Global research trends and Artificial Intelligence applications in Fibro-Osseous Lesions: A Scientometric Analysis

Global research trends and Artificial Intelligence applications in Fibro-Osseous Lesions: A Scientometric Analysis | 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 Global research trends and Artificial Intelligence applications in Fibro-Osseous Lesions: A Scientometric Analysis Bruno Špiljak, Monal Yuwanati, Gowri Sivaramkrishnan, Julien Issa, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9573539/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 Objectives To assess global research trends on fibro-osseous lesions (FOLs) involving artificial intelligence (AI) by analysing publication patterns, key contributors, collaboration networks, citation impact, and thematic development. Methods A comprehensive bibliometric and scientometric analysis was performed using Scopus data and evaluated through Bibliometrix (R) and VOSviewer. Indicators included annual publication output, productive authors, institutions, and countries, journal impact, citation metrics, keyword co-occurrence, thematic mapping, and bibliographic coupling. All articles and reviews related to FOLs and AI published between 1995 and January 2025 were included. Results A total of 194 documents met the inclusion criteria. Annual scientific production remained low until 2016, followed by a marked increase from 2017 onward, with a peak in 2024. China, the United States, and Germany were the most productive countries, while Switzerland and Belgium had the highest citation rates per document. Leading institutions included Shanghai Jiao Tong University and the Technical University of Munich. Keyword analyses revealed a shift from traditional diagnostic terms to AI-driven concepts, including “radiomics,” “deep learning,” and “bone tumors.” Thematic and bibliographic coupling analyses revealed increasing integration of FOL research with radiology, oncology, and computational imaging methodologies. Conclusion Research on FOLs increasingly incorporates AI, reflecting a broader shift toward quantitative, image-based diagnostic methods. Although global output continues to rise, collaboration networks remain fragmented, and standardized AI workflows for FOL assessment are limited. The findings of this study offer a framework for future innovation in skeletal pathology and computer-assisted diagnosis. Dentistry Fibro-osseous lesions bibliometric analysis artificial intelligence deep learning diagnostic imaging scientometric trends Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Introduction Fibro-osseous lesions (FOLs) are a heterogeneous group of conditions in which normal bone is replaced by fibrous tissue (bone, osteoid, cementum) and varying mineralized components. Entities such as fibrous dysplasia, cemento-osseous dysplasia, and ossifying fibroma predominantly affect the oral and maxillofacial areas and often present with overlapping radiographic features, making diagnosis challenging [ 1 – 4 ]. Although typically benign, their clinical behavior ranges from stable to locally aggressive, requiring careful radiologic assessment and individualized management [ 5 ]. Advanced imaging modalities, particularly cone-beam computed tomography (CBCT) and computed tomography (CT), have enhanced the characterization of these lesions. Magnetic resonance imaging (MRI) and positron emission tomography (PET) may be used selectively in specific clinical scenarios; however, interpretation across modalities remains subjective and dependent on specialist experience. In recent years, artificial intelligence (AI), particularly deep learning and radiomics, has been increasingly applied to oral and maxillofacial imaging to improve lesion detection, segmentation, and classification. AI-based approaches have shown promise in evaluating bone lesions, including metastases, sarcomas, and primary bone tumors, yet their application to FOLs remains limited and fragmented [ 6 ]. Given the rarity and diagnostic ambiguity of FOLs, research in this domain is often fragmented, with limited cross-institutional collaboration and diverse methodological approaches. Scientometric and bibliometric analyses provide a structured approach to examining such fields by quantifying research output, identifying influential contributors, and tracing the evolution of themes over time [ 7 ]. To date, no comprehensive scientometric evaluation has assessed global research activity on FOLs in the context of AI and radiologic diagnostics. This study addresses this gap by analyzing publication trends, collaboration patterns, citation impact, and thematic development over the past three decades. Using Bibliometrix and VOSviewer, we aim to characterize the intellectual structure of this field and identify emerging research directions at the intersection of craniofacial pathology, radiology, and artificial intelligence. Materials and Methods Eligibility Criteria Eligible publications included peer-reviewed articles and reviews indexed in Scopus that explicitly addressed fibro-osseous lesions in the craniofacial region or related oral and maxillofacial pathology, and that incorporated artificial intelligence, machine learning, deep learning, or radiomics in their methodology or diagnostic applications. No language or publication-year restrictions were applied. Exclusion criteria comprised publications not indexed in Scopus, studies unrelated to fibro-osseous lesions (such as general bone tumors or systemic skeletal disorders without craniofacial focus), AI studies in dentistry or medicine that did not involve fibro-osseous lesions, as well as editorials, letters, notes, and other non-peer-reviewed material. Data Source and Search Strategy A comprehensive search was conducted on 9 January 2025 in the Scopus database, selected for its broad coverage of peer-reviewed scientific literature and suitability for bibliometric analysis. The search strategy combined controlled vocabulary and free-text terms related to FOLs and artificial intelligence. Considering the variation in terminology across the literature, the search strategy was intentionally broad to maximize sensitivity. The final query included: (“fibro-osseous lesion” OR “fibrous dysplasia” OR “ossifying fibroma” OR “osseous dysplasia”) AND (“artificial intelligence” OR “deep learning” OR “machine learning” OR “radiomics” OR “computer-aided diagnosis”). Complete Scopus search syntax is provided in Appendix I (Table S1). The detailed screening and selection process is depicted in Fig. 1 . Data Collection and Screening All retrieved records were exported in BibTeX and CSV formats for compatibility with bibliometric software. Initial screening involved removing duplicates and assessing all retrieved records for relevance; only those consistent with fibro-osseous lesions and artificial intelligence applications were retained.Titles and abstracts were manually reviewed by two independent reviewers to confirm thematic relevance. Any disagreements that arose between the reviewers at any stage of the process were resolved through discussion or with the involvement of a third reviewer. Bibliometric Analysis Tools and Parameters Data were analyzed using the Bibliometrix R package (version 4.2.0) and VOSviewer (version 1.6.20). Descriptive bibliometric indicators included annual scientific production, most relevant authors, journals, affiliations, and countries, citation metrics (local and global citations), authors' h-index, Lotka’s law of productivity, and source impact based on Bradford’s law. Network analyses were performed to visualize co-authorship patterns at the author, institutional, and country levels, as well as to examine keyword co-occurrence, bibliographic coupling among authors, documents, and countries, and thematic evolution over time. Longitudinal analyses were conducted in 5-year intervals to identify temporal shifts in research themes and collaboration patterns. Data Visualization and Quality Control Visual outputs, including collaboration maps, Sankey diagrams, word clouds, thematic maps, and trend charts, were generated using ggplot2, VOSviewer, and Biblioshiny, the interactive web interface of Bibliometrix. To ensure quality and reproducibility, all data extraction and analysis steps were independently verified by two reviewers. Results General Publication Trends A total of 194 documents met the inclusion criteria, comprising 153 original research articles and 41 review papers published between 1995 and January 2025 (Fig. 1 ). As illustrated in Fig. 2 , annual scientific production remained relatively low until 2016, followed by a marked increase beginning in 2017. This upward trajectory peaked in 2023 (n = 28) and 2024 (n = 36), indicating a substantial rise in interdisciplinary interest in applying artificial intelligence to FOLs. The cumulative citation count for the included documents was 2,439 with an average of 12.57 citations per document. The mean document age was 6.2 years, indicating that most publications were produced within the past decade, consistent with the recent expansion of AI-based diagnostic approaches in oral and maxillofacial radiology. Lotka’s law revealed a high proportion of one-time authorship (78%), while a small core group of researchers contributed multiple publications, reflecting concentrated expertise in this niche field. An analysis of author productivity similarly identified a limited but active cluster of contributors with five or more publications. Most studies were published in Q1 and Q2 journals, with a growing preference for open-access outlets in recent years, reflecting broader trends toward transparent and accessible AI-related medical research. Most Relevant Authors, Institutions, and Countries A total of 1,140 unique authors contributed to the 194 documents analyzed. The most prolific authors were Sconfienza, Luca Maria (Scopus ID: 24448438200; 8 publications), Albano, Domenico (Scopus ID: 57212327232; 7 publications), Gitto, Salvatore (Scopus ID: 57188980176; 7 publications), and Messina, Carmelo (Scopus ID: 55557560000) (Fig. 3 ), predominantly publishing on deep learning and radiomics applications in bone tumor diagnostics, with affiliations concentrated in major Chinese institutions. The most active institutions were the Technical University of Munich, Germany (17 documents ), Università Degli Studi Di Milano, Italy (15 documents), and Central South University, China (14 documents) (Fig. 4 ). These centers also demonstrated high centrality in co-authorship networks, indicating their roles as key research hubs in AI-based diagnostic research on FOLs. To avoid inflation in country production caused by multi-country collaborations in the count, we applied the fractional counting analysis method, in which each publication’s credit was proportionally divided among all contributing countries. Fractional counting analysis revealed that China contributed the largest share of publications (63.52), followed by the United States (26.14), Japan (22.42), Germany (15.20), and Italy (8.45). Other notable contributors included Netherlands (5.60), Sweden (5.44), South Korea (5.13), France (4.14), and Saudi Arabia (1.38). (Fig. 5 and Appendix Table S2). A similar trend was observed for citations, where China tops with 641 citations, followed by the United States (297 citations), Japan (273 citations), and Italy (214 citations) (Fig. 6 ). When adjusted for citation impact, however, in terms of average citation, Belgium and Switzerland showed the highest average citations per document, at 44.70 (134 citations; 3 documents) and 41.20 (165 citations; 8 documents), respectively. This suggests that although their output is smaller, their contributions are highly influential. The country collaboration map demonstrated robust international cooperation, particularly the United States (n = 28), Germany (n = 10), and Italy (n = 10) (Table 1 ). Notably, most transcontinental collaborations arose from multidisciplinary projects involving radiologists, computer scientists, and oral and maxillofacial surgeons. However, there is limited collaboration between South Asia and Asia with the rest of the world. Table 1 Country Collaboration Map. Rank Country Total documents 1 USA 28 2 Germany 10 3 Italy 10 4 China 9 5 Canada 8 Most Relevant Journals and Source Impact The included studies were published across 128 journals, reflecting broad interdisciplinary interest in FOLs and their diagnostic evaluation using artificial intelligence. The most productive journals were Oral Surgery, Oral Medicine, Oral Pathology, and Oral Radiology (n = 14), Journal of Cranio-Maxillofacial Surgery (n = 11), European Journal of Radiology (n = 9), Clinical Oral Investigations (n = 9), and Frontiers in Oncology (n = 8). In terms of source impact, as evaluated by total citations and h-index, the European Journal of Radiology and Oral Radiology recorded the highest local citation scores, while Scientific Reports and Frontiers in Oncology demonstrated strong global visibility, with frequent citations per article. Bradford’s law indicated that a small core of approximately eight journals accounted for nearly one-third of all publications, suggesting a concentration of influential research within select radiology and oral pathology outlets. Peripheral journals contributed to thematic diversity, though with lower citation accumulation. Longitudinal analysis of source activity showed a notable shift from earlier publications concentrated in oral pathology journals to more recent studies appearing in multidisciplinary and radiology-focused platforms such as Frontiers in Artificial Intelligence, Diagnostics, and Cancers. This reflects the increasing integration of AI methodologies into diagnostic research on FOLs. Keyword and Thematic Analysis A total of 728 unique author keywords were identified. The most frequent terms were “fibrous dysplasia” (n = 42), “ossifying fibroma” (n = 28), “artificial intelligence” (n = 26), “radiomics” (n = 21), and “deep learning” (n = 19). The word cloud visualization highlighted the prominence of AI-related keywords such as “machine learning,” “convolutional neural network,” and “segmentation,” indicating growing emphasis on automated image analysis and data-driven diagnostics. Terms such as “bone tumors” and “CBCT” (cone-beam computed tomography) also appeared frequently, suggesting the integration of FOL-focused research into broader oncologic and dental imaging domains (Fig. 7 ). The temporal keyword analysis revealed a clear thematic transition. Earlier publications were dominated by foundational diagnostic terms such as “fibrous dysplasia,” “histopathology,” and “cemento-osseous dysplasia,” whereas from 2017 onward, keywords increasingly reflected technologically advanced concepts including “radiomics,” “3D reconstruction,” “machine learning,” and “quantitative imaging” (Fig. 8 ). Most Globally and Locally Cited Documents The most globally cited document was a 2010 study published in the Journal of Magnetic Resonance Imaging by Juntu et al. [ 8 ], which focused on deep learning–based classification of craniofacial bone lesions and accumulated 121 citations (Fig. 9 b). In contrast, the most locally cited document, referring to citations within the analyzed corpus, was a 2021 radiological review of fibro-osseous lesions by Von Schacky [ 9 ], with 17 local citations (Fig. 9 a). Collaboration and Bibliographic Coupling Analysis Author and Institutional Collaboration Networks The co-authorship analysis showed that most studies were conducted within relatively small, institution-based teams, with limited intercontinental collaboration beyond China, the United States, and select European countries. The author collaboration map identified a central cluster led by Zhang Z. [Scopus ID: 57473490600] and Li X. [Scopus ID: 57895381000], characterized by strong internal linkage but limited external connectivity, reflecting concentrated productivity within a few high-output research groups. At the institutional level, the strongest collaborative links were observed between Shanghai Jiao Tong University and Peking University, followed by the Technical University of Munich and its European partners. Nevertheless, many institutions appeared isolated or only weakly connected, suggesting fragmentation in global research efforts and highlighting the need for stronger cross-institutional collaboration. Bibliographic Coupling of Documents, Authors, and Countries Bibliographic coupling analysis revealed several distinct document clusters, each representing thematic areas such as radiomics, CBCT segmentation, and deep learning–based classification. Documents with higher citation impact formed denser clusters, reflecting shared methodological frameworks and overlapping reference bases (Fig. 10 ). Author-level bibliographic coupling revealed three major clusters: one focused on diagnostic AI pipelines, another on FOL taxonomy and pathology, and a third on radiographic pattern recognition. Zhang Z. [Scopus ID: 57473490600] appeared across multiple clusters, indicating broad thematic involvement and methodological influence. Country-level coupling confirmed the dominant roles of China and the United States, with additional but smaller contributions from research communities in Europe, the Middle East, and Asia. Citation-normalized analyses further showed that Switzerland, Canada, and Belgium produced high-impact outputs despite lower publication volumes. Discussion This scientometric and bibliographic analysis provides the first comprehensive mapping of global research trends on FOLs in the context of AI and diagnostic imaging. The findings reveal a field in transition from traditional histopathological approaches to data-driven methodologies, mirroring broader trends in precision medicine and computational radiology [ 5 , 1 ]. The sharp increase in publications since 2017 reflects the growing convergence between diagnostic imaging, computational sciences, and oral and maxillofacial pathology. This trend aligns with the broader literature, which documents a similar rise in AI applications across oncology, radiology, and dentistry over the past decade [ 6 , 10 ]. The peaks observed in 2023 and 2024 coincide with the rapid proliferation of deep learning methods, particularly convolutional neural networks, applied to CBCT and CT datasets for lesion classification and segmentation [ 11 ]. China and the United States emerged as global leaders in scientific production, institutional centrality, and author impact. This reflects their substantial investments in AI research, computational infrastructure, and access to large imaging datasets. European countries, although contributing fewer publications, demonstrated higher citation impact per paper, indicating a focus on producing high-quality work. This pattern is consistent with previous scientometric findings in AI-driven diagnostics across oncology and neurology [ 12 , 7 ]. However, limited intercontinental and cross-disciplinary collaboration was observed. Most co-authorship networks were institution-based or intra-national, and the lack of multinational research frameworks may restrict the development of broadly applicable diagnostic models, particularly given ethnic, anatomical, and imaging protocol variability across regions [ 3 ]. Keyword analysis revealed a thematic shift over time from descriptive pathology terms, such as “fibrous dysplasia” and “ossifying fibroma,” to methodological terms, including “radiomics,” “machine learning,” and “quantitative imaging.” This trend reflects the growing incorporation of AI and computational radiology into disease characterization pipelines [ 11 ]. However, several important areas remain underrepresented. Few studies address the clinical validation of AI models, the standardization of imaging datasets, or the integration of multimodal data such as histology and imaging, all of which are essential for clinical translation [ 4 ]. Additionally, although FOLs are included in the WHO odontogenic tumor classification, limited work has examined their overlap with neoplastic or reactive lesions, which contributes to ongoing diagnostic ambiguity [ 13 ]. Nevertheless, several critical gaps remain in the current literature. Although AI-based methods show promise in automating lesion detection and improving diagnostic consistency, relatively few studies have conducted rigorous validation against expert human diagnosis or implemented prospective clinical trials. The reproducibility of these models across imaging modalities, patient populations, and institutional settings also remains insufficiently explored. Moreover, the multimodal integration of imaging, histopathology, and clinical data is rarely addressed, despite its potential to enhance classification accuracy and support therapeutic decision-making. Patient-centered outcomes, such as diagnostic confidence, time to diagnosis, and quality-of-life measures, are largely absent from existing research [ 14 ]. Crucially, the clinical ramifications of AI-based classification of FOLs remain ambiguous. While numerous recent studies indicate high diagnostic accuracy for automated detection and radiological differentiation, relatively few examine whether these tools affect clinical decision-making. In everyday practice, the management of FOLs often depends on a combination of radiological interpretation, histopathological confirmation, and longitudinal follow-up rather than solely on imaging results. As a result, even the most accurate AI-driven classification systems may have a limited impact unless incorporated into more comprehensive diagnostic workflows that incorporate clinicopathological correlation. Therefore, the integration of AI-driven classification systems into standard clinical decision-making remains limited, and forthcoming studies should assess their impact on treatment planning, follow-up approaches, and patient outcomes. Another limitation is the underrepresentation of certain lesion types and clinical contexts. While fibrous dysplasia and ossifying fibroma are frequently investigated, other FOLs variants, such as focal cemento-osseous dysplasia and juvenile trabecular ossifying fibroma, receive comparatively little attention. These gaps limit the generalizability of proposed diagnostic algorithms and underscore the need for more granular taxonomical and phenotypic stratification within datasets used for AI development [ 1 , 5 ]. A further challenge is the lack of open-access, annotated datasets specific to FOLs, which restricts innovation and reproducibility. Establishing standardized repositories and imaging benchmarks, similar to those available in oncology and neurology, would help stimulate broader engagement from the AI community and promote methodological transparency. Such resources may also facilitate closer collaboration among radiologists, dental specialists, pathologists, and computational scientists, who must work collectively to advance both the scientific foundation and clinical implementation of AI in FOL diagnosis [ 6 , 7 ]. Another methodological consideration pertains to the selection of databases. This analysis was conducted solely using the Scopus database, which is widely recognized in bibliometric research for its extensive journal coverage and compatibility with bibliometric software. Nevertheless, dependence on a single database may affect the observed patterns of publication and citation. Research in AI is often shared through conference proceedings, engineering journals, or interdisciplinary platforms that may be indexed differently in databases such as Web of Science, PubMed, or IEEE Xplore. As a result, some pertinent publications may not have been included in the current dataset. Future scientometric studies could benefit from including multiple databases to provide a more comprehensive representation of global research on AI-driven diagnostics for fibro-osseous lesions. Taken together, this review suggests that although the field of AI-enhanced diagnostics for fibro-osseous lesions is expanding rapidly, it remains methodologically fragmented and lacks sufficient validation. There is an urgent need for more robust, collaborative, and clinically grounded research to translate these promising technologies from academic development into routine diagnostic practice. Conclusion This scientometric review highlights a rapidly expanding body of research on fibro-osseous lesions at the interface of diagnostic imaging and artificial intelligence. A clear thematic shift toward quantitative, image-based diagnostics is evident, driven by advances in machine learning and radiomics. China and the United States lead in research output, while several European countries contribute high-impact work, yet overall collaboration remains limited, and the field lacks standardized datasets and clinically validated AI models. Future work should prioritize external validation, multimodal data integration, and the establishment of open-access annotated datasets to improve reproducibility. Strengthening international and cross-disciplinary collaboration will be essential to accelerate the clinical translation of AI-based tools for the diagnosis and management of fibro-osseous lesions. Declarations Conflicts of Interest The authors declare no competing financial or non-financial interests related to the content of this manuscript. Data Availability Statement The data used in this study were obtained from Scopusdatabase. All metadata generated during the bibliometric analysis, visualizations, and analytical outputs created using Bibliometrix (R) and VOSviewer are available in the main text or as supplementary files. Author Contributions Bruno Špiljak, Monal Yuwanati, Gowri Sivaramkrishnan, Julien Issa Anand Ramanathan, Akhilanand Chaurasia: Conceptualization, Database search, Methodology design, Writing – original draft. Monal Yuwanati, Gowri Sivaramkrishnan: Data curation, Software analysis (Bibliometrix and VOSviewer), Visualization. Bruno Špiljak, Monal Yuwanati, Julien Issa, Akhilanand Chaurasia: Review and editing, Interpretation of results, Supervision. All authors have read and approved the final manuscript. Ethical Considerations This study did not involve human participants, animal subjects, or patient data and was therefore exempt from ethical review. Acknowledgments Funding sources: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. References MacDonald DS. Maxillofacial fibro-osseous lesions. Clin Radiol . 2015;70(1):25–36. https://doi.org/10.1016/j.crad.2014.06.022 McCarthy EF. Fibro-osseous lesions of the maxillofacial bones. Head Neck Pathol . 2013;7(1):5–10. https://doi.org/10.1007/s12105-013-0430-7 Speight PM, Carlos R. Maxillofacial fibro-osseous lesions. Curr Diagn Pathol . 2006;12(1):1–10. Waldron CA. Fibro-osseous lesions of the jaws. J Oral Maxillofac Surg . 1993;51(8):828–35. Brannon RB, Fowler CB. Benign fibro-osseous lesions: a review of current concepts. Adv Anat Pathol . 2001;8(3):126–43. Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. Nat Rev Cancer . 2018;18(8):500–10. https://doi.org/10.1038/s41568-018-0016-5 Donthu N, Kumar S, Mukherjee D, Pandey N, Lim WM. How to conduct a bibliometric analysis: an overview and guidelines. J Bus Res . 2021;133:285–96. Juntu J, Sijbers J, De Backer S, Rajan J, Van Dyck D. Machine learning study of several classifiers trained with texture analysis features to differentiate benign from malignant soft-tissue tumors in T1-MRI images. J Magn Reson Imaging . 2010;31(3):680–9. https://doi.org/10.1002/jmri.22095 von Schacky CE, Wilhelm NJ, Schäfer VS, Leonhardt Y, Gassert FG, Foreman SC, et al. Multitask deep learning for segmentation and classification of primary bone tumors on radiographs. Radiology . 2021;301(2):398–406. https://doi.org/10.1148/radiol.2021204531 Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med . 2019;25(1):44–56. https://doi.org/10.1038/s41591-018-0300-7 Zhou SK, Greenspan H, Davatzikos C, Duncan JS, van Ginneken B, Madabhushi A, et al. A review of deep learning in medical imaging: imaging traits, technology trends, case studies with progress highlights, and future promises. Proc IEEE . 2021;109(5):820–38. https://doi.org/10.1109/JPROC.2021.3054390 Davenport T, Kalakota R. The potential for artificial intelligence in healthcare. Future Healthc J . 2019;6(2):94–8. https://doi.org/10.7861/futurehosp.6-2-94 Slootweg PJ. Maxillofacial fibro-osseous lesions: classification and differential diagnosis. Semin Diagn Pathol . 1996;13(2):104–12. Aerts HJWL, Velazquez ER, Leijenaar RTH, Parmar C, Grossmann P, Carvalho S, et al. Corrigendum: Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun . 2014;5:4644. https://doi.org/10.1038/ncomms5644 Additional Declarations The authors declare no competing interests. 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-9573539","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":632277515,"identity":"22d9c8df-f20c-4842-859d-8978c599258b","order_by":0,"name":"Bruno Špiljak","email":"","orcid":"https://orcid.org/0000-0003-2545-4474","institution":"Department of Oral Medicine, University of Zagreb School of Dental Medicine, Zagreb, 10000, Croatia","correspondingAuthor":false,"prefix":"","firstName":"Bruno","middleName":"","lastName":"Špiljak","suffix":""},{"id":632277516,"identity":"912fffb0-deab-4827-b473-2d96376d5d6a","order_by":1,"name":"Monal Yuwanati","email":"","orcid":"https://orcid.org/0000-0001-9233-8292","institution":"Department of Oral and Maxillofacial Pathology, Saveetha Dental College and Hospitals, Saveetha Institute of Medical and Technical Sciences, Saveetha University, Chennai, India.","correspondingAuthor":false,"prefix":"","firstName":"Monal","middleName":"","lastName":"Yuwanati","suffix":""},{"id":632277517,"identity":"97ba349f-45a2-4985-8245-570ed500e5e9","order_by":2,"name":"Gowri Sivaramkrishnan","email":"","orcid":"https://orcid.org/0000-0002-5877-205X","institution":"Specialty Dental Residency Program, Primary Health Care Centers, Bahrain","correspondingAuthor":false,"prefix":"","firstName":"Gowri","middleName":"","lastName":"Sivaramkrishnan","suffix":""},{"id":632277518,"identity":"2ceda433-c55c-4bde-bcfa-f4a8f3496c9f","order_by":3,"name":"Julien Issa","email":"","orcid":"https://orcid.org/0000-0002-6498-7989","institution":"Department of Oral Radiology \u0026 Digital Dentistry, Academic Centre for Dentistry Amsterdam (ACTA), University of Amsterdam \u0026 Vrije Universiteit Amsterdam, Gustav Mahlerlaan, Amsterdam, The Netherlands","correspondingAuthor":false,"prefix":"","firstName":"Julien","middleName":"","lastName":"Issa","suffix":""},{"id":632277519,"identity":"a1baad30-9539-4bde-9b72-176181fc63a1","order_by":4,"name":"Akhilanand Chaurasia","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0002-8356-9512","institution":"Department of Oral Medicine and Radiology King George’s Medical University, India","correspondingAuthor":true,"prefix":"","firstName":"Akhilanand","middleName":"","lastName":"Chaurasia","suffix":""}],"badges":[],"createdAt":"2026-04-30 07:30:19","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9573539/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9573539/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108411290,"identity":"5e9071f0-e2b9-4648-a2e0-7f8b2c4918dd","added_by":"auto","created_at":"2026-05-04 10:21:08","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":131820,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePRISMA flow diagram.\u003c/strong\u003e Illustrating the identification, screening, and inclusion of studies for the bibliometric analysis of FOL and AI.\u003c/p\u003e","description":"","filename":"image1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/416e89b394f9546348cebde8.jpeg"},{"id":108492880,"identity":"603447d8-1e3a-4dcd-be25-21b452ebd73f","added_by":"auto","created_at":"2026-05-05 09:58:53","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":74375,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAnnual Scientific Production (January\u003c/strong\u003e \u003cstrong\u003e1995–2025).\u003c/strong\u003e Number of publications per year on fibro-osseous lesions in relation to artificial intelligence and diagnostic imaging.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/52d77abb6a9b1e361da7bef8.png"},{"id":108493415,"identity":"7b311b43-f0e1-47ec-aa75-1d8b150cf3d2","added_by":"auto","created_at":"2026-05-05 10:00:19","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":59381,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMost Relevant Authors.\u003c/strong\u003e Top contributing authors by number of publications.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/89eebc3417f559cd1106b1fa.png"},{"id":108411293,"identity":"c2247be4-786a-488f-a143-5e5b413443e7","added_by":"auto","created_at":"2026-05-04 10:21:08","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":82692,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMost Relevant Affiliations.\u003c/strong\u003e Top contributing affiliation by number of publications.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/e0d7f5e9a9d2b344b822c13d.png"},{"id":108493482,"identity":"bcafe54f-f331-4dfe-987b-b566a1851632","added_by":"auto","created_at":"2026-05-05 10:00:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":21514,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCountry production of publications based on author affiliations\u003c/strong\u003e. Results are shown using fractional counting, which proportionally distributes credit among contributing countries. Full counting results, which highlight collaboration intensity but inflate totals, are provided in the supplementary table.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/e4fbcc63a9647581fabe1eef.png"},{"id":108493350,"identity":"5f63e0a8-44c7-472d-a5b4-fd26e53a899e","added_by":"auto","created_at":"2026-05-05 10:00:00","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":62201,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTop 10 most cited countries.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/b941b700ed2efcd5f341fa58.png"},{"id":108493483,"identity":"225a24c1-3839-4af5-a5df-eae1602bafb6","added_by":"auto","created_at":"2026-05-05 10:00:40","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":237069,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAuthor Keywords Word Cloud\u003c/strong\u003e: Visualization of the most frequent author-provided keywords.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/05e6eb84b0d27e5f7c4729c2.png"},{"id":108492886,"identity":"f7bfe82d-f12d-436f-a544-4a9b851d67c8","added_by":"auto","created_at":"2026-05-05 09:58:53","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":52178,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTrending Keywords Over Time. \u003c/strong\u003eTemporal evolution of keyword usage, illustrating a shift from histopathology-related terms to AI-focused concepts beginning after 2018.\u003c/p\u003e","description":"","filename":"image7.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/c013689e69e280844f2f9047.png"},{"id":108411295,"identity":"ef0f4864-0dea-466f-81cc-cbe18bb3481f","added_by":"auto","created_at":"2026-05-04 10:21:08","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":181415,"visible":true,"origin":"","legend":"\u003cp\u003eMost Cited Documents (a. Local and b. Global). Top-cited documents in terms of both global citations and intra-corpus (local) citations.\u003c/p\u003e","description":"","filename":"image8.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/0637b961d57bb7d5aff5af62.png"},{"id":108492889,"identity":"886917d7-8ae4-4b38-9afc-a8e7c9f853e9","added_by":"auto","created_at":"2026-05-05 09:58:54","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":276307,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBibliographic Coupling of Documents\u003c/strong\u003e. \u003cem\u003eClusters of documents based on shared references. Node size reflects citation impact, and color indicates cluster affiliation.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image9.png","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/d62542d3c75a83306c6bce0e.png"},{"id":108495201,"identity":"eddbba72-f423-400d-984e-05baea95791f","added_by":"auto","created_at":"2026-05-05 10:09:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1245022,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9573539/v1/47f09219-1424-4b44-9f6a-c90b506f5afb.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eGlobal research trends and Artificial Intelligence applications in Fibro-Osseous Lesions: A Scientometric Analysis\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eFibro-osseous lesions (FOLs) are a heterogeneous group of conditions in which normal bone is replaced by fibrous tissue (bone, osteoid, cementum) and varying mineralized components. Entities such as fibrous dysplasia, cemento-osseous dysplasia, and ossifying fibroma predominantly affect the oral and maxillofacial areas and often present with overlapping radiographic features, making diagnosis challenging [\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Although typically benign, their clinical behavior ranges from stable to locally aggressive, requiring careful radiologic assessment and individualized management [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAdvanced imaging modalities, particularly cone-beam computed tomography (CBCT) and computed tomography (CT), have enhanced the characterization of these lesions. Magnetic resonance imaging (MRI) and positron emission tomography (PET) may be used selectively in specific clinical scenarios; however, interpretation across modalities remains subjective and dependent on specialist experience. In recent years, artificial intelligence (AI), particularly deep learning and radiomics, has been increasingly applied to oral and maxillofacial imaging to improve lesion detection, segmentation, and classification. AI-based approaches have shown promise in evaluating bone lesions, including metastases, sarcomas, and primary bone tumors, yet their application to FOLs remains limited and fragmented [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eGiven the rarity and diagnostic ambiguity of FOLs, research in this domain is often fragmented, with limited cross-institutional collaboration and diverse methodological approaches. Scientometric and bibliometric analyses provide a structured approach to examining such fields by quantifying research output, identifying influential contributors, and tracing the evolution of themes over time [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. To date, no comprehensive scientometric evaluation has assessed global research activity on FOLs in the context of AI and radiologic diagnostics. This study addresses this gap by analyzing publication trends, collaboration patterns, citation impact, and thematic development over the past three decades. Using Bibliometrix and VOSviewer, we aim to characterize the intellectual structure of this field and identify emerging research directions at the intersection of craniofacial pathology, radiology, and artificial intelligence.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eEligibility Criteria\u003c/h2\u003e \u003cp\u003eEligible publications included peer-reviewed articles and reviews indexed in Scopus that explicitly addressed fibro-osseous lesions in the craniofacial region or related oral and maxillofacial pathology, and that incorporated artificial intelligence, machine learning, deep learning, or radiomics in their methodology or diagnostic applications. No language or publication-year restrictions were applied. Exclusion criteria comprised publications not indexed in Scopus, studies unrelated to fibro-osseous lesions (such as general bone tumors or systemic skeletal disorders without craniofacial focus), AI studies in dentistry or medicine that did not involve fibro-osseous lesions, as well as editorials, letters, notes, and other non-peer-reviewed material.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eData Source and Search Strategy\u003c/h3\u003e\n\u003cp\u003eA comprehensive search was conducted on 9 January 2025 in the Scopus database, selected for its broad coverage of peer-reviewed scientific literature and suitability for bibliometric analysis. The search strategy combined controlled vocabulary and free-text terms related to FOLs and artificial intelligence. Considering the variation in terminology across the literature, the search strategy was intentionally broad to maximize sensitivity. The final query included: (\u0026ldquo;fibro-osseous lesion\u0026rdquo; OR \u0026ldquo;fibrous dysplasia\u0026rdquo; OR \u0026ldquo;ossifying fibroma\u0026rdquo; OR \u0026ldquo;osseous dysplasia\u0026rdquo;) AND (\u0026ldquo;artificial intelligence\u0026rdquo; OR \u0026ldquo;deep learning\u0026rdquo; OR \u0026ldquo;machine learning\u0026rdquo; OR \u0026ldquo;radiomics\u0026rdquo; OR \u0026ldquo;computer-aided diagnosis\u0026rdquo;). Complete Scopus search syntax is provided in Appendix I (Table S1). The detailed screening and selection process is depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eData Collection and Screening\u003c/h3\u003e\n\u003cp\u003eAll retrieved records were exported in BibTeX and CSV formats for compatibility with bibliometric software. Initial screening involved removing duplicates and assessing all retrieved records for relevance; only those consistent with fibro-osseous lesions and artificial intelligence applications were retained.Titles and abstracts were manually reviewed by two independent reviewers to confirm thematic relevance. Any disagreements that arose between the reviewers at any stage of the process were resolved through discussion or with the involvement of a third reviewer.\u003c/p\u003e\n\u003ch3\u003eBibliometric Analysis Tools and Parameters\u003c/h3\u003e\n\u003cp\u003eData were analyzed using the Bibliometrix R package (version 4.2.0) and VOSviewer (version 1.6.20). Descriptive bibliometric indicators included annual scientific production, most relevant authors, journals, affiliations, and countries, citation metrics (local and global citations), authors' h-index, Lotka\u0026rsquo;s law of productivity, and source impact based on Bradford\u0026rsquo;s law.\u003c/p\u003e \u003cp\u003eNetwork analyses were performed to visualize co-authorship patterns at the author, institutional, and country levels, as well as to examine keyword co-occurrence, bibliographic coupling among authors, documents, and countries, and thematic evolution over time. Longitudinal analyses were conducted in 5-year intervals to identify temporal shifts in research themes and collaboration patterns.\u003c/p\u003e\n\u003ch3\u003eData Visualization and Quality Control\u003c/h3\u003e\n\u003cp\u003eVisual outputs, including collaboration maps, Sankey diagrams, word clouds, thematic maps, and trend charts, were generated using ggplot2, VOSviewer, and Biblioshiny, the interactive web interface of Bibliometrix. To ensure quality and reproducibility, all data extraction and analysis steps were independently verified by two reviewers.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eGeneral Publication Trends\u003c/h2\u003e \u003cp\u003eA total of 194 documents met the inclusion criteria, comprising 153 original research articles and 41 review papers published between 1995 and January 2025 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). As illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, annual scientific production remained relatively low until 2016, followed by a marked increase beginning in 2017. This upward trajectory peaked in 2023 (n\u0026thinsp;=\u0026thinsp;28) and 2024 (n\u0026thinsp;=\u0026thinsp;36), indicating a substantial rise in interdisciplinary interest in applying artificial intelligence to FOLs.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe cumulative citation count for the included documents was 2,439 with an average of 12.57 citations per document. The mean document age was 6.2 years, indicating that most publications were produced within the past decade, consistent with the recent expansion of AI-based diagnostic approaches in oral and maxillofacial radiology. Lotka\u0026rsquo;s law revealed a high proportion of one-time authorship (78%), while a small core group of researchers contributed multiple publications, reflecting concentrated expertise in this niche field. An analysis of author productivity similarly identified a limited but active cluster of contributors with five or more publications. Most studies were published in Q1 and Q2 journals, with a growing preference for open-access outlets in recent years, reflecting broader trends toward transparent and accessible AI-related medical research.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMost Relevant Authors, Institutions, and Countries\u003c/h3\u003e\n\u003cp\u003eA total of 1,140 unique authors contributed to the 194 documents analyzed. The most prolific authors were Sconfienza, Luca Maria (Scopus ID: 24448438200; 8 publications), Albano, Domenico (Scopus ID: 57212327232; 7 publications), Gitto, Salvatore (Scopus ID: 57188980176; 7 publications), and Messina, Carmelo (Scopus ID: 55557560000) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), predominantly publishing on deep learning and radiomics applications in bone tumor diagnostics, with affiliations concentrated in major Chinese institutions. The most active institutions were the Technical University of Munich, Germany (17 documents ), Universit\u0026agrave; Degli Studi Di Milano, Italy (15 documents), and Central South University, China (14 documents) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). These centers also demonstrated high centrality in co-authorship networks, indicating their roles as key research hubs in AI-based diagnostic research on FOLs.\u003c/p\u003e \u003cp\u003eTo avoid inflation in country production caused by multi-country collaborations in the count, we applied the fractional counting analysis method, in which each publication\u0026rsquo;s credit was proportionally divided among all contributing countries. Fractional counting analysis revealed that \u003cem\u003eChina\u003c/em\u003e contributed the largest share of publications (63.52), followed by the \u003cem\u003eUnited States\u003c/em\u003e (26.14), \u003cem\u003eJapan\u003c/em\u003e (22.42), \u003cem\u003eGermany\u003c/em\u003e (15.20), and \u003cem\u003eItaly\u003c/em\u003e (8.45). Other notable contributors included \u003cem\u003eNetherlands\u003c/em\u003e (5.60), \u003cem\u003eSweden\u003c/em\u003e (5.44), \u003cem\u003eSouth Korea\u003c/em\u003e (5.13), \u003cem\u003eFrance\u003c/em\u003e (4.14), and \u003cem\u003eSaudi Arabia\u003c/em\u003e (1.38). (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e and Appendix Table S2). A similar trend was observed for citations, where China tops with 641 citations, followed by the United States (297 citations), Japan (273 citations), and Italy (214 citations) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). When adjusted for citation impact, however, in terms of average citation, Belgium and Switzerland showed the highest average citations per document, at 44.70 (134 citations; 3 documents) and 41.20 (165 citations; 8 documents), respectively. This suggests that although their output is smaller, their contributions are highly influential.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe country collaboration map demonstrated robust international cooperation, particularly the United States (n\u0026thinsp;=\u0026thinsp;28), Germany (n\u0026thinsp;=\u0026thinsp;10), and Italy (n\u0026thinsp;=\u0026thinsp;10) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Notably, most transcontinental collaborations arose from multidisciplinary projects involving radiologists, computer scientists, and oral and maxillofacial surgeons. However, there is limited collaboration between South Asia and Asia with the rest of the world.\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\u003eCountry Collaboration Map.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\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=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRank\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCountry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTotal documents\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUSA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e28\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGermany\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eItaly\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eChina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCanada\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMost Relevant Journals and Source Impact\u003c/h2\u003e \u003cp\u003eThe included studies were published across 128 journals, reflecting broad interdisciplinary interest in FOLs and their diagnostic evaluation using artificial intelligence. The most productive journals were Oral Surgery, Oral Medicine, Oral Pathology, and Oral Radiology (n\u0026thinsp;=\u0026thinsp;14), Journal of Cranio-Maxillofacial Surgery (n\u0026thinsp;=\u0026thinsp;11), European Journal of Radiology (n\u0026thinsp;=\u0026thinsp;9), Clinical Oral Investigations (n\u0026thinsp;=\u0026thinsp;9), and Frontiers in Oncology (n\u0026thinsp;=\u0026thinsp;8). In terms of source impact, as evaluated by total citations and h-index, the European Journal of Radiology and Oral Radiology recorded the highest local citation scores, while Scientific Reports and Frontiers in Oncology demonstrated strong global visibility, with frequent citations per article.\u003c/p\u003e \u003cp\u003eBradford\u0026rsquo;s law indicated that a small core of approximately eight journals accounted for nearly one-third of all publications, suggesting a concentration of influential research within select radiology and oral pathology outlets. Peripheral journals contributed to thematic diversity, though with lower citation accumulation. Longitudinal analysis of source activity showed a notable shift from earlier publications concentrated in oral pathology journals to more recent studies appearing in multidisciplinary and radiology-focused platforms such as Frontiers in Artificial Intelligence, Diagnostics, and Cancers. This reflects the increasing integration of AI methodologies into diagnostic research on FOLs.\u003c/p\u003e \u003cp\u003e \u003cem\u003eKeyword and Thematic Analysis\u003c/em\u003e \u003c/p\u003e \u003cp\u003eA total of 728 unique author keywords were identified. The most frequent terms were \u0026ldquo;fibrous dysplasia\u0026rdquo; (n\u0026thinsp;=\u0026thinsp;42), \u0026ldquo;ossifying fibroma\u0026rdquo; (n\u0026thinsp;=\u0026thinsp;28), \u0026ldquo;artificial intelligence\u0026rdquo; (n\u0026thinsp;=\u0026thinsp;26), \u0026ldquo;radiomics\u0026rdquo; (n\u0026thinsp;=\u0026thinsp;21), and \u0026ldquo;deep learning\u0026rdquo; (n\u0026thinsp;=\u0026thinsp;19). The word cloud visualization highlighted the prominence of AI-related keywords such as \u0026ldquo;machine learning,\u0026rdquo; \u0026ldquo;convolutional neural network,\u0026rdquo; and \u0026ldquo;segmentation,\u0026rdquo; indicating growing emphasis on automated image analysis and data-driven diagnostics. Terms such as \u0026ldquo;bone tumors\u0026rdquo; and \u0026ldquo;CBCT\u0026rdquo; (cone-beam computed tomography) also appeared frequently, suggesting the integration of FOL-focused research into broader oncologic and dental imaging domains (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe temporal keyword analysis revealed a clear thematic transition. Earlier publications were dominated by foundational diagnostic terms such as \u0026ldquo;fibrous dysplasia,\u0026rdquo; \u0026ldquo;histopathology,\u0026rdquo; and \u0026ldquo;cemento-osseous dysplasia,\u0026rdquo; whereas from 2017 onward, keywords increasingly reflected technologically advanced concepts including \u0026ldquo;radiomics,\u0026rdquo; \u0026ldquo;3D reconstruction,\u0026rdquo; \u0026ldquo;machine learning,\u0026rdquo; and \u0026ldquo;quantitative imaging\u0026rdquo; (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eMost Globally and Locally Cited Documents\u003c/h2\u003e \u003cp\u003eThe most globally cited document was a 2010 study published in the Journal of Magnetic Resonance Imaging by Juntu et al. [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], which focused on deep learning\u0026ndash;based classification of craniofacial bone lesions and accumulated 121 citations (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eb). In contrast, the most locally cited document, referring to citations within the analyzed corpus, was a 2021 radiological review of fibro-osseous lesions by Von Schacky [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e], with 17 local citations (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ea).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eCollaboration and Bibliographic Coupling Analysis\u003c/h2\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003eAuthor and Institutional Collaboration Networks\u003c/h2\u003e \u003cp\u003eThe co-authorship analysis showed that most studies were conducted within relatively small, institution-based teams, with limited intercontinental collaboration beyond China, the United States, and select European countries. The author collaboration map identified a central cluster led by Zhang Z. [Scopus ID: 57473490600] and Li X. [Scopus ID: 57895381000], characterized by strong internal linkage but limited external connectivity, reflecting concentrated productivity within a few high-output research groups.\u003c/p\u003e \u003cp\u003eAt the institutional level, the strongest collaborative links were observed between Shanghai Jiao Tong University and Peking University, followed by the Technical University of Munich and its European partners. Nevertheless, many institutions appeared isolated or only weakly connected, suggesting fragmentation in global research efforts and highlighting the need for stronger cross-institutional collaboration.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eBibliographic Coupling of Documents, Authors, and Countries\u003c/h2\u003e \u003cp\u003eBibliographic coupling analysis revealed several distinct document clusters, each representing thematic areas such as radiomics, CBCT segmentation, and deep learning\u0026ndash;based classification. Documents with higher citation impact formed denser clusters, reflecting shared methodological frameworks and overlapping reference bases (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAuthor-level bibliographic coupling revealed three major clusters: one focused on diagnostic AI pipelines, another on FOL taxonomy and pathology, and a third on radiographic pattern recognition. Zhang Z. [Scopus ID: 57473490600] appeared across multiple clusters, indicating broad thematic involvement and methodological influence. Country-level coupling confirmed the dominant roles of China and the United States, with additional but smaller contributions from research communities in Europe, the Middle East, and Asia. Citation-normalized analyses further showed that Switzerland, Canada, and Belgium produced high-impact outputs despite lower publication volumes.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis scientometric and bibliographic analysis provides the first comprehensive mapping of global research trends on FOLs in the context of AI and diagnostic imaging. The findings reveal a field in transition from traditional histopathological approaches to data-driven methodologies, mirroring broader trends in precision medicine and computational radiology [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe sharp increase in publications since 2017 reflects the growing convergence between diagnostic imaging, computational sciences, and oral and maxillofacial pathology. This trend aligns with the broader literature, which documents a similar rise in AI applications across oncology, radiology, and dentistry over the past decade [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. The peaks observed in 2023 and 2024 coincide with the rapid proliferation of deep learning methods, particularly convolutional neural networks, applied to CBCT and CT datasets for lesion classification and segmentation [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eChina and the United States emerged as global leaders in scientific production, institutional centrality, and author impact. This reflects their substantial investments in AI research, computational infrastructure, and access to large imaging datasets. European countries, although contributing fewer publications, demonstrated higher citation impact per paper, indicating a focus on producing high-quality work. This pattern is consistent with previous scientometric findings in AI-driven diagnostics across oncology and neurology [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eHowever, limited intercontinental and cross-disciplinary collaboration was observed. Most co-authorship networks were institution-based or intra-national, and the lack of multinational research frameworks may restrict the development of broadly applicable diagnostic models, particularly given ethnic, anatomical, and imaging protocol variability across regions [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eKeyword analysis revealed a thematic shift over time from descriptive pathology terms, such as \u0026ldquo;fibrous dysplasia\u0026rdquo; and \u0026ldquo;ossifying fibroma,\u0026rdquo; to methodological terms, including \u0026ldquo;radiomics,\u0026rdquo; \u0026ldquo;machine learning,\u0026rdquo; and \u0026ldquo;quantitative imaging.\u0026rdquo; This trend reflects the growing incorporation of AI and computational radiology into disease characterization pipelines [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. However, several important areas remain underrepresented. Few studies address the clinical validation of AI models, the standardization of imaging datasets, or the integration of multimodal data such as histology and imaging, all of which are essential for clinical translation [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Additionally, although FOLs are included in the WHO odontogenic tumor classification, limited work has examined their overlap with neoplastic or reactive lesions, which contributes to ongoing diagnostic ambiguity [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eNevertheless, several critical gaps remain in the current literature. Although AI-based methods show promise in automating lesion detection and improving diagnostic consistency, relatively few studies have conducted rigorous validation against expert human diagnosis or implemented prospective clinical trials. The reproducibility of these models across imaging modalities, patient populations, and institutional settings also remains insufficiently explored. Moreover, the multimodal integration of imaging, histopathology, and clinical data is rarely addressed, despite its potential to enhance classification accuracy and support therapeutic decision-making. Patient-centered outcomes, such as diagnostic confidence, time to diagnosis, and quality-of-life measures, are largely absent from existing research [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Crucially, the clinical ramifications of AI-based classification of FOLs remain ambiguous. While numerous recent studies indicate high diagnostic accuracy for automated detection and radiological differentiation, relatively few examine whether these tools affect clinical decision-making. In everyday practice, the management of FOLs often depends on a combination of radiological interpretation, histopathological confirmation, and longitudinal follow-up rather than solely on imaging results. As a result, even the most accurate AI-driven classification systems may have a limited impact unless incorporated into more comprehensive diagnostic workflows that incorporate clinicopathological correlation. Therefore, the integration of AI-driven classification systems into standard clinical decision-making remains limited, and forthcoming studies should assess their impact on treatment planning, follow-up approaches, and patient outcomes.\u003c/p\u003e \u003cp\u003eAnother limitation is the underrepresentation of certain lesion types and clinical contexts. While fibrous dysplasia and ossifying fibroma are frequently investigated, other FOLs variants, such as focal cemento-osseous dysplasia and juvenile trabecular ossifying fibroma, receive comparatively little attention. These gaps limit the generalizability of proposed diagnostic algorithms and underscore the need for more granular taxonomical and phenotypic stratification within datasets used for AI development [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. A further challenge is the lack of open-access, annotated datasets specific to FOLs, which restricts innovation and reproducibility. Establishing standardized repositories and imaging benchmarks, similar to those available in oncology and neurology, would help stimulate broader engagement from the AI community and promote methodological transparency. Such resources may also facilitate closer collaboration among radiologists, dental specialists, pathologists, and computational scientists, who must work collectively to advance both the scientific foundation and clinical implementation of AI in FOL diagnosis [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Another methodological consideration pertains to the selection of databases. This analysis was conducted solely using the Scopus database, which is widely recognized in bibliometric research for its extensive journal coverage and compatibility with bibliometric software. Nevertheless, dependence on a single database may affect the observed patterns of publication and citation. Research in AI is often shared through conference proceedings, engineering journals, or interdisciplinary platforms that may be indexed differently in databases such as Web of Science, PubMed, or IEEE Xplore. As a result, some pertinent publications may not have been included in the current dataset. Future scientometric studies could benefit from including multiple databases to provide a more comprehensive representation of global research on AI-driven diagnostics for fibro-osseous lesions.\u003c/p\u003e \u003cp\u003eTaken together, this review suggests that although the field of AI-enhanced diagnostics for fibro-osseous lesions is expanding rapidly, it remains methodologically fragmented and lacks sufficient validation. There is an urgent need for more robust, collaborative, and clinically grounded research to translate these promising technologies from academic development into routine diagnostic practice.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis scientometric review highlights a rapidly expanding body of research on fibro-osseous lesions at the interface of diagnostic imaging and artificial intelligence. A clear thematic shift toward quantitative, image-based diagnostics is evident, driven by advances in machine learning and radiomics. China and the United States lead in research output, while several European countries contribute high-impact work, yet overall collaboration remains limited, and the field lacks standardized datasets and clinically validated AI models.\u003c/p\u003e \u003cp\u003eFuture work should prioritize external validation, multimodal data integration, and the establishment of open-access annotated datasets to improve reproducibility. Strengthening international and cross-disciplinary collaboration will be essential to accelerate the clinical translation of AI-based tools for the diagnosis and management of fibro-osseous lesions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eConflicts of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing financial or non-financial interests related to the content of this manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used in this study were obtained from Scopusdatabase. All metadata generated during the bibliometric analysis, visualizations, and analytical outputs created using Bibliometrix (R) and VOSviewer are available in the main text or as supplementary files.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBruno \u0026Scaron;piljak, Monal Yuwanati,\u0026nbsp;Gowri Sivaramkrishnan, Julien Issa Anand Ramanathan, Akhilanand Chaurasia: Conceptualization, Database search, Methodology design, Writing \u0026ndash; original draft.\u003c/p\u003e\n\u003cp\u003eMonal Yuwanati,\u0026nbsp;Gowri Sivaramkrishnan: Data curation, Software analysis (Bibliometrix and VOSviewer), Visualization.\u003c/p\u003e\n\u003cp\u003eBruno \u0026Scaron;piljak,\u0026nbsp;Monal Yuwanati, Julien Issa, Akhilanand Chaurasia: Review and editing, Interpretation of results, Supervision.\u003c/p\u003e\n\u003cp\u003eAll authors have read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u003cstrong\u003eEthical Considerations\u003c/strong\u003e\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThis study did not involve human participants, animal subjects, or patient data and was therefore exempt from ethical review.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFunding sources: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMacDonald DS. Maxillofacial fibro-osseous lesions. \u003cem\u003eClin Radiol\u003c/em\u003e. 2015;70(1):25\u0026ndash;36. https://doi.org/10.1016/j.crad.2014.06.022 \u003c/li\u003e\n\u003cli\u003eMcCarthy EF. Fibro-osseous lesions of the maxillofacial bones. \u003cem\u003eHead Neck Pathol\u003c/em\u003e. 2013;7(1):5\u0026ndash;10. https://doi.org/10.1007/s12105-013-0430-7 \u003c/li\u003e\n\u003cli\u003eSpeight PM, Carlos R. Maxillofacial fibro-osseous lesions. \u003cem\u003eCurr Diagn Pathol\u003c/em\u003e. 2006;12(1):1\u0026ndash;10.\u003c/li\u003e\n\u003cli\u003eWaldron CA. Fibro-osseous lesions of the jaws. \u003cem\u003eJ Oral Maxillofac Surg\u003c/em\u003e. 1993;51(8):828\u0026ndash;35.\u003c/li\u003e\n\u003cli\u003eBrannon RB, Fowler CB. Benign fibro-osseous lesions: a review of current concepts. \u003cem\u003eAdv Anat Pathol\u003c/em\u003e. 2001;8(3):126\u0026ndash;43.\u003c/li\u003e\n\u003cli\u003eHosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. \u003cem\u003eNat Rev Cancer\u003c/em\u003e. 2018;18(8):500\u0026ndash;10. https://doi.org/10.1038/s41568-018-0016-5 \u003c/li\u003e\n\u003cli\u003eDonthu N, Kumar S, Mukherjee D, Pandey N, Lim WM. How to conduct a bibliometric analysis: an overview and guidelines. \u003cem\u003eJ Bus Res\u003c/em\u003e. 2021;133:285\u0026ndash;96.\u003c/li\u003e\n\u003cli\u003eJuntu J, Sijbers J, De Backer S, Rajan J, Van Dyck D. Machine learning study of several classifiers trained with texture analysis features to differentiate benign from malignant soft-tissue tumors in T1-MRI images. \u003cem\u003eJ Magn Reson Imaging\u003c/em\u003e. 2010;31(3):680\u0026ndash;9. https://doi.org/10.1002/jmri.22095 \u003c/li\u003e\n\u003cli\u003evon Schacky CE, Wilhelm NJ, Sch\u0026auml;fer VS, Leonhardt Y, Gassert FG, Foreman SC, et al. Multitask deep learning for segmentation and classification of primary bone tumors on radiographs. \u003cem\u003eRadiology\u003c/em\u003e. 2021;301(2):398\u0026ndash;406. https://doi.org/10.1148/radiol.2021204531 \u003c/li\u003e\n\u003cli\u003eTopol EJ. High-performance medicine: the convergence of human and artificial intelligence. \u003cem\u003eNat Med\u003c/em\u003e. 2019;25(1):44\u0026ndash;56. https://doi.org/10.1038/s41591-018-0300-7 \u003c/li\u003e\n\u003cli\u003eZhou SK, Greenspan H, Davatzikos C, Duncan JS, van Ginneken B, Madabhushi A, et al. A review of deep learning in medical imaging: imaging traits, technology trends, case studies with progress highlights, and future promises. \u003cem\u003eProc IEEE\u003c/em\u003e. 2021;109(5):820\u0026ndash;38. https://doi.org/10.1109/JPROC.2021.3054390\u003c/li\u003e\n\u003cli\u003eDavenport T, Kalakota R. The potential for artificial intelligence in healthcare. \u003cem\u003eFuture Healthc J\u003c/em\u003e. 2019;6(2):94\u0026ndash;8. https://doi.org/10.7861/futurehosp.6-2-94\u003c/li\u003e\n\u003cli\u003eSlootweg PJ. Maxillofacial fibro-osseous lesions: classification and differential diagnosis. \u003cem\u003eSemin Diagn Pathol\u003c/em\u003e. 1996;13(2):104\u0026ndash;12.\u003c/li\u003e\n\u003cli\u003eAerts HJWL, Velazquez ER, Leijenaar RTH, Parmar C, Grossmann P, Carvalho S, et al. Corrigendum: Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. \u003cem\u003eNat Commun\u003c/em\u003e. 2014;5:4644. https://doi.org/10.1038/ncomms5644\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"King George's Medical University","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":"Fibro-osseous lesions, bibliometric analysis, artificial intelligence, deep learning, diagnostic imaging, scientometric trends","lastPublishedDoi":"10.21203/rs.3.rs-9573539/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9573539/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjectives\u003c/h2\u003e \u003cp\u003eTo assess global research trends on fibro-osseous lesions (FOLs) involving artificial intelligence (AI) by analysing publication patterns, key contributors, collaboration networks, citation impact, and thematic development.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA comprehensive bibliometric and scientometric analysis was performed using Scopus data and evaluated through Bibliometrix (R) and VOSviewer. Indicators included annual publication output, productive authors, institutions, and countries, journal impact, citation metrics, keyword co-occurrence, thematic mapping, and bibliographic coupling. All articles and reviews related to FOLs and AI published between 1995 and January 2025 were included.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA total of 194 documents met the inclusion criteria. Annual scientific production remained low until 2016, followed by a marked increase from 2017 onward, with a peak in 2024. China, the United States, and Germany were the most productive countries, while Switzerland and Belgium had the highest citation rates per document. Leading institutions included Shanghai Jiao Tong University and the Technical University of Munich. Keyword analyses revealed a shift from traditional diagnostic terms to AI-driven concepts, including \u0026ldquo;radiomics,\u0026rdquo; \u0026ldquo;deep learning,\u0026rdquo; and \u0026ldquo;bone tumors.\u0026rdquo; Thematic and bibliographic coupling analyses revealed increasing integration of FOL research with radiology, oncology, and computational imaging methodologies.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eResearch on FOLs increasingly incorporates AI, reflecting a broader shift toward quantitative, image-based diagnostic methods. Although global output continues to rise, collaboration networks remain fragmented, and standardized AI workflows for FOL assessment are limited. The findings of this study offer a framework for future innovation in skeletal pathology and computer-assisted diagnosis.\u003c/p\u003e","manuscriptTitle":"Global research trends and Artificial Intelligence applications in Fibro-Osseous Lesions: A Scientometric Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-04 10:21:03","doi":"10.21203/rs.3.rs-9573539/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":"f6c4ffeb-3941-44cf-a231-0dd6fd9c7133","owner":[],"postedDate":"May 4th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":67300015,"name":"Dentistry"}],"tags":[],"updatedAt":"2026-05-04T10:21:03+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-04 10:21:03","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9573539","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9573539","identity":"rs-9573539","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}