AI-Assisted Comparative Analysis of Pediatric Urology Guidelines (EAU-AUA-NICE): A Cross-Guideline NLP-Based Study

AI-Assisted Comparative Analysis of Pediatric Urology Guidelines (EAU-AUA-NICE): A Cross-Guideline NLP-Based Study | 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 AI-Assisted Comparative Analysis of Pediatric Urology Guidelines (EAU-AUA-NICE): A Cross-Guideline NLP-Based Study Çiğdem ARSLAN ALICI, Aykut AYKAÇ This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7068363/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: Pediatric urology guidelines by EAU, AUA, and NICE differ in terminology and recommendations, leading to inconsistencies in clinical practice. Harmonizing these standards can enhance decision-making in global pediatric care. Objective: This study aimed to compare major pediatric urology guidelines using artificial intelligence (AI) and natural language processing (NLP) to identify areas of concordance and divergence. Methods: Full-text guidelines from EAU, AUA, and NICE (2023–2025) were analyzed across four domains: vesicoureteral reflux (VUR), enuresis, hydronephrosis, and imaging. NLP tools (BERT, GPT, SciSpacy) were applied to extract and classify recommendations. Semantic similarity metrics (Jaccard and cosine similarity) and expert panel validation were used to assess alignment. Results: High concordance was observed between EAU and AUA guidelines (Jaccard 0.82; cosine 0.91), while NICE diverged moderately. VUR recommendations showed the highest agreement (9 of 12 statements), whereas hydronephrosis had greater variation. Enuresis terminology varied, especially between AUA and NICE. Expert validation confirmed AI-derived findings in 90% of sampled statements. Visualizations clearly illustrated alignment and conflict areas. Conclusion: This study demonstrates the utility of AI in systematically comparing pediatric urology guidelines. The approach may support future harmonization efforts and inform clinical decision support tools by highlighting both consensus and variation across international recommendations. Pediatric urology EAU AUA NICE Artificial intelligence Natural language processing Figures Figure 1 Figure 2 Figure 3 Introduction Pediatric urology spans a wide array of diagnoses with varied management strategies. Despite the availability of international guidelines such as those from the European Association of Urology (EAU) [ 1 ], the American Urological Association (AUA) [ 2 ], and the National Institute for Health and Care Excellence (NICE) [ 3 ] substantial heterogeneity exists. For clinicians working across systems, these inconsistencies complicate evidence-based care delivery and may lead to practice variability and suboptimal outcomes [ 4 ]. The rise of artificial intelligence (AI), particularly natural language processing (NLP), offers a systematic and scalable approach to evaluate large textual datasets, such as medical guidelines, by extracting, classifying, and comparing clinical recommendations [ 5 , 6 ]. These methods have shown promise in other areas of medicine, including oncology and cardiology guideline harmonization, but remain underexplored in pediatric urology [ 7 , 8 ]. This study presents a methodology to apply NLP in comparing pediatric urology guidelines, aiming to identify convergence, clarify differences, and assist in harmonized decision-making. Materials and Methods Pediatric urology guidelines from EAU, AUA, and NICE were selected based on their international recognition, broad clinical applicability, recent updates within the last 2 years (2023–2025), and explicit coverage of pediatric urology topics. Inclusion criteria required that guidelines must be publicly available, regularly updated, and cited frequently in clinical practice. Guidelines were specifically chosen for their comprehensive coverage of vesicoureteral reflux, enuresis, hydronephrosis, and diagnostic (Table 1 ). Table 1 Comparative Overview of Common Topics in Pediatric Urology Guidelines Topic EAU/ESPU AUA NICE Urinary Tract Infections (UTIs) + + + Vesicoureteral Reflux (VUR) + + +* Prenatal/Postnatal Hydronephrosis + + +* Urinary Incontinence / Nocturnal Enuresis + + + Diagnostic Imaging Approaches + + + Legend: ‘+’ indicates the topic is explicitly addressed in the guideline., ‘+*’ indicates the topic is covered indirectly or to a limited extent. The textual content of each guideline was processed using advanced NLP tools, including pretrained BERT and GPT-based models, and medical-specific tokenizers available in SciSpacy. Key sections were segmented and labeled using entity recognition to classify guideline content by disease, intervention, and recommendation strength. Manual review ensured the preservation of clinical context during automated processing. To quantify the similarity between guidelines, Jaccard similarity and cosine similarity metrics were calculated across disease-specific content. Sentence embeddings were used to vectorize the recommendations, and clustering algorithms (e.g., k-means) were applied to detect thematic proximity. These metrics facilitated the detection of semantic overlaps and divergences among the guidelines. An expert panel of three board-certified pediatric urologists independently reviewed a randomized subset of the extracted recommendations, particularly those flagged as conflicting or low in semantic alignment. Expert validation served to confirm clinical relevance and improve the specificity of the NLP-driven classification. To visually represent guideline similarities and differences, heatmaps and network graphs were generated using Python-based data visualization libraries. Additionally, decision trees summarizing harmonized recommendations were proposed for each clinical domain based on clustering results and expert interpretation. Results The comparative similarity analysis revealed a moderate-to-high degree of alignment between the pediatric urology guidelines from EAU, AUA, and NICE. Using Jaccard similarity as a semantic overlap metric, the highest alignment was observed between EAU and AUA guidelines (0.82), followed by AUA and NICE (0.70), and EAU and NICE (0.65). Cosine similarity values showed similar trends, with the highest similarity between EAU and AUA (0.91), and lower values for AUA–NICE (0.78) and EAU–NICE (0.75). These findings quantitatively demonstrate the degree of textual and thematic overlap across guideline pairs without accounting for contextual interpretation. Cosine similarity, reflecting vectorized content alignment, showed even higher values: 0.91 between EAU and AUA, 0.78 between AUA and NICE, and 0.75 between EAU and NICE. These values indicate that, despite terminological differences, the thematic content and overall treatment approach remain relatively consistent across guidelines. To further validate the NLP-derived outputs, precision, recall, and F1-score metrics were calculated based on expert-validated recommendation classifications. The models achieved a precision of 0.88, recall of 0.91, and an F1-score of 0.895. These metrics demonstrate the robustness of the AI approach in capturing semantically aligned and clinically relevant recommendations. A breakdown of guideline recommendations across four key domains—VUR, enuresis, hydronephrosis, and imaging strategies—revealed varied patterns of concordance. Of 12 VUR-related recommendations, 9 were concordant across all three guidelines, while 2 showed divergence, primarily regarding surgical thresholds and indications for continuous antibiotic prophylaxis. Notably, 1 recommendation concerning post-operative imaging protocols was missing from the NICE guideline, likely due to its preference for conservative follow-up and minimized imaging burden. For enuresis, 9 of 14 extracted recommendations were concordant. However, 3 statements exhibited terminological discrepancies—such as “enuresis nocturna” used by EAU versus “nighttime urinary incontinence” in NICE—without clearly aligned diagnostic criteria. Moreover, 2 recommendations concerning the use of combination therapy (alarm plus pharmacologic agent) and urodynamic assessment in refractory cases were absent from the AUA guideline, potentially reflecting a more stepwise treatment philosophy focused on first-line behavioral interventions. These discrepancies are clinically relevant as they may influence practitioner interpretation and treatment sequencing, especially in health systems with limited access to subspecialty care. Terminological heterogeneity also risks diagnostic ambiguity, highlighting the importance of adopting standardized pediatric continence terminology across international guidelines. Hydronephrosis yielded a lower concordance rate, with 8 of 15 recommendations aligned and 4 showing differences related to the timing of surgical referral and follow-up imaging protocols. Imaging strategies were generally more consistent, with 10 of 14 recommendations concordant and 2 instances of discordance involving preferences for initial imaging modality. The remaining recommendations were either unique to a single guideline or excluded from one or more documents. Expert panel validation supported the majority of NLP-derived similarity findings, with only 2 of 20 reviewed recommendation pairs flagged as “clinically discordant” despite semantic proximity. This highlights the importance of domain expertise in interpreting AI-derived outputs. Visualizations supported these quantitative results. Jaccard and cosine similarity heatmaps (Figs. 1 and 2) clearly depicted the proximity between guideline pairs. A stacked bar chart (Fig. 3) provided an overview of concordant, discordant, and missing recommendations by topic, revealing that VUR exhibited the highest agreement, while hydronephrosis showed the greatest variability. Discussion This study provides a novel, AI-assisted comparative analysis of major pediatric urology guidelines and highlights both areas of alignment and divergence across EAU, AUA, and NICE recommendations. Our findings demonstrate a generally high level of concordance between the EAU and AUA guidelines, particularly in domains such as vesicoureteral reflux (VUR), initial imaging strategies, and criteria for conservative management. This is consistent with the shared clinical priorities and overlapping evidence bases that inform North American and European guideline development [ 1 , 2 ]. These findings suggest that European and American guidelines share more terminological and structural similarities. However, NICE guidelines exhibited moderate divergence, particularly in the follow-up protocols for antenatal hydronephrosis, use of VCUG, and diagnostic imaging strategies. Notably, NICE tends to prioritize population-level cost-effectiveness and reduced radiation exposure, resulting in fewer imaging follow-ups and a more conservative management approach [ 3 ]. These differences reflect a distinct methodological framework compared to the more intervention-oriented perspectives of the EAU and AUA guidelines [ 9 ]. Future harmonization efforts should consider these underlying policy paradigms to better align international recommendations. Discrepancies in terminology were particularly evident in the management of functional disorders such as enuresis and dysfunctional voiding. Terms such as “enuresis nocturna” and “nighttime urinary incontinence” were used interchangeably across guidelines, yet lacked unified definitions or consistent diagnostic thresholds [ 10 ]. These terminological inconsistencies may contribute to clinical misinterpretation and highlight the necessity of a standardized lexicon, as previously advocated by the International Children’s Continence Society [ 11 ]. Through the integration of NLP tools, this study allowed for a systematic and scalable comparison across extensive guideline texts. High cosine and Jaccard similarity indices indicated that, despite structural and stylistic variation, core content alignment was substantial between EAU and AUA documents. These results were validated by a blinded expert panel, confirming > 90% concordance between AI-derived classifications and human assessment, thus supporting the clinical reliability of NLP applications in medical literature synthesis [ 5 , 12 ]. One of the major implications of this research lies in the potential for AI-based technologies to support international harmonization of pediatric urology guidelines. By identifying semantically concordant recommendation clusters and flagging divergent areas, these tools may assist multidisciplinary task forces in achieving cross-institutional alignment and minimizing practice variability [ 13 ]. This framework could also serve as the foundation for real-time, AI-powered clinical decision support systems tailored to pediatric urology, with the ability to integrate multiple guideline recommendations in a user-specific, contextualized format [ 14 ]. Nonetheless, several limitations merit acknowledgment. While NLP models demonstrated strong performance in textual clustering and semantic extraction, they may not fully capture nuanced clinical context, such as disease severity modifiers, expert consensus deviations, or rare comorbid scenarios without domain-specific training and manual curation. Additionally, the current analysis was limited to a select group of guideline topics—primarily VUR, enuresis, hydronephrosis, and initial imaging. Future extensions should include broader urologic conditions such as hypospadias, posterior urethral valves, neurogenic bladder, and pediatric stone disease, to ensure comprehensive harmonization efforts [ 15 ]. In conclusion, this AI-assisted guideline comparison illustrates how NLP tools can reliably detect semantic convergence and divergence among pediatric urology recommendations. We propose that future international panels incorporate AI-based cross-guideline alignment metrics to streamline consensus processes, particularly in areas where terminology and policy perspectives diverge, such as hydronephrosis imaging and functional urological disorders. Expansion of this framework to cover additional domains (e.g., hypospadias, neurogenic bladder) will enable a more robust and generalizable foundation for global pediatric urology practice harmonization. The authors did not use productive AI or AI-assisted technologies in the writing and preparation of this article. Declarations Funding There is no funding in this study. No ethical approval is required for this study. Author Contribution 1. Study Design: Çiğdem Arslan Alıcı, Aykut AYKAÇ2. Data Collection: Çiğdem Arslan Alıcı3. Data Analysis: Çiğdem Arslan Alıcı4. Data Interpretation: Çiğdem Arslan Alıcı, Aykut AYKAÇ5. Manuscript Preparation: Çiğdem Arslan Alıcı, Aykut AYKAÇ6. Literature Search: Çiğdem Arslan Alıcı, Aykut AYKAÇ7. Funds Collection: n/a Acknowledgement We would like to thank Assoc. Prof. Dr. Özer Baran, Prof. Dr. Ural Oğuz, and Prof. Dr. Baran Tokar for their support in the review process. References Tekgül S, Riedmiller H, Hoebeke P, Kočvara R, Nijman RJ, Radmayr C, Stein R, Dogan HS (2012) European Association of Urology. EAU guidelines on vesicoureteral reflux in children. Eur Urol. ;62(3):534 – 42. 10.1016/j.eururo.2012.05.059 . Epub 2012 Jun 5. PMID: 22698573 Peters CA, Skoog SJ, Arant BS Jr, Copp HL, Elder JS, Hudson RG, Khoury AE, Lorenzo AJ, Pohl HG, Shapiro E, Snodgrass WT, Diaz M (2010) Summary of the AUA Guideline on Management of Primary Vesicoureteral Reflux in Children. J Urol 184(3):1134–1144 Epub 2010 Jul 21. PMID: 20650499 Urinary tract infection in under 16s: diagnosis and management. London, National Institute for Health and Care Excellence (NICE) (2018) Oct : (NICE Clinical Guidelines, No. 54.) Available from: https://www.ncbi.nlm.nih.gov/books/NBK553083/ Meier KM, Mata C, Kaar JL, Rensing AJ, Dudley AG, Carrasco A Jr, Drzewiecki BA, VanderBrink BA, Streur CS, Bagli DJ, Chalmers DJ, Wilcox DT, Yerkes EB, Lau GA, Vricella GJ, Hecht SL, Copp HL, Pohl HG, Franco I, Ahn J, Wiener JS, Singer JS, Long CJ, Keays MA, Daugherty MR, Fuchs ME, Austin PF, Wu CQ, Zee RS, Misseri R, Tanaka ST, Bauer SB, Rove KO (2024) Expert Consensus on Pediatric Urodynamics Reporting Using Modified Delphi Technique. J Urol 212(1):165–174. 10.1097/JU.0000000000004000 Epub 2024 May 3. PMID: 38700226 Chan L, Beers K, Yau AA, Chauhan K, Duffy Á, Chaudhary K, Debnath N, Saha A, Pattharanitima P, Cho J, Kotanko P, Federman A, Coca SG, Van Vleck T, Nadkarni GN (2020) Natural language processing of electronic health records is superior to billing codes to identify symptom burden in hemodialysis patients. Kidney Int 97(2):383–392 Epub 2019 Nov 9. PMID: 31883805; PMCID: PMC7001114 Hossain E, Rana R, Higgins N, Soar J, Barua PD, Pisani AR, Turner K (2023) Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review. Comput Biol Med 155:106649. 10.1016/j.compbiomed.2023.106649 Epub 2023 Feb 10. PMID: 36805219 Khondker A, Kwong JC, Ahmad I, Rajesh Z, Dhalla R, MacNevin W, Rickard M, Erdman L, Gabrielson AT, Nguyen DD, Kim JK, Abbas T, Fernandez N, Fischer K, Hoen T, Keefe LA, Nelson DT, Viteri CP, Wang B, Weaver HS, Yadav J, Lorenzo P AJ. A living scoping review and online repository of artificial intelligence models in pediatric urology: Results from the AI-PEDURO collaborative. J Pediatr Urol. 2025 Feb 5:S1477-5131(25)00038 – 5. 10.1016/j.jpurol.2025.01.035 . Epub ahead of print. PMID: 39956703 Khondker A, Kwong JCC, Malik S, Erdman L, Keefe DT, Fernandez N, Tasian GE, Wang H-HS, Estrada CR Jr, Nelson CP, Lorenzo AJ, Rickard M (2022) The state of artificial intelligence in pediatric urology. Front Urol 2:1024662. 10.3389/fruro.2022.1024662 Silay MS, Undre S, Nambiar AK, Dogan HS, Kocvara R, Nijman RJM, Stein R, Tekgul S, Radmayr C (2017) Role of antibiotic prophylaxis in antenatal hydronephrosis: A systematic review from the European Association of Urology/European Society for Paediatric Urology Guidelines Panel. J Pediatr Urol 13(3):306–315 Epub 2017 Mar 19. PMID: 28462806 Nevéus T, von Gontard A, Hoebeke P, Hjälmås K, Bauer S, Bower W, Jørgensen TM, Rittig S, Walle JV, Yeung CK, Djurhuus JC (2006) The standardization of terminology of lower urinary tract function in children and adolescents: report from the Standardisation Committee of the International Children's Continence Society. J Urol. ;176(1):314 – 24. 10.1016/S0022-5347(06)00305-3 . PMID: 16753432 Austin PF, Bauer SB, Bower W, Chase J, Franco I, Hoebeke P, Rittig S, Walle JV, von Gontard A, Wright A, Yang SS, Nevéus T (2016) The standardization of terminology of lower urinary tract function in children and adolescents: Update report from the standardization committee of the International Children's Continence Society. Neurourol Urodyn 35(4):471–481. 10.1002/nau.22751 Epub 2015 Mar 14. PMID: 25772695 Winkler-Schwartz A, Bissonnette V, Mirchi N, Ponnudurai N, Yilmaz R, Ledwos N, Siyar S, Azarnoush H, Karlik B, Del Maestro RF (2019 Nov-Dec) Artificial Intelligence in Medical Education: Best Practices Using Machine Learning to Assess Surgical Expertise in Virtual Reality Simulation. J Surg Educ 76(6):1681–1690 Epub 2019 Jun 13. PMID: 31202633 Lekadir K, Frangi AF, Porras AR, Glocker B, Cintas C, Langlotz CP, Weicken E, Asselbergs FW, Prior F, Collins GS, Kaissis G, Tsakou G, Buvat I, Kalpathy-Cramer J, Mongan J, Schnabel JA, Kushibar K, Riklund K, Marias K, Amugongo LM, Fromont LA, Maier-Hein L, Cerdá-Alberich L, Martí-Bonmatí L, Cardoso MJ, Bobowicz M, Shabani M, Tsiknakis M, Zuluaga MA, Fritzsche MC, Camacho M, Linguraru MG, Wenzel M, De Bruijne M, Tolsgaard MG, Goisauf M, Cano Abadía M, Papanikolaou N, Lazrak N, Pujol O, Osuala R, Napel S, Colantonio S, Joshi S, Klein S, Aussó S, Rogers WA, Salahuddin Z, Starmans MPA, FUTURE-AI Consortium (2025). FUTURE-AI: international consensus guideline for trustworthy and deployable artificial intelligence in healthcare. BMJ. ;388:e081554. 10.1136/bmj-2024-081554 . Erratum in: BMJ. 2025;388:r340. doi: 10.1136/bmj.r340. PMID: 39909534; PMCID: PMC11795397 Chowdhury AT, Salam A, Naznine M, Abdalla D, Erdman L, Chowdhury MEH, Abbas TO (2024) Artificial Intelligence Tools in Pediatric Urology: A Comprehensive Review of Recent Advances. Diagnostics (Basel) 14(18):2059. 10.3390/diagnostics14182059 PMID: 39335738; PMCID: PMC11431426 Kwong JC, Khondker A, Kim JK, Chua M, Keefe DT, Dos Santos J, Skreta M, Erdman L, D'Souza N, Selman AF, Weaver J, Weiss DA, Long C, Tasian G, Teoh CW, Rickard M, Lorenzo AJ (2022) Posterior Urethral Valves Outcomes Prediction (PUVOP): a machine learning tool to predict clinically relevant outcomes in boys with posterior urethral valves. Pediatr Nephrol 37(5):1067–1074. 10.1007/s00467-021-05321-3 Epub 2021 Oct 22. PMID: 34686914 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-7068363","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":501870584,"identity":"75c49b45-5c13-403d-b910-7973fd4e231f","order_by":0,"name":"Çiğdem ARSLAN ALICI","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIiWNgGAWjYBACPgY2IGnAwMDP3nwAyJKQIaiFDaZFsudYAkgLD5FaQLpu+BiAaCK0sB9L/FxQYGPPcIPn86sbNRY8DOyHj27Aq4Un7bD0DIO0xMbZvdusc44BHcaTlnYDv8PSG6R5DA4nMMuc3WacwwbUIsFjhl8L//Pm3zwG/+3ZJHKeGef8I0aLRNoxoC0HGHskcpgf57YRpeVZmvUMg+TEGTzHzJhz+yR42Aj5hZ8/zfh2wR87e/vjzY8/53yrk+NnP3wMrxYQYIbbCCYJKUfWwvyBGNWjYBSMglEw8gAA8HxBSSqmNNgAAAAASUVORK5CYII=","orcid":"","institution":"Health Sciences University Eskisehir City Hospital, Pediatric Urology Clinic","correspondingAuthor":true,"prefix":"","firstName":"Çiğdem","middleName":"ARSLAN","lastName":"ALICI","suffix":""},{"id":501870590,"identity":"2743a268-2bc1-4fd6-9529-9ae6edffd6e0","order_by":1,"name":"Aykut AYKAÇ","email":"","orcid":"","institution":"Health Sciences University Eskisehir City Hospital","correspondingAuthor":false,"prefix":"","firstName":"Aykut","middleName":"","lastName":"AYKAÇ","suffix":""}],"badges":[],"createdAt":"2025-07-07 19:23:26","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7068363/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7068363/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89983572,"identity":"6ebf4faf-4b8a-4555-bb51-6d89eb16e41d","added_by":"auto","created_at":"2025-08-27 06:34:03","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":44708,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"Figure1JaccardHeatmap.png","url":"https://assets-eu.researchsquare.com/files/rs-7068363/v1/fe980606394f660559b9198e.png"},{"id":89985272,"identity":"8f077cf6-9e6e-4414-83de-6d12e0947bbe","added_by":"auto","created_at":"2025-08-27 06:42:03","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":42833,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"Figure2CosineHeatmap.png","url":"https://assets-eu.researchsquare.com/files/rs-7068363/v1/9a50519dd918488d3e959bc1.png"},{"id":89983630,"identity":"81711919-da45-49d5-b8bf-042cb090af40","added_by":"auto","created_at":"2025-08-27 06:34:04","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":87380,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"Figure3RecommendationConcordance.png","url":"https://assets-eu.researchsquare.com/files/rs-7068363/v1/ba0334169a2b74699934305f.png"},{"id":92186011,"identity":"2d9fa248-be76-4713-907e-223dd123d05f","added_by":"auto","created_at":"2025-09-25 14:18:07","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":510198,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7068363/v1/adb1da87-88ab-4ed5-afdb-36b2176edb39.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"AI-Assisted Comparative Analysis of Pediatric Urology Guidelines (EAU-AUA-NICE): A Cross-Guideline NLP-Based Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePediatric urology spans a wide array of diagnoses with varied management strategies. Despite the availability of international guidelines such as those from the European Association of Urology (EAU) [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e], the American Urological Association (AUA) [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e], and the National Institute for Health and Care Excellence (NICE) [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] substantial heterogeneity exists. For clinicians working across systems, these inconsistencies complicate evidence-based care delivery and may lead to practice variability and suboptimal outcomes [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe rise of artificial intelligence (AI), particularly natural language processing (NLP), offers a systematic and scalable approach to evaluate large textual datasets, such as medical guidelines, by extracting, classifying, and comparing clinical recommendations [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. These methods have shown promise in other areas of medicine, including oncology and cardiology guideline harmonization, but remain underexplored in pediatric urology [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. This study presents a methodology to apply NLP in comparing pediatric urology guidelines, aiming to identify convergence, clarify differences, and assist in harmonized decision-making.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003ePediatric urology guidelines from EAU, AUA, and NICE were selected based on their international recognition, broad clinical applicability, recent updates within the last 2 years (2023\u0026ndash;2025), and explicit coverage of pediatric urology topics. Inclusion criteria required that guidelines must be publicly available, regularly updated, and cited frequently in clinical practice. Guidelines were specifically chosen for their comprehensive coverage of vesicoureteral reflux, enuresis, hydronephrosis, and diagnostic (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eComparative Overview of Common Topics in Pediatric Urology Guidelines\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTopic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEAU/ESPU\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAUA\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNICE\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUrinary Tract Infections (UTIs)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVesicoureteral Reflux (VUR)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e+*\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePrenatal/Postnatal Hydronephrosis\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e+*\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUrinary Incontinence / Nocturnal Enuresis\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDiagnostic Imaging Approaches\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e+\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"4\"\u003eLegend: \u0026lsquo;+\u0026rsquo; indicates the topic is explicitly addressed in the guideline., \u0026lsquo;+*\u0026rsquo; indicates the topic is covered indirectly or to a limited extent.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe textual content of each guideline was processed using advanced NLP tools, including pretrained BERT and GPT-based models, and medical-specific tokenizers available in SciSpacy. Key sections were segmented and labeled using entity recognition to classify guideline content by disease, intervention, and recommendation strength. Manual review ensured the preservation of clinical context during automated processing.\u003c/p\u003e\u003cp\u003eTo quantify the similarity between guidelines, Jaccard similarity and cosine similarity metrics were calculated across disease-specific content. Sentence embeddings were used to vectorize the recommendations, and clustering algorithms (e.g., k-means) were applied to detect thematic proximity. These metrics facilitated the detection of semantic overlaps and divergences among the guidelines.\u003c/p\u003e\u003cp\u003eAn expert panel of three board-certified pediatric urologists independently reviewed a randomized subset of the extracted recommendations, particularly those flagged as conflicting or low in semantic alignment. Expert validation served to confirm clinical relevance and improve the specificity of the NLP-driven classification.\u003c/p\u003e\u003cp\u003eTo visually represent guideline similarities and differences, heatmaps and network graphs were generated using Python-based data visualization libraries. Additionally, decision trees summarizing harmonized recommendations were proposed for each clinical domain based on clustering results and expert interpretation.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eThe comparative similarity analysis revealed a moderate-to-high degree of alignment between the pediatric urology guidelines from EAU, AUA, and NICE. Using Jaccard similarity as a semantic overlap metric, the highest alignment was observed between EAU and AUA guidelines (0.82), followed by AUA and NICE (0.70), and EAU and NICE (0.65). Cosine similarity values showed similar trends, with the highest similarity between EAU and AUA (0.91), and lower values for AUA\u0026ndash;NICE (0.78) and EAU\u0026ndash;NICE (0.75). These findings quantitatively demonstrate the degree of textual and thematic overlap across guideline pairs without accounting for contextual interpretation.\u003c/p\u003e\u003cp\u003eCosine similarity, reflecting vectorized content alignment, showed even higher values: 0.91 between EAU and AUA, 0.78 between AUA and NICE, and 0.75 between EAU and NICE. These values indicate that, despite terminological differences, the thematic content and overall treatment approach remain relatively consistent across guidelines. To further validate the NLP-derived outputs, precision, recall, and F1-score metrics were calculated based on expert-validated recommendation classifications. The models achieved a precision of 0.88, recall of 0.91, and an F1-score of 0.895. These metrics demonstrate the robustness of the AI approach in capturing semantically aligned and clinically relevant recommendations.\u003c/p\u003e\u003cp\u003eA breakdown of guideline recommendations across four key domains\u0026mdash;VUR, enuresis, hydronephrosis, and imaging strategies\u0026mdash;revealed varied patterns of concordance. Of 12 VUR-related recommendations, 9 were concordant across all three guidelines, while 2 showed divergence, primarily regarding surgical thresholds and indications for continuous antibiotic prophylaxis. Notably, 1 recommendation concerning post-operative imaging protocols was missing from the NICE guideline, likely due to its preference for conservative follow-up and minimized imaging burden.\u003c/p\u003e\u003cp\u003eFor enuresis, 9 of 14 extracted recommendations were concordant. However, 3 statements exhibited terminological discrepancies\u0026mdash;such as \u0026ldquo;enuresis nocturna\u0026rdquo; used by EAU versus \u0026ldquo;nighttime urinary incontinence\u0026rdquo; in NICE\u0026mdash;without clearly aligned diagnostic criteria. Moreover, 2 recommendations concerning the use of combination therapy (alarm plus pharmacologic agent) and urodynamic assessment in refractory cases were absent from the AUA guideline, potentially reflecting a more stepwise treatment philosophy focused on first-line behavioral interventions.\u003c/p\u003e\u003cp\u003eThese discrepancies are clinically relevant as they may influence practitioner interpretation and treatment sequencing, especially in health systems with limited access to subspecialty care. Terminological heterogeneity also risks diagnostic ambiguity, highlighting the importance of adopting standardized pediatric continence terminology across international guidelines.\u003c/p\u003e\u003cp\u003eHydronephrosis yielded a lower concordance rate, with 8 of 15 recommendations aligned and 4 showing differences related to the timing of surgical referral and follow-up imaging protocols. Imaging strategies were generally more consistent, with 10 of 14 recommendations concordant and 2 instances of discordance involving preferences for initial imaging modality. The remaining recommendations were either unique to a single guideline or excluded from one or more documents.\u003c/p\u003e\u003cp\u003eExpert panel validation supported the majority of NLP-derived similarity findings, with only 2 of 20 reviewed recommendation pairs flagged as \u0026ldquo;clinically discordant\u0026rdquo; despite semantic proximity. This highlights the importance of domain expertise in interpreting AI-derived outputs.\u003c/p\u003e\u003cp\u003eVisualizations supported these quantitative results. Jaccard and cosine similarity heatmaps (Figs.\u0026nbsp;1 and 2) clearly depicted the proximity between guideline pairs. A stacked bar chart (Fig.\u0026nbsp;3) provided an overview of concordant, discordant, and missing recommendations by topic, revealing that VUR exhibited the highest agreement, while hydronephrosis showed the greatest variability.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study provides a novel, AI-assisted comparative analysis of major pediatric urology guidelines and highlights both areas of alignment and divergence across EAU, AUA, and NICE recommendations. Our findings demonstrate a generally high level of concordance between the EAU and AUA guidelines, particularly in domains such as vesicoureteral reflux (VUR), initial imaging strategies, and criteria for conservative management. This is consistent with the shared clinical priorities and overlapping evidence bases that inform North American and European guideline development [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. These findings suggest that European and American guidelines share more terminological and structural similarities.\u003c/p\u003e\u003cp\u003eHowever, NICE guidelines exhibited moderate divergence, particularly in the follow-up protocols for antenatal hydronephrosis, use of VCUG, and diagnostic imaging strategies. Notably, NICE tends to prioritize population-level cost-effectiveness and reduced radiation exposure, resulting in fewer imaging follow-ups and a more conservative management approach [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. These differences reflect a distinct methodological framework compared to the more intervention-oriented perspectives of the EAU and AUA guidelines [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Future harmonization efforts should consider these underlying policy paradigms to better align international recommendations.\u003c/p\u003e\u003cp\u003eDiscrepancies in terminology were particularly evident in the management of functional disorders such as enuresis and dysfunctional voiding. Terms such as \u0026ldquo;enuresis nocturna\u0026rdquo; and \u0026ldquo;nighttime urinary incontinence\u0026rdquo; were used interchangeably across guidelines, yet lacked unified definitions or consistent diagnostic thresholds [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. These terminological inconsistencies may contribute to clinical misinterpretation and highlight the necessity of a standardized lexicon, as previously advocated by the International Children\u0026rsquo;s Continence Society [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThrough the integration of NLP tools, this study allowed for a systematic and scalable comparison across extensive guideline texts. High cosine and Jaccard similarity indices indicated that, despite structural and stylistic variation, core content alignment was substantial between EAU and AUA documents. These results were validated by a blinded expert panel, confirming\u0026thinsp;\u0026gt;\u0026thinsp;90% concordance between AI-derived classifications and human assessment, thus supporting the clinical reliability of NLP applications in medical literature synthesis [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eOne of the major implications of this research lies in the potential for AI-based technologies to support international harmonization of pediatric urology guidelines. By identifying semantically concordant recommendation clusters and flagging divergent areas, these tools may assist multidisciplinary task forces in achieving cross-institutional alignment and minimizing practice variability [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. This framework could also serve as the foundation for real-time, AI-powered clinical decision support systems tailored to pediatric urology, with the ability to integrate multiple guideline recommendations in a user-specific, contextualized format [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eNonetheless, several limitations merit acknowledgment. While NLP models demonstrated strong performance in textual clustering and semantic extraction, they may not fully capture nuanced clinical context, such as disease severity modifiers, expert consensus deviations, or rare comorbid scenarios without domain-specific training and manual curation. Additionally, the current analysis was limited to a select group of guideline topics\u0026mdash;primarily VUR, enuresis, hydronephrosis, and initial imaging. Future extensions should include broader urologic conditions such as hypospadias, posterior urethral valves, neurogenic bladder, and pediatric stone disease, to ensure comprehensive harmonization efforts [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn conclusion, this AI-assisted guideline comparison illustrates how NLP tools can reliably detect semantic convergence and divergence among pediatric urology recommendations. We propose that future international panels incorporate AI-based cross-guideline alignment metrics to streamline consensus processes, particularly in areas where terminology and policy perspectives diverge, such as hydronephrosis imaging and functional urological disorders. Expansion of this framework to cover additional domains (e.g., hypospadias, neurogenic bladder) will enable a more robust and generalizable foundation for global pediatric urology practice harmonization.\u003c/p\u003e\u003cp\u003eThe authors did not use productive AI or AI-assisted technologies in the writing and preparation of this article.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eThere is no funding in this study.\u003c/p\u003e\u003cp\u003eNo ethical approval is required for this study.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003e1. Study Design: \u0026Ccedil;iğdem Arslan Alıcı, Aykut AYKA\u0026Ccedil;2. Data Collection: \u0026Ccedil;iğdem Arslan Alıcı3. Data Analysis: \u0026Ccedil;iğdem Arslan Alıcı4. Data Interpretation: \u0026Ccedil;iğdem Arslan Alıcı, Aykut AYKA\u0026Ccedil;5. Manuscript Preparation: \u0026Ccedil;iğdem Arslan Alıcı, Aykut AYKA\u0026Ccedil;6. Literature Search: \u0026Ccedil;iğdem Arslan Alıcı, Aykut AYKA\u0026Ccedil;7. Funds Collection: n/a\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe would like to thank Assoc. Prof. Dr. \u0026Ouml;zer Baran, Prof. Dr. Ural Oğuz, and Prof. Dr. Baran Tokar for their support in the review process.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eTekg\u0026uuml;l S, Riedmiller H, Hoebeke P, Kočvara R, Nijman RJ, Radmayr C, Stein R, Dogan HS (2012) European Association of Urology. EAU guidelines on vesicoureteral reflux in children. Eur Urol. ;62(3):534\u0026thinsp;\u0026ndash;\u0026thinsp;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.eururo.2012.05.059\u003c/span\u003e\u003cspan address=\"10.1016/j.eururo.2012.05.059\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2012 Jun 5. PMID: 22698573\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePeters CA, Skoog SJ, Arant BS Jr, Copp HL, Elder JS, Hudson RG, Khoury AE, Lorenzo AJ, Pohl HG, Shapiro E, Snodgrass WT, Diaz M (2010) Summary of the AUA Guideline on Management of Primary Vesicoureteral Reflux in Children. J Urol 184(3):1134\u0026ndash;1144 Epub 2010 Jul 21. PMID: 20650499\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUrinary tract infection in under 16s: diagnosis and management. London, National Institute for Health and Care Excellence (NICE) (2018) Oct : (NICE Clinical Guidelines, No. 54.) Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/books/NBK553083/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/books/NBK553083/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeier KM, Mata C, Kaar JL, Rensing AJ, Dudley AG, Carrasco A Jr, Drzewiecki BA, VanderBrink BA, Streur CS, Bagli DJ, Chalmers DJ, Wilcox DT, Yerkes EB, Lau GA, Vricella GJ, Hecht SL, Copp HL, Pohl HG, Franco I, Ahn J, Wiener JS, Singer JS, Long CJ, Keays MA, Daugherty MR, Fuchs ME, Austin PF, Wu CQ, Zee RS, Misseri R, Tanaka ST, Bauer SB, Rove KO (2024) Expert Consensus on Pediatric Urodynamics Reporting Using Modified Delphi Technique. J Urol 212(1):165\u0026ndash;174. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/JU.0000000000004000\u003c/span\u003e\u003cspan address=\"10.1097/JU.0000000000004000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003eEpub 2024 May 3. PMID: 38700226\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChan L, Beers K, Yau AA, Chauhan K, Duffy \u0026Aacute;, Chaudhary K, Debnath N, Saha A, Pattharanitima P, Cho J, Kotanko P, Federman A, Coca SG, Van Vleck T, Nadkarni GN (2020) Natural language processing of electronic health records is superior to billing codes to identify symptom burden in hemodialysis patients. Kidney Int 97(2):383\u0026ndash;392 Epub 2019 Nov 9. PMID: 31883805; PMCID: PMC7001114\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHossain E, Rana R, Higgins N, Soar J, Barua PD, Pisani AR, Turner K (2023) Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review. Comput Biol Med 155:106649. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.compbiomed.2023.106649\u003c/span\u003e\u003cspan address=\"10.1016/j.compbiomed.2023.106649\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003eEpub 2023 Feb 10. PMID: 36805219\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhondker A, Kwong JC, Ahmad I, Rajesh Z, Dhalla R, MacNevin W, Rickard M, Erdman L, Gabrielson AT, Nguyen DD, Kim JK, Abbas T, Fernandez N, Fischer K, Hoen T, Keefe LA, Nelson DT, Viteri CP, Wang B, Weaver HS, Yadav J, Lorenzo P AJ. A living scoping review and online repository of artificial intelligence models in pediatric urology: Results from the AI-PEDURO collaborative. J Pediatr Urol. 2025 Feb 5:S1477-5131(25)00038\u0026thinsp;\u0026ndash;\u0026thinsp;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jpurol.2025.01.035\u003c/span\u003e\u003cspan address=\"10.1016/j.jpurol.2025.01.035\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub ahead of print. PMID: 39956703\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhondker A, Kwong JCC, Malik S, Erdman L, Keefe DT, Fernandez N, Tasian GE, Wang H-HS, Estrada CR Jr, Nelson CP, Lorenzo AJ, Rickard M (2022) The state of artificial intelligence in pediatric urology. Front Urol 2:1024662. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fruro.2022.1024662\u003c/span\u003e\u003cspan address=\"10.3389/fruro.2022.1024662\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSilay MS, Undre S, Nambiar AK, Dogan HS, Kocvara R, Nijman RJM, Stein R, Tekgul S, Radmayr C (2017) Role of antibiotic prophylaxis in antenatal hydronephrosis: A systematic review from the European Association of Urology/European Society for Paediatric Urology Guidelines Panel. J Pediatr Urol 13(3):306\u0026ndash;315 Epub 2017 Mar 19. PMID: 28462806\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eNev\u0026eacute;us T, von Gontard A, Hoebeke P, Hj\u0026auml;lm\u0026aring;s K, Bauer S, Bower W, J\u0026oslash;rgensen TM, Rittig S, Walle JV, Yeung CK, Djurhuus JC (2006) The standardization of terminology of lower urinary tract function in children and adolescents: report from the Standardisation Committee of the International Children's Continence Society. J Urol. ;176(1):314\u0026thinsp;\u0026ndash;\u0026thinsp;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S0022-5347(06)00305-3\u003c/span\u003e\u003cspan address=\"10.1016/S0022-5347(06)00305-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. PMID: 16753432\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAustin PF, Bauer SB, Bower W, Chase J, Franco I, Hoebeke P, Rittig S, Walle JV, von Gontard A, Wright A, Yang SS, Nev\u0026eacute;us T (2016) The standardization of terminology of lower urinary tract function in children and adolescents: Update report from the standardization committee of the International Children's Continence Society. Neurourol Urodyn 35(4):471\u0026ndash;481. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/nau.22751\u003c/span\u003e\u003cspan address=\"10.1002/nau.22751\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003eEpub 2015 Mar 14. PMID: 25772695\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWinkler-Schwartz A, Bissonnette V, Mirchi N, Ponnudurai N, Yilmaz R, Ledwos N, Siyar S, Azarnoush H, Karlik B, Del Maestro RF (2019 Nov-Dec) Artificial Intelligence in Medical Education: Best Practices Using Machine Learning to Assess Surgical Expertise in Virtual Reality Simulation. J Surg Educ 76(6):1681\u0026ndash;1690 Epub 2019 Jun 13. PMID: 31202633\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLekadir K, Frangi AF, Porras AR, Glocker B, Cintas C, Langlotz CP, Weicken E, Asselbergs FW, Prior F, Collins GS, Kaissis G, Tsakou G, Buvat I, Kalpathy-Cramer J, Mongan J, Schnabel JA, Kushibar K, Riklund K, Marias K, Amugongo LM, Fromont LA, Maier-Hein L, Cerd\u0026aacute;-Alberich L, Mart\u0026iacute;-Bonmat\u0026iacute; L, Cardoso MJ, Bobowicz M, Shabani M, Tsiknakis M, Zuluaga MA, Fritzsche MC, Camacho M, Linguraru MG, Wenzel M, De Bruijne M, Tolsgaard MG, Goisauf M, Cano Abad\u0026iacute;a M, Papanikolaou N, Lazrak N, Pujol O, Osuala R, Napel S, Colantonio S, Joshi S, Klein S, Auss\u0026oacute; S, Rogers WA, Salahuddin Z, Starmans MPA, FUTURE-AI Consortium (2025). FUTURE-AI: international consensus guideline for trustworthy and deployable artificial intelligence in healthcare. BMJ. ;388:e081554. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmj-2024-081554\u003c/span\u003e\u003cspan address=\"10.1136/bmj-2024-081554\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Erratum in: BMJ. 2025;388:r340. doi: 10.1136/bmj.r340. PMID: 39909534; PMCID: PMC11795397\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChowdhury AT, Salam A, Naznine M, Abdalla D, Erdman L, Chowdhury MEH, Abbas TO (2024) Artificial Intelligence Tools in Pediatric Urology: A Comprehensive Review of Recent Advances. Diagnostics (Basel) 14(18):2059. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/diagnostics14182059\u003c/span\u003e\u003cspan address=\"10.3390/diagnostics14182059\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003ePMID: 39335738; PMCID: PMC11431426\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKwong JC, Khondker A, Kim JK, Chua M, Keefe DT, Dos Santos J, Skreta M, Erdman L, D'Souza N, Selman AF, Weaver J, Weiss DA, Long C, Tasian G, Teoh CW, Rickard M, Lorenzo AJ (2022) Posterior Urethral Valves Outcomes Prediction (PUVOP): a machine learning tool to predict clinically relevant outcomes in boys with posterior urethral valves. Pediatr Nephrol 37(5):1067\u0026ndash;1074. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00467-021-05321-3\u003c/span\u003e\u003cspan address=\"10.1007/s00467-021-05321-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003eEpub 2021 Oct 22. PMID: 34686914\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":"Pediatric urology, EAU, AUA, NICE, Artificial intelligence, Natural language processing","lastPublishedDoi":"10.21203/rs.3.rs-7068363/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7068363/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Pediatric urology guidelines by EAU, AUA, and NICE differ in terminology and recommendations, leading to inconsistencies in clinical practice. Harmonizing these standards can enhance decision-making in global pediatric care.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObjective:\u003c/strong\u003e This study aimed to compare major pediatric urology guidelines using artificial intelligence (AI) and natural language processing (NLP) to identify areas of concordance and divergence.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e Full-text guidelines from EAU, AUA, and NICE (2023–2025) were analyzed across four domains: vesicoureteral reflux (VUR), enuresis, hydronephrosis, and imaging. NLP tools (BERT, GPT, SciSpacy) were applied to extract and classify recommendations. Semantic similarity metrics (Jaccard and cosine similarity) and expert panel validation were used to assess alignment.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e High concordance was observed between EAU and AUA guidelines (Jaccard 0.82; cosine 0.91), while NICE diverged moderately. VUR recommendations showed the highest agreement (9 of 12 statements), whereas hydronephrosis had greater variation. Enuresis terminology varied, especially between AUA and NICE. Expert validation confirmed AI-derived findings in 90% of sampled statements. Visualizations clearly illustrated alignment and conflict areas.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e This study demonstrates the utility of AI in systematically comparing pediatric urology guidelines. The approach may support future harmonization efforts and inform clinical decision support tools by highlighting both consensus and variation across international recommendations.\u003c/p\u003e","manuscriptTitle":"AI-Assisted Comparative Analysis of Pediatric Urology Guidelines (EAU-AUA-NICE): A Cross-Guideline NLP-Based Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-27 06:33:58","doi":"10.21203/rs.3.rs-7068363/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":"ee22f90e-6142-41aa-a4cf-152cdf80302b","owner":[],"postedDate":"August 27th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-09-25T14:09:59+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-27 06:33:58","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7068363","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7068363","identity":"rs-7068363","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}