CT tomographic image analysis, discussing its clinical application value and the factors influencing errors, thereby providing a reference for the enhancement of the diagnostic system. 1. Material and methods 1.1 Research object A total of 56 patients with cervical cancer, confirmed by biopsy or surgery, were enrolled from the Department of Gynecologic Radiotherapy and Gynecologic Oncology at Jiangsu Cancer Hospital between March and December 2017. The patients ranged in age from 22 to 73 years, with an average age of 43.2 ± 13.5 years. None had received radiotherapy or chemotherapy prior to the CT examination. 1.2 CT scanning method Scans were performed using either the GE Discovery CT 750 HDCT or the GE Hispeed NX/i scanner, both operating at a tube voltage of 120 kV. The Discovery CT 750 HDCT utilizes automatic milliampere-seconds technology, while the Hispeed NX/i scanner operates at a tube current of 400 mAs. The layer thickness and spacing were set to 5 mm, with a reconstruction layer thickness of 1 mm. A contrast agent of 100 ml iodohexol was administered at an injection rate of 2 ml/s. The scan covered the area from the upper margin of the diaphragm to the lower margin of the symphysis pubis. An abdominal window was employed for image observation, with a window width of 350 and a window position of 40. 1.3 The imaging diagnostic system delineated lymph nodes and facilitated artificial image observation (UCC imaging diagnostic system based on convolutional neural networks) In this system, the GoogLeNet network was employed for binary classification, while the Faster R-CNN was utilized for target detection, with lymph nodes delineated using boxes appropriate for their size. The outlined lymph node images were manually reviewed and documented. To ensure the accuracy of the manual review, two associate chief physicians from the imaging department conducted a double-blind review and recorded the lymph node length, diameter, location, and presence of necrosis after reaching a consensus. The lymph nodes were categorized into three groups based on the nodule recognition capability of the CNN system from previous studies and the clinical probability of lymph node occurrence according to size 5 , as well as into two groups based on common locations and the presence of necrosis. The diagnostic system mapped the number of correctly identified, incorrectly identified, and missed lymph nodes. 1.4 Statistical methods Statistical analysis was performed using SPSS 17.0 software to calculate the coverage of lymph nodes delineated in each group under various influencing factors, with the chi-square (χ²) test employed to assess differences in coverage among the groups. A p-value of < 0.05 was considered statistically significant. The Kappa test was utilized to analyze the consistency between the diagnostic system and manual film reading, with a p-value of < 0.05 indicating statistical significance. The sensitivity and specificity of the diagnostic system for lymph node delineation were calculated, and the receiver operating characteristic (ROC) curve was generated, with the area under the curve also calculated. 2. Results 2.1 There was no significant difference in the coverage rates of lymph node delineation among different sizes, and all lymph nodes could be identified and delineated by the imaging diagnostic system. However, the contour coverage of lymph nodes located between the main lumen vessels, on the left side of the abdominal aorta, and behind the bilateral external iliac vessels in the common drainage location group was significantly higher than that in the unusual location group, with a statistically significant difference between the two groups. The delineation coverage rate of the lymph node group with necrosis was lower than that of the group without necrosis, and this difference was statistically significant. The specific values and delineation coverage of lymph nodes are presented in Table 1 and Figure 1. 2.2 Consistency Analysis Results: The Kappa test indicated that the Kappa value for the consistency analysis between the diagnostic system and manual film reading was 0.813, demonstrating a high level of consistency (P < 0.01). Based on this standard, the sensitivity of the imaging diagnostic system for lymph node delineation was 94.70%, the specificity was 86.07%, and the area under the ROC curve was 0.904, as illustrated in Figure 3. Table 1：Delineation and coverage rates of 679 lymph nodes across different groups in 56 patients. System Factor Number of systemically identified lymph nodes/total number of lymph nodes Identification coverage rates 2 P Size 2cm 80/86 93.02% Location Common 575/587 97.96% 91.57 <0.001 Uncommon 68/92 73.91% Necrosis With 23/32 71.88% 34.84 <0.001 Without 620/647 95.83% Image diagnosis showed no statistically significant difference in contour coverage of lymph nodes of different sizes; however, it demonstrated statistical significance in contour coverage differences for lymph nodes located in different regions and those without necrosis (p < 0.01) 3. Discussion With advancements in science and technology, the application of artificial intelligence in medical diagnosis 6 has made significant progress 7 . The integration of machine learning and image processing technologies with computer-aided diagnosis has emerged as a prominent research focus in the medical field. Deep learning techniques can automatically learn high-dimensional features and internal relationships within datasets to facilitate data analysis and classification. Its performance typically surpasses that of traditional manually extracted image features 8 , making it an effective image analysis algorithm 9 that saves both manpower and time 10 . In this study, a CNN-based UCC image diagnosis system was selected. By utilizing extensive learning and recognition of UCC lymph node images, along with the process of re-evaluation and correction by radiologists, and simultaneously adjusting the relevant parameters during model development, target lymph nodes can be identified from the vast amount of information provided by CT images. The recognition coverage and accuracy of the system are highly dependent on the sample size during the initial learning phase; specifically, the more frequently the corresponding characteristic information of the learned lymph nodes appears, the higher the coverage and accuracy of the diagnostic system in recognizing similar lymph nodes. In this study, it was found that the diagnostic system failed to recognize lymph nodes primarily due to the following reasons: 1. Lymph nodes in atypical positions. The draining lymph nodes of UCC are predominantly located below the level of the renal hilum, situated between the vessels of the posterior peritoneal cavity and the left side and bilateral pelvic walls of the abdominal aorta 11 . Additionally, lymph nodes located on the right side or anterior to the inferior vena cava, as well as those near the internal iliac vessels and anterior to the external iliac vessels, have a low probability of occurrence. In this study, it was found that the diagnostic system failed to recognize lymph nodes primarily due to the following reasons: 1. Lymph nodes in atypical positions. The draining lymph nodes of UCC are predominantly located below the level of the renal hilum, situated between the vessels of the posterior peritoneal cavity and the left side and bilateral pelvic walls of the abdominal aorta. Additionally, lymph nodes located on the right side or anterior to the inferior vena cava, as well as those near the internal iliac vessels and anterior to the external iliac vessels, have a low probability of occurrence. The presence of complex lymph nodes in areas with numerous vascular cross-sections results in a relatively low identification rate for these lymph nodes within the imaging diagnostic system. Specifically, small lymph nodes may not be distinguishable amidst the vascular cross-sections, particularly when there is significant lymph node enhancement and minimal differences in vascular density values. Consequently, the diagnostic system struggles to identify lymph nodes based on both morphological and density characteristics. This limitation contributes to the system’s failure to recognize nodes in common locations. 3. UCC often presents at a later stage with a higher number of lymph nodes, resulting in the imaging diagnostic system being able to identify only a portion of the lymph nodes, primarily those in common locations, while others may be overlooked. The author retrospectively analyzed CT images from randomly selected patients and found that there were relatively few cases with obvious metastasis, potentially due to improved medical conditions and increased public awareness of health. Consequently, it is pertinent to conduct examinations in the early and middle stages of the disease in most cases 12. However, the sample size of images featuring a large number of lymph node metastases during the training of the imaging diagnostic system is insufficient. Thus, the results of this study indicated that the differences in diagnostic coverage of the lymph node imaging diagnostic system at various locations were statistically significant. 4. The coverage rate of the imaging diagnostic system for swollen lymph nodes with necrosis was low, and the difference in coverage rate was statistically significant when compared to that of the non-necrotic lymph node group. The author posited that necrotic lymph nodes were not recognized as lymph nodes due to their decreased internal density, leading the imaging diagnostic system to classify their density values as distinct from those of previously learned lymph nodes. In addition, during the study, the author observed that in the early stages of learning, the diagnostic system would mistakenly identify small vascular sections located in the drainage area of cervical cancer lymph nodes (i.e., the common location) with only subtle enhancement as lymph nodes. This misidentification occurred because these vascular sections are very similar to lymph nodes in terms of shape, density, and location, as illustrated in the images of Group A and Group C in Figure 1. However, as the sample size increased during the later stages of learning, the diagnostic system’s ability to distinguish the subtle differences between the two improved, effectively reducing the occurrence of such misjudgments. In this study, lymph nodes from patients with UCC were selected as the research subjects, as accurate delineation of lymph nodes is essential for both surgical and radiation therapy 13 . Subsequently, lymph node dissection or treatment should be guided by the local staging of UCC lesions, the presence of high-risk factors, the size of the lymph nodes, and their location in high-risk drainage areas. CT images effectively display lymph nodes in each drainage area of UCC 14 . The imaging department should first assess the presence of lymph nodes when manually interpreting the images. Subsequently, further evaluations should be made regarding their characteristics, including the long and short diameters, their ratio, shape, edge clarity, and uniformity of enhancement, for clinical reference 15 . The delineation principle for radiotherapy physicians is that if the short diameter exceeds 1 cm or if PET-CT results are positive, it qualifies for targeted irradiation 16 . Therefore, the short diameter of lymph nodes serves as a critical factor in determining treatment options. Consequently, the initial step in clinical management is to confirm the presence of lymph nodes, meaning that imaging examinations must effectively detect them. CT remains the primary imaging modality for cervical cancer patients due to its advantages of rapid examination speed, broad coverage, and relatively low cost, effectively aligning with the aforementioned clinical treatment principles. Currently, the image diagnosis system utilized in this study can accurately delineate lymph nodes of varying sizes within the drainage areas, aiding in the formulation of clinical treatment plans. This capability reduces the likelihood of overlooking lymph nodes when non-imaging specialists delineate target areas, thereby enhancing clinician efficiency and improving patient therapeutic outcomes. Senior imaging specialists can accurately identify lymph nodes in CT images, which can serve as a benchmark for assessing the coverage rate of the image diagnosis system. In summary, the convolutional neural network-based image diagnosis system for UCC can accurately delineate lymph nodes in UCC patients by simulating animal visual cognition functions and learning from previous cases. While it demonstrates a high degree of consistency with manual diagnoses, its coverage for recognizing relatively rare lymph nodes remains low. Expanding the sample size for training and adjusting model parameters are essential steps to enhance recognition coverage, enabling the image diagnosis system to serve as a valuable aid in clinical practice. Declarations Acknowledgements Not applicable. Author contributions Yi han Feng: Conceptualization, Methodology, Software, Investigation, Formal Analysis, Writing - Original Draft, Data Curation, Writing - Original Draft, Visualization, Investigation, Resources, Supervision, Software, Validation; Visualization, Writing -Review& Editing; Chun han Pan: Conceptualization, Funding Acquisition, Resources, Supervision, Writing -Review& Editing. Funding Not applicable. Data availability The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors. Ethics approval and consent to participate This retrospective study was approved by the Ethics Committee of Jiangsu Cancer Hospital and the methods were carried out in accordance with the approved guidelines. All the patients have been informed and signed informed consent before the experiments. Consent for publication Not applicable. Competing interests The authors declare that there is no competing interests regarding the publication of this article. Open access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/ licenses/ by/4. 0/. References MARQUINA G, MANZANO A, CASADO A. 2018. Targeted Agents in Cervical Cancer: Beyond Bevacizumab. Curr Oncol Rep , 20, pp.40. KUMAR L, HARISH P, MALIK PS, KHURANA S. Chemotherapy and targeted therapy in the management of cervical cancer. Curr Probl Cancer. 2018;42:120–8. LIU B, GAO S, LI S. A Comprehensive Comparison of CT, MRI, Positron Emission Tomography or Positron Emission Tomography/CT, and Diffusion Weighted Imaging-MRI for Detecting the Lymph Nodes Metastases in Patients with Cervical Cancer: A Meta-Analysis Based on 67 Studies. Gynecol Obstet Invest. 2017;82:209–22. YAMASHITA R, NISHIO M, DO, R. K. G., TOGASHI K. Convolutional neural networks: an overview and application in radiology. Insights Imaging. 2018;9:611–29. KAWAGUCHI Y, MATSUURA Y, KONDO Y, ICHINOSE J, NAKAO M, OKUMURA S, MUN M. The predictive power of artificial intelligence on mediastinal lymphnode metastasis. Gen Thorac Cardiovasc Surg. 2021;69:1545–52. DAS N, TOPALOVIC, M., JANSSENS W. Artificial intelligence in diagnosis of obstructive lung disease: current status and future potential. Curr Opin Pulm Med. 2018;24:117–23. WENG S, XU X, LI J, WONG STC. Combining deep learning and coherent anti-Stokes Raman scattering imaging for automated differential diagnosis of lung cancer. J Biomed Opt. 2017;22:1–10. LEE YH. Efficiency Improvement in a Busy Radiology Practice: Determination of Musculoskeletal Magnetic Resonance Imaging Protocol Using Deep-Learning Convolutional Neural Networks. J Digit Imaging. 2018;31:604–10. JUN Y, EO T, KIM T, SHIN H, HWANG D, PARK BAESH, LEE YW, CHOI HJ. B. W. & AHN, S. S. 2018. Deep-learned 3D black-blood imaging using automatic labelling technique and 3D convolutional neural networks for detecting metastatic brain tumors. Sci Rep , 8, pp.9450. CHANG P, BARDIS GRINBANDJWEINBERGBD, KHY M, CADENA M, CHA GSUMY, BOTA SFILIPPICG, BALDI D, JAIN PPOISSONLM, R., CHOW D. Deep-Learning Convolutional Neural Networks Accurately Classify Genetic Mutations in Gliomas. AJNR Am J Neuroradiol. 2018;39:1201–7. DU R, LI L, MA S, TAN X, ZHONG S, WU M. Lymph nodes metastasis in cervical cancer: Incidences, risk factors, consequences and imaging evaluations. Asia Pac J Clin Oncol. 2018;14:e380–5. MINIG L, SCAMBIA FAGOTTIA, SALVO G, PATRONO G, HAIDOPOULOS MG, ZAPARDIEL D, DOMINGO I, SOTIROPOULOU S, CHISHOLM M, G., RAMIREZ PT. Incidence of Lymph Node Metastases in Women With Low-Risk Early Cervical Cancer (< 2 cm) Without Lymph-Vascular Invasion. Int J Gynecol Cancer. 2018;28:788–93. JUNG W, LEE PARKKR, KIM KJ, LEE K, JEONG J, KIM S, KIM YJ, KANG JYOONHJ, KOO BC, CHO HSSUNGSH, M. S., PARK S. Value of imaging study in predicting pelvic lymph node metastases of uterine cervical cancer. Radiat Oncol J. 2017;35:340–8. DHAMIJA E, BABY A, BHATLA N, KUMAR PULAPPADIVP, KUMAR M, KUMAR S, L., SHARMA D. Radiological evaluation of metastatic lymph nodes in carcinoma cervix with emphasis on their infiltrative pattern. Indian J Med Res. 2021;154:383–90. CHEN J, DONG HEB, LIU D, DUAN P, LI H, LI W, WANG P, ZHANG LFANHWANGS, HUANG LTIANJ, Z., CHEN C. 2020. Noninvasive CT radiomic model for preoperative prediction of lymph node metastasis in early cervical carcinoma. Br J Radiol , 93, pp.20190558. BECKMANN MW, VORDERMARK STUEBSFA, KOCH D, HORN MC, L. C., FEHM T. The Diagnosis, Treatment, and Aftercare of Cervical Carcinoma. Dtsch Arztebl Int. 2021;118:806–12. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 12 Feb, 2026 Reviews received at journal 01 Dec, 2025 Reviewers agreed at journal 18 Nov, 2025 Reviews received at journal 12 Nov, 2025 Reviewers agreed at journal 08 Nov, 2025 Reviewers invited by journal 15 Sep, 2025 Editor invited by journal 10 Sep, 2025 Editor assigned by journal 29 Aug, 2025 Submission checks completed at journal 28 Aug, 2025 First submitted to journal 28 Aug, 2025 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-7292070","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":515036783,"identity":"ff3faa9f-655c-4281-8f69-64bdd4406387","order_by":0,"name":"Yi han Feng","email":"","orcid":"","institution":"The Affiliated Cancer Hospital of Nanjing Medical University, Jiangsu Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"han","lastName":"Feng","suffix":""},{"id":515036786,"identity":"e33cc7f4-4f84-4de0-bffb-4863ba676c03","order_by":1,"name":"Chun han Pan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1klEQVRIie3RIQsCMRTA8XcIswysJ+KJH0CYDIQD0a/yhnDJYDSeWLUr+CEGwuXBZJbZL57FZLFpEcWk6WYT3L//2Ht7AD7fD1YDvb+Kez8i1blyI/WZwbAgCa9Rg26EacvqBdFivRozx8mMZQxpIqS1l/wMg6iTlohgYSaIYZ/L/XIbb2DEe6qEVEItFbKkKe0ha1BQIisjpFV0U0QdyHx8ciMUdhxQ6fZzfeJGQjAJiPT1yTzeMIddhkqb4Ja+TnnMz9NBVEo+n6Sup3kj3wqfz+f7ix4SwErtYjfuKgAAAABJRU5ErkJggg==","orcid":"","institution":"The Affiliated Cancer Hospital of Nanjing Medical University, Jiangsu Institute of Cancer Research","correspondingAuthor":true,"prefix":"","firstName":"Chun","middleName":"han","lastName":"Pan","suffix":""}],"badges":[],"createdAt":"2025-08-04 14:08:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7292070/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7292070/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91897756,"identity":"31c41856-9a71-4e13-bc66-dea89518ccaf","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"doc","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1332736,"visible":true,"origin":"","legend":"","description":"","filename":"FactorsInfluencingtheAccuracyandCoverageofCTBasedLymphNodeDelineationinUterineCervicalCarcinomaADeepLearningApproach0806.doc","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/071452cecb091af70f405ce0.doc"},{"id":91898293,"identity":"8b3f81e0-b0ac-4973-8c50-b60c192b13cf","added_by":"auto","created_at":"2025-09-22 19:10:30","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5777,"visible":true,"origin":"","legend":"","description":"","filename":"3392f172294a4005acfbc6ef40248078.json","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/41aa21ca456f408af001a6aa.json"},{"id":91897767,"identity":"4b83312d-9b2c-41ea-aae7-2af6040903cf","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"zip","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1583223,"visible":true,"origin":"","legend":"","description":"","filename":"Figuresandtables.zip","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/ccf93ba1b8cb74346f4b964d.zip"},{"id":91898295,"identity":"a1af742b-8902-4be6-8eea-939d91f6f8f6","added_by":"auto","created_at":"2025-09-22 19:10:30","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59870,"visible":true,"origin":"","legend":"","description":"","filename":"3392f172294a4005acfbc6ef402480781enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/2f885f4a6e35d3e7bd8a6e08.xml"},{"id":91898396,"identity":"02fe4c24-fcbb-44c9-91f2-21c1de0513ab","added_by":"auto","created_at":"2025-09-22 19:18:30","extension":"eps","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44740,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/2ffa248eebce49e6e028c853.eps"},{"id":91897761,"identity":"b147bbd2-eb64-4c9b-8b26-1d2583cc4047","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"eps","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44532,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage2.eps","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/bf29ab72eeb13e468364ddbd.eps"},{"id":91898294,"identity":"31d3e9d4-e592-4547-8048-6bdefdefafbb","added_by":"auto","created_at":"2025-09-22 19:10:30","extension":"eps","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44108,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage3.eps","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/355a1c8a9dc29da6dfc2ad27.eps"},{"id":91897763,"identity":"d1ce3dc8-d7e9-4d19-b0ea-eed099ff8e1f","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"eps","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44166,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage4.eps","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/b8edc70038d3b1db1b79ee4e.eps"},{"id":91898065,"identity":"f1ad63f2-8608-4e2a-884e-ac526f3a7220","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"eps","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44746,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage5.eps","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/250dd5a5de5e97d0172be0f0.eps"},{"id":91897773,"identity":"2a5bed36-65a3-4fef-a9aa-61439d5f25da","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"eps","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44526,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage6.eps","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/b9a65a66fae84e12c395e023.eps"},{"id":91898060,"identity":"9f756f29-6a36-4156-b5df-72f49c8e6f74","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":88587,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/b58947b2d30329cb97820ec3.jpeg"},{"id":91898069,"identity":"88aa3a12-d9ec-445a-82a3-a4f4b7ec62a6","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"jpeg","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":91418,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/ac66ce1e6b9a1e0b7f4b4675.jpeg"},{"id":91897787,"identity":"004019ac-4232-4ea4-aa11-4100bcb53cb7","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"jpeg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102374,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/1efee3482ffa7b80d46f3017.jpeg"},{"id":91898399,"identity":"3cdbbd1b-c613-46c4-8b21-bf5e4ad9afdc","added_by":"auto","created_at":"2025-09-22 19:18:30","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":137747,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/eac4dd75e17edcc1a53d08cb.png"},{"id":91897783,"identity":"9650c9d6-c1cb-4813-ad2e-c096a59a44cf","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"jpeg","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94270,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/e2a35d581984a20d6445c509.jpeg"},{"id":91898672,"identity":"58ea3362-8f3a-4622-a8e9-3802d0c15f0b","added_by":"auto","created_at":"2025-09-22 19:26:30","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":153541,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/5fbd10cc4a62c7ed1e151db4.png"},{"id":91963682,"identity":"b9948ce8-be21-422c-9be5-c835a8503b5e","added_by":"auto","created_at":"2025-09-23 08:10:07","extension":"jpeg","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94994,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/f9ecf8097fb28728d5d4543e.jpeg"},{"id":91897785,"identity":"f66747ad-1d49-4d05-b75c-77e1cdfb9ac7","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":193318,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/108218d4576dc5a51ff39ce0.png"},{"id":91897789,"identity":"553a286a-2f07-427a-aa99-a394b514f782","added_by":"auto","created_at":"2025-09-22 18:54:31","extension":"jpeg","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":82787,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/8f685cfd05e163d0ad6f8830.jpeg"},{"id":91898301,"identity":"e20c8eee-851c-4b47-8210-314ed83aca68","added_by":"auto","created_at":"2025-09-22 19:10:30","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":173263,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/d713f443e50dfbb3b6d51e03.png"},{"id":91898062,"identity":"d08f09f2-6934-4b7f-907d-194eabe3d7e2","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"jpeg","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106237,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/443e3ef51cea8f6a237aab2b.jpeg"},{"id":91898296,"identity":"12f0d773-d8a8-4625-be1f-f75a9ec99ad0","added_by":"auto","created_at":"2025-09-22 19:10:30","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27592,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/a974ef7d8c5bd0d6ea18e926.png"},{"id":91897770,"identity":"7f87ce57-3d84-41a3-ab16-22e726423124","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":46539,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/93780207ed39a9a1285f71ee.png"},{"id":91897784,"identity":"530a4007-b0ce-4227-b237-f6038a4f3180","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":47339,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/96b683e3afa7c598edb5d526.png"},{"id":91898397,"identity":"10e982e8-09be-4e4e-9d5d-e75e3b80a6d1","added_by":"auto","created_at":"2025-09-22 19:18:30","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29876,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/763a474edf4c9c4139ecedf0.png"},{"id":91898674,"identity":"0257f7d9-ad7f-4271-b4e6-819c47811f72","added_by":"auto","created_at":"2025-09-22 19:26:30","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":52141,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/9ad3a1021b918d9f6e540218.png"},{"id":91897788,"identity":"a3036b30-12a0-4910-8c99-084d5b8352c6","added_by":"auto","created_at":"2025-09-22 18:54:31","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":40745,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/8d85b55bf5f5e72720b359e6.png"},{"id":91898297,"identity":"1939c66e-ead8-4b79-9e83-05e371e0dcdd","added_by":"auto","created_at":"2025-09-22 19:10:30","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":32004,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/c7c4d960f66440b90bc86683.png"},{"id":91897780,"identity":"e41f5b58-53c9-44bf-b680-8c471d804be2","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42365,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/0e20b0e78aeb0d03cbd2bbc5.png"},{"id":91897786,"identity":"2ea81b40-732e-4dbf-9f76-97fb5ead9f34","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43501,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/6789032b6742baa8ff274419.png"},{"id":91898074,"identity":"11502eb2-3652-403f-8850-941e276ef376","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41211,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/4a4c04720990d4c278dc55f5.png"},{"id":91897792,"identity":"b29fe9db-f083-46ad-bd74-9bb1fff5557b","added_by":"auto","created_at":"2025-09-22 18:54:31","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":58200,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/644ef3b4884889b587e42eac.png"},{"id":91898076,"identity":"f2e58d99-8435-4532-ad4c-8aca82918c95","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"xml","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":57626,"visible":true,"origin":"","legend":"","description":"","filename":"3392f172294a4005acfbc6ef402480781structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/57b8583271d21d1b9cad1af6.xml"},{"id":91897791,"identity":"a648a11c-db63-4b95-8d90-0b8106db3e06","added_by":"auto","created_at":"2025-09-22 18:54:31","extension":"html","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":65442,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/3d4e84fc5e1e6831b4d1c042.html"},{"id":91898057,"identity":"fecdf206-ea48-4173-9120-8a5d06a1228a","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":955970,"visible":true,"origin":"","legend":"\u003cp\u003eFemale, 46 years old, demonstrating both incorrectly and correctly delineated lymph nodes by the image diagnostic system. The red box on the left column is the lymph node situation outlined by the image diagnosis system, and the green box on the right column is the same CT image corresponding to the left column, and the correct lymph node picture outlined by the deputy chief physician of the imaging department. The Group A diagnostic system mistook the small vessels on the left side of the abdominal aorta for lymph nodes. The Group B diagnostic system failed to recognize a small lymph node behind the larger retroperitoneal lymph node. The diagnostic system in group C mistaken the small vessels in the left pelvic wall for small lymph nodes. There were several lymph nodes in the left pelvic wall of group D, only one of which was delineated by the diagnostic system.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/34bc1b9649a9970b5b098bbd.jpg"},{"id":91897759,"identity":"083e3ce3-a571-4043-9813-b74f03719cfb","added_by":"auto","created_at":"2025-09-22 18:54:30","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":439410,"visible":true,"origin":"","legend":"\u003cp\u003eFemale, 52 years old, evaluated using an imaging diagnostic system that successfully delineated the correct lymph nodes. “A” represents the outline of the retroperitoneal lymph nodes, while “B” indicates the outline of the bilateral pelvic wall lymph nodes.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/ca69472d7d35e610c7459cbb.jpg"},{"id":91898058,"identity":"30c38807-8691-4fc5-9071-fed6b54b38b4","added_by":"auto","created_at":"2025-09-22 19:02:30","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":137033,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver Operating Characteristic (ROC) curve of the imaging diagnostic system.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/b6fd0134510bbc1c15c909f8.jpg"},{"id":91964491,"identity":"d8a52831-d926-4192-b938-76a1e9c18821","added_by":"auto","created_at":"2025-09-23 08:11:56","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1945665,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7292070/v1/d6c02919-97e0-430c-911b-3ed6663402d9.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Factors Influencing the Accuracy and Coverage of CT-Based Lymph Node Delineation in Uterine Cervical Carcinoma: A Deep Learning Approach","fulltext":[{"header":"Introduction","content":"\u003cp\u003eUterine cervical carcinoma (UCC) is a common malignancy among women, with its incidence ranking second only to breast cancer\u003csup\u003e1\u003c/sup\u003e. The primary treatment for UCC in early-stage patients is surgical resection. For patients who are not candidates for resection, radiotherapy is administered\u003csup\u003e2\u003c/sup\u003e. Both treatment strategies necessitate an accurate assessment of lymph node status prior to intervention. CT scans can effectively visualize the lymph nodes in the drainage area of UCC, serving as a routine examination method for patients with this condition\u003csup\u003e3\u003c/sup\u003e and providing the basis for delineating the target area for radiotherapy. With advancements in artificial intelligence, intelligent image analysis has emerged as a research hotspot\u003csup\u003e4\u003c/sup\u003e. Several mature artificial intelligence applications, such as those for lung nodule detection and rib fracture recognition, have been integrated into clinical practice, greatly facilitating clinicians’ work. This paper analyzes the accuracy and coverage rate of the UCC imaging diagnostic system based on convolutional neural networks in relation to lymph node delineation in CT tomographic image analysis, discussing its clinical application value and the factors influencing errors, thereby providing a reference for the enhancement of the diagnostic system.\u003c/p\u003e"},{"header":"1. Material and methods","content":"\u003cp\u003e1.1 Research object\u003c/p\u003e\n\u003cp\u003eA total of 56 patients with cervical cancer, confirmed by biopsy or surgery, were enrolled from the Department of Gynecologic Radiotherapy and Gynecologic Oncology at Jiangsu Cancer Hospital between March and December 2017. The patients ranged in age from 22 to 73 years, with an average age of 43.2 ± 13.5 years. None had received radiotherapy or chemotherapy prior to the CT examination.\u003c/p\u003e\n\u003cp\u003e1.2 CT scanning method\u003c/p\u003e\n\u003cp\u003eScans were performed using either the GE Discovery CT 750 HDCT or the GE Hispeed NX/i scanner, both operating at a tube voltage of 120 kV. The Discovery CT 750 HDCT utilizes automatic milliampere-seconds technology, while the Hispeed NX/i scanner operates at a tube current of 400 mAs. The layer thickness and spacing were set to 5 mm, with a reconstruction layer thickness of 1 mm. A contrast agent of 100 ml iodohexol was administered at an injection rate of 2 ml/s. The scan covered the area from the upper margin of the diaphragm to the lower margin of the symphysis pubis. An abdominal window was employed for image observation, with a window width of 350 and a window position of 40.\u003c/p\u003e\n\u003cp\u003e1.3 The imaging diagnostic system delineated lymph nodes and facilitated artificial image observation (UCC imaging diagnostic system based on convolutional neural networks)\u003c/p\u003e\n\u003cp\u003eIn this system, the GoogLeNet network was employed for binary classification, while the Faster R-CNN was utilized for target detection, with lymph nodes delineated using boxes appropriate for their size. The outlined lymph node images were manually reviewed and documented. To ensure the accuracy of the manual review, two associate chief physicians from the imaging department conducted a double-blind review and recorded the lymph node length, diameter, location, and presence of necrosis after reaching a consensus. The lymph nodes were categorized into three groups based on the nodule recognition capability of the CNN system from previous studies and the clinical probability of lymph node occurrence according to size\u003csup\u003e5\u003c/sup\u003e, as well as into two groups based on common locations and the presence of necrosis. The diagnostic system mapped the number of correctly identified, incorrectly identified, and missed lymph nodes.\u003c/p\u003e\n\u003cp\u003e1.4 Statistical methods\u003c/p\u003e\n\u003cp\u003eStatistical analysis was performed using SPSS 17.0 software to calculate the coverage of lymph nodes delineated in each group under various influencing factors, with the chi-square (χ²) test employed to assess differences in coverage among the groups. A p-value of \u0026lt; 0.05 was considered statistically significant. The Kappa test was utilized to analyze the consistency between the diagnostic system and manual film reading, with a p-value of \u0026lt; 0.05 indicating statistical significance. The sensitivity and specificity of the diagnostic system for lymph node delineation were calculated, and the receiver operating characteristic (ROC) curve was generated, with the area under the curve also calculated.\u003c/p\u003e\n\n\n\n\n\n"},{"header":"2. Results","content":"\u003cp\u003e2.1 \u0026nbsp;There was no significant difference in the coverage rates of lymph node delineation among different sizes, and all lymph nodes could be identified and delineated by the imaging diagnostic system. However, the contour coverage of lymph nodes located between the main lumen vessels, on the left side of the abdominal aorta, and behind the bilateral external iliac vessels in the common drainage location group was significantly higher than that in the unusual location group, with a statistically significant difference between the two groups. The delineation coverage rate of the lymph node group with necrosis was lower than that of the group without necrosis, and this difference was statistically significant. The specific values and delineation coverage of lymph nodes are presented in Table 1 and Figure 1.\u003c/p\u003e\n\u003cp\u003e2.2 \u0026nbsp;Consistency Analysis Results: The Kappa test indicated that the Kappa value for the consistency analysis between the diagnostic system and manual film reading was 0.813, demonstrating a high level of consistency (P \u0026lt; 0.01). Based on this standard, the sensitivity of the imaging diagnostic system for lymph node delineation was 94.70%, the specificity was 86.07%, and the area under the ROC curve was 0.904, as illustrated in Figure 3.\u003c/p\u003e\n\u003cp\u003eTable 1：Delineation and coverage rates of 679 lymph nodes across different groups in 56 patients.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"587\" class=\"fr-table-selection-hover\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 73px;\"\u003e\n \u003cp\u003eSystem \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 18.7394%;\"\u003e\n \u003cp\u003eFactor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003eNumber of systemically identified lymph nodes/total number of lymph nodes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003eIdentification coverage rates\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003csup\u003e2\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003eSize\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003e\u0026lt; 0.5 cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e265/281\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e94.31%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e0.61\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003e0.5-2 cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e298/312\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e95.51%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003e\u0026gt; 2cm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e80/86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e93.02%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003eLocation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003eCommon\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e575/587\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e97.96%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e91.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003eUncommon\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e68/92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e73.91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003eNecrosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003eWith\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e23/32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e71.88%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e34.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\" style=\"width: 83px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 17.0358%;\"\u003e\n \u003cp\u003eWithout\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 27.9386%;\"\u003e\n \u003cp\u003e620/647\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 84px;\"\u003e\n \u003cp\u003e95.83%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 69px;\"\u003e\n \u003cp\u003e \u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eImage diagnosis showed no statistically significant difference in contour coverage of lymph nodes of different sizes; however, it demonstrated statistical significance in contour coverage differences for lymph nodes located in different regions and those without necrosis (p \u0026lt; 0.01)\u003c/p\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eWith advancements in science and technology, the application of artificial intelligence in medical diagnosis\u003csup\u003e6\u003c/sup\u003e has made significant progress\u003csup\u003e7\u003c/sup\u003e. The integration of machine learning and image processing technologies with computer-aided diagnosis has emerged as a prominent research focus in the medical field. Deep learning techniques can automatically learn high-dimensional features and internal relationships within datasets to facilitate data analysis and classification. Its performance typically surpasses that of traditional manually extracted image features\u003csup\u003e8\u003c/sup\u003e, making it an effective image analysis algorithm\u003csup\u003e9\u003c/sup\u003e that saves both manpower and time\u003csup\u003e10\u003c/sup\u003e. In this study, a CNN-based UCC image diagnosis system was selected. By utilizing extensive learning and recognition of UCC lymph node images, along with the process of re-evaluation and correction by radiologists, and simultaneously adjusting the relevant parameters during model development, target lymph nodes can be identified from the vast amount of information provided by CT images.\u0026nbsp;The recognition coverage and accuracy of the system are highly dependent on the sample size during the initial learning phase; specifically, the more frequently the corresponding characteristic information of the learned lymph nodes appears, the higher the coverage and accuracy of the diagnostic system in recognizing similar lymph nodes.\u003c/p\u003e\n\u003cp\u003eIn this study, it was found that the diagnostic system failed to recognize lymph nodes primarily due to the following reasons: 1. Lymph nodes in atypical positions. The draining lymph nodes of UCC are predominantly located below the level of the renal hilum, situated between the vessels of the posterior peritoneal cavity and the left side and bilateral pelvic walls of the abdominal aorta\u003csup\u003e11\u003c/sup\u003e. Additionally, lymph nodes located on the right side or anterior to the inferior vena cava, as well as those near the internal iliac vessels and anterior to the external iliac vessels, have a low probability of occurrence.\u0026nbsp;In this study, it was found that the diagnostic system failed to recognize lymph nodes primarily due to the following reasons: 1. Lymph nodes in atypical positions. The draining lymph nodes of UCC are predominantly located below the level of the renal hilum, situated between the vessels of the posterior peritoneal cavity and the left side and bilateral pelvic walls of the abdominal aorta. Additionally, lymph nodes located on the right side or anterior to the inferior vena cava, as well as those near the internal iliac vessels and anterior to the external iliac vessels, have a low probability of occurrence. The presence of complex lymph nodes in areas with numerous vascular cross-sections results in a relatively low identification rate for these lymph nodes within the imaging diagnostic system. Specifically, small lymph nodes may not be distinguishable amidst the vascular cross-sections, particularly when there is significant lymph node enhancement and minimal differences in vascular density values. Consequently, the diagnostic system struggles to identify lymph nodes based on both morphological and density characteristics. This limitation contributes to the system’s failure to recognize nodes in common locations.\u0026nbsp;3. UCC often presents at a later stage with a higher number of lymph nodes, resulting in the imaging diagnostic system being able to identify only a portion of the lymph nodes, primarily those in common locations, while others may be overlooked. The author retrospectively analyzed CT images from randomly selected patients and found that there were relatively few cases with obvious metastasis, potentially due to improved medical conditions and increased public awareness of health. Consequently, it is pertinent to conduct examinations in the early and middle stages of the disease in most cases\u003csup\u003e12.\u003c/sup\u003e However, the sample size of images featuring a large number of lymph node metastases during the training of the imaging diagnostic system is insufficient. Thus, the results of this study indicated that the differences in diagnostic coverage of the lymph node imaging diagnostic system at various locations were statistically significant. 4. The coverage rate of the imaging diagnostic system for swollen lymph nodes with necrosis was low, and the difference in coverage rate was statistically significant when compared to that of the non-necrotic lymph node group. The author posited that necrotic lymph nodes were not recognized as lymph nodes due to their decreased internal density, leading the imaging diagnostic system to classify their density values as distinct from those of previously learned lymph nodes. In addition, during the study, the author observed that in the early stages of learning, the diagnostic system would mistakenly identify small vascular sections located in the drainage area of cervical cancer lymph nodes (i.e., the common location) with only subtle enhancement as lymph nodes. This misidentification occurred because these vascular sections are very similar to lymph nodes in terms of shape, density, and location, as illustrated in the images of Group A and Group C in Figure 1. However, as the sample size increased during the later stages of learning, the diagnostic system’s ability to distinguish the subtle differences between the two improved, effectively reducing the occurrence of such misjudgments.\u003c/p\u003e\n\u003cp\u003eIn this study, lymph nodes from patients with UCC were selected as the research subjects, as accurate delineation of lymph nodes is essential for both surgical and radiation therapy\u003csup\u003e13\u003c/sup\u003e. Subsequently, lymph node dissection or treatment should be guided by the local staging of UCC lesions, the presence of high-risk factors, the size of the lymph nodes, and their location in high-risk drainage areas. CT images effectively display lymph nodes in each drainage area of UCC\u003csup\u003e14\u003c/sup\u003e. The imaging department should first assess the presence of lymph nodes when manually interpreting the images. Subsequently, further evaluations should be made regarding their characteristics, including the long and short diameters, their ratio, shape, edge clarity, and uniformity of enhancement, for clinical reference\u003csup\u003e15\u003c/sup\u003e. The delineation principle for radiotherapy physicians is that if the short diameter exceeds 1 cm or if PET-CT results are positive, it qualifies for targeted irradiation\u003csup\u003e16\u003c/sup\u003e. Therefore, the short diameter of lymph nodes serves as a critical factor in determining treatment options. Consequently, the initial step in clinical management is to confirm the presence of lymph nodes, meaning that imaging examinations must effectively detect them. CT remains the primary imaging modality for cervical cancer patients due to its advantages of rapid examination speed, broad coverage, and relatively low cost, effectively aligning with the aforementioned clinical treatment principles. Currently, the image diagnosis system utilized in this study can accurately delineate lymph nodes of varying sizes within the drainage areas, aiding in the formulation of clinical treatment plans. This capability reduces the likelihood of overlooking lymph nodes when non-imaging specialists delineate target areas, thereby enhancing clinician efficiency and improving patient therapeutic outcomes. Senior imaging specialists can accurately identify lymph nodes in CT images, which can serve as a benchmark for assessing the coverage rate of the image diagnosis system.\u003c/p\u003e\n\u003cp\u003eIn summary, the convolutional neural network-based image diagnosis system for UCC can accurately delineate lymph nodes in UCC patients by simulating animal visual cognition functions and learning from previous cases. While it demonstrates a high degree of consistency with manual diagnoses, its coverage for recognizing relatively rare lymph nodes remains low. Expanding the sample size for training and adjusting model parameters are essential steps to enhance recognition coverage, enabling the image diagnosis system to serve as a valuable aid in clinical practice.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e Yi han Feng: Conceptualization, Methodology, Software, Investigation, Formal Analysis, Writing - Original Draft, Data Curation, Writing - Original Draft, Visualization, Investigation, Resources, Supervision, Software, Validation; Visualization, Writing -Review\u0026amp; Editing; Chun han Pan: Conceptualization, Funding Acquisition, Resources, Supervision, Writing -Review\u0026amp; Editing.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e This retrospective study was approved by the Ethics Committee of Jiangsu Cancer Hospital and the methods were carried out in accordance with the approved guidelines. All the patients have been informed and signed informed consent before the experiments.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e The authors declare that there is no competing interests regarding the publication of this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOpen access\u003c/strong\u003e This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article\u0026rsquo;s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article\u0026rsquo;s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/ licenses/ by/4. 0/.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eMARQUINA G, MANZANO A, CASADO A. 2018. Targeted Agents in Cervical Cancer: Beyond Bevacizumab. \u003cem\u003eCurr Oncol Rep\u003c/em\u003e, 20, pp.40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKUMAR L, HARISH P, MALIK PS, KHURANA S. Chemotherapy and targeted therapy in the management of cervical cancer. Curr Probl Cancer. 2018;42:120\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLIU B, GAO S, LI S. A Comprehensive Comparison of CT, MRI, Positron Emission Tomography or Positron Emission Tomography/CT, and Diffusion Weighted Imaging-MRI for Detecting the Lymph Nodes Metastases in Patients with Cervical Cancer: A Meta-Analysis Based on 67 Studies. Gynecol Obstet Invest. 2017;82:209\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYAMASHITA R, NISHIO M, DO, R. K. G., TOGASHI K. Convolutional neural networks: an overview and application in radiology. Insights Imaging. 2018;9:611\u0026ndash;29.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKAWAGUCHI Y, MATSUURA Y, KONDO Y, ICHINOSE J, NAKAO M, OKUMURA S, MUN M. The predictive power of artificial intelligence on mediastinal lymphnode metastasis. Gen Thorac Cardiovasc Surg. 2021;69:1545\u0026ndash;52.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDAS N, TOPALOVIC, M., JANSSENS W. Artificial intelligence in diagnosis of obstructive lung disease: current status and future potential. Curr Opin Pulm Med. 2018;24:117\u0026ndash;23.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWENG S, XU X, LI J, WONG STC. Combining deep learning and coherent anti-Stokes Raman scattering imaging for automated differential diagnosis of lung cancer. J Biomed Opt. 2017;22:1\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLEE YH. Efficiency Improvement in a Busy Radiology Practice: Determination of Musculoskeletal Magnetic Resonance Imaging Protocol Using Deep-Learning Convolutional Neural Networks. J Digit Imaging. 2018;31:604\u0026ndash;10.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJUN Y, EO T, KIM T, SHIN H, HWANG D, PARK BAESH, LEE YW, CHOI HJ. B. W. \u0026amp; AHN, S. S. 2018. Deep-learned 3D black-blood imaging using automatic labelling technique and 3D convolutional neural networks for detecting metastatic brain tumors. \u003cem\u003eSci Rep\u003c/em\u003e, 8, pp.9450.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCHANG P, BARDIS GRINBANDJWEINBERGBD, KHY M, CADENA M, CHA GSUMY, BOTA SFILIPPICG, BALDI D, JAIN PPOISSONLM, R., CHOW D. Deep-Learning Convolutional Neural Networks Accurately Classify Genetic Mutations in Gliomas. AJNR Am J Neuroradiol. 2018;39:1201\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDU R, LI L, MA S, TAN X, ZHONG S, WU M. Lymph nodes metastasis in cervical cancer: Incidences, risk factors, consequences and imaging evaluations. Asia Pac J Clin Oncol. 2018;14:e380\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMINIG L, SCAMBIA FAGOTTIA, SALVO G, PATRONO G, HAIDOPOULOS MG, ZAPARDIEL D, DOMINGO I, SOTIROPOULOU S, CHISHOLM M, G., RAMIREZ PT. Incidence of Lymph Node Metastases in Women With Low-Risk Early Cervical Cancer (\u0026lt;\u0026thinsp;2 cm) Without Lymph-Vascular Invasion. Int J Gynecol Cancer. 2018;28:788\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJUNG W, LEE PARKKR, KIM KJ, LEE K, JEONG J, KIM S, KIM YJ, KANG JYOONHJ, KOO BC, CHO HSSUNGSH, M. S., PARK S. Value of imaging study in predicting pelvic lymph node metastases of uterine cervical cancer. Radiat Oncol J. 2017;35:340\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDHAMIJA E, BABY A, BHATLA N, KUMAR PULAPPADIVP, KUMAR M, KUMAR S, L., SHARMA D. Radiological evaluation of metastatic lymph nodes in carcinoma cervix with emphasis on their infiltrative pattern. Indian J Med Res. 2021;154:383\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCHEN J, DONG HEB, LIU D, DUAN P, LI H, LI W, WANG P, ZHANG LFANHWANGS, HUANG LTIANJ, Z., CHEN C. 2020. Noninvasive CT radiomic model for preoperative prediction of lymph node metastasis in early cervical carcinoma. \u003cem\u003eBr J Radiol\u003c/em\u003e, 93, pp.20190558.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBECKMANN MW, VORDERMARK STUEBSFA, KOCH D, HORN MC, L. C., FEHM T. The Diagnosis, Treatment, and Aftercare of Cervical Carcinoma. Dtsch Arztebl Int. 2021;118:806\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Uterine cervical carcinoma (UCC), Convolutional neural networks (CNN), Image diagnosis system, Computed tomography","lastPublishedDoi":"10.21203/rs.3.rs-7292070/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7292070/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo analyze the factors influencing the accuracy and errors in the delineation of draining lymph nodes for uterine cervical carcinoma using a CT imaging diagnostic system based on two convolutional neural networks—GoogLeNet and Faster R-CNN.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMaterial and methods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 679 lymph nodes in 56 patients with Uterine Cervical Carcinoma (UCC) below the renal hilar level, around the main lumen vessels, and in the pelvic cavity were delineated by the image diagnostic system. Then, two associate chief physicians in the imaging department evaluated the outlined lymph nodes to check for any missed lymph nodes. The lymph nodes were categorized into the following groups based on their basic characteristics: 1. Size. Based on previous research regarding the capability of convolutional neural networks (CNN) to recognize nodular nodes and the likelihood of lymph nodes appearing in images, they were classified into three groups: those with a length and diameter of less than 0.5 cm, between 0.5 cm and 2 cm, and greater than 2 cm; 2. Location. Considering the probability of metastatic lymph nodes in cervical cancer, they were divided into bilateral pelvic wall, common iliac vessels, retroperitoneum (common location group), and pelvic lymph nodes (uncommon location group); 3. Presence or absence of necrosis. We analyzed the coverage rate of lymph nodes identified by the imaging diagnostic system and its consistency with manual diagnoses. We analyzed the coverage rate of lymph nodes identified by the imaging diagnostic system and its consistency with manual diagnoses.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA total of 679 lymph nodes were identified in this study, of which 281 were smaller than 0.5 cm, and 265 were outlined by the imaging diagnostic system. There were 312 lymph nodes measuring between 0.5 cm and 2 cm, of which 298 were outlined by the imaging diagnostic system. A total of 86 lymph nodes were larger than 2 cm, with 80 outlined. The diagnostic coverage rates for the three groups were 94.31%, 95.51%, and 93.02%, respectively, showing no statistically significant difference. Among the lymph nodes, 587 were located in common areas, with 575 outlined, resulting in a coverage rate of 97.96%. There were 92 unusual lymph nodes, of which 68 were outlined, yielding a coverage rate of 73.91% (p ≤ 0.01), indicating a statistically significant difference. There were 32 necrotic lymph nodes, with 23 outlined, resulting in a coverage rate of 71.88%. Additionally, there were 647 lymph nodes without necrosis, and 620 were outlined, yielding a coverage rate of 95.83% (p ≤ 0.01), indicating a statistically significant difference.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe imaging diagnostic system for (UCC) based on convolutional neural networks demonstrates a high degree of consistency with manual diagnoses. The coverage rate of lymph node delineation is significantly influenced by the location of lymph nodes and the presence or absence of necrosis, whereas the size of lymph nodes does not have a significant effect on the diagnostic system’s coverage rate. Therefore, the recognition capability of the diagnostic system for lymph nodes with necrosis and those in unusual locations should be enhanced to improve the overall coverage rate of diagnosis.\u003c/p\u003e","manuscriptTitle":"Factors Influencing the Accuracy and Coverage of CT-Based Lymph Node Delineation in Uterine Cervical Carcinoma: A Deep Learning Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-22 18:54:25","doi":"10.21203/rs.3.rs-7292070/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-02-12T09:41:53+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-01T15:10:54+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"42342949456495469708805259484291935896","date":"2025-11-19T00:19:22+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-13T04:32:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"310895835786105054241142484161887452065","date":"2025-11-08T12:06:56+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-15T05:30:16+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-10T16:39:41+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-29T04:55:53+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-28T15:07:08+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Oncology","date":"2025-08-28T14:17:52+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c72cc7ec-8afd-4c60-abcf-5f1d386efdee","owner":[],"postedDate":"September 22nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-03-11T09:23:15+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-22 18:54:25","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7292070","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7292070","identity":"rs-7292070","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}