Integration of artificial intelligence in orthodontic diagnosis and treatment planning a PRISMA ScR guided scoping review

Integration of artificial intelligence in orthodontic diagnosis and treatment planning a PRISMA ScR guided scoping review | 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 Systematic Review Integration of artificial intelligence in orthodontic diagnosis and treatment planning a PRISMA ScR guided scoping review Salah Bin Hafedh, Ahmad Hashridz Bin Ruslan, Ramy Ishaq, Rozita Hassan, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8641111/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 Artificial intelligence (AI) applications in orthodontics are rapidly expanding across diagnosis, image analysis, and treatment planning. Methods A PRISMA-ScR–guided scoping review was conducted. PubMed/MEDLINE, Scopus, Web of Science, and Google Scholar were searched from 2010 to 16 September 2025. Original studies in orthodontics that used AI or machine learning for diagnosis, prediction, image analysis, or treatment planning were eligible. Two reviewers independently screened records, extracted data, and assessed risk of bias using QUADAS-2 for diagnostic accuracy studies and PROBAST for prediction model studies. Owing to heterogeneity in study design, datasets, and outcome metrics, results were synthesized narratively. Results Of 1,162 records identified, 1,008 remained after duplicate removal and were screened by title and abstract. A total of 154 full-text articles were assessed for eligibility, and 45 met the inclusion criteria. Frequent AI tasks included cephalometric landmark detection, malocclusion classification, extraction-decision support, treatment duration prediction, and cone-beam computed tomography (CBCT)-based segmentation. Many studies reported high accuracies for cephalometric landmark detection (mean radial error 80%) and malocclusion classification (accuracies > 85%). However, risk-of-bias concerns, particularly in analysis and validation domains, were common, and external validation was infrequent. Conclusions AI models show promising performance for orthodontic diagnosis and treatment planning and may enhance efficiency and standardization of care. Nevertheless, non-standardized outcome measures, limited external validation, and insufficient reporting of model development and evaluation currently restrict clinical translation. Larger, multicenter datasets, standardized benchmarks, and robust validation—ideally following AI-specific reporting guidelines—are required before routine clinical adoption. artificial intelligence machine learning deep learning orthodontics diagnosis treatment planning scoping review Figures Figure 1 1. Introduction In recent years, artificial intelligence (AI) has moved from being a futuristic concept to an everyday reality in healthcare. Orthodontics, where precision in diagnosis and treatment planning is essential for long-term functional and aesthetic outcomes, has been particularly affected by this transformation [ 1 – 3 ]. AI tools can help reduce diagnostic errors, streamline clinical workflows, and allow orthodontists to dedicate more time to direct patient care [ 4 – 6 ]. Machine learning (ML) and deep learning (DL), as subfields of AI, are particularly powerful for analyzing diverse clinical datasets, including cephalometric radiographs, CBCT scans, panoramic images, intraoral photographs, and structured patient records [ 7 – 11 ]. Convolutional neural networks (CNNs), among the most widely applied DL architectures, have shown high accuracy in cephalometric landmark detection, malocclusion classification, and image segmentation, often matching or surpassing the performance of human examiners [ 12 – 16 ]. Landmark detection studies frequently report mean radial errors below 2 mm and successful detection rates above 80%, while classification accuracies for malocclusion often exceed 85% [ 17 – 20 ]. Beyond diagnosis, AI is increasingly applied to complex clinical decision-making, such as predicting treatment duration, identifying when extractions are necessary, and estimating the need for orthognathic surgery [ 21 – 27 ]. These applications indicate that AI is evolving from a purely experimental tool to a clinically relevant adjunct in routine orthodontic practice [ 28 – 30 ]. However, important challenges remain. Many models are trained on single-center or demographically narrow datasets, which may limit generalizability. Ethical and regulatory issues—including data privacy, algorithmic bias, and medico-legal responsibility—also require careful consideration. Moreover, the “black-box” nature of many DL models can impede clinician trust in AI-generated recommendations [ 31 – 35 ]. This PRISMA-ScR–guided scoping review synthesizes evidence from original studies published between 2010 and 2025 that evaluated AI for orthodontic diagnosis and treatment planning. The objectives are to (1) summarize the types of AI tasks and methodologies used in orthodontics, (2) evaluate the reported diagnostic and predictive performance of these systems, and (3) identify current limitations and future research priorities for safe and effective AI integration into orthodontic care [ 36 – 40 ]. In contrast to prior narrative and systematic syntheses in this area, this scoping review is designed as an evidence map that charts the primary-study landscape by clinical task, data source, algorithm family, reference standard, and validation approach (internal versus external validation), and highlights methodological and reporting gaps that currently limit clinical translation. 2. Materials and Methods 2.1 Protocol and registration This scoping review was conducted in accordance with the PRISMA-ScR statement and relevant AI-focused reporting extensions when applicable. The review protocol was prospectively registered with PROSPERO (registration number CRD420251134644). 2.2 Information sources Electronic searches were carried out in PubMed/MEDLINE, Scopus, Web of Science (Core Collection), and Google Scholar. Databases were searched for articles published between 1 January 2010 and 16 September 2025. Search strategies combined keywords and MeSH terms related to “artificial intelligence,” “machine learning,” “deep learning,” “orthodontics,” “diagnosis,” “treatment planning,” and “cephalometrics.” Detailed, database-specific search strings are provided in the Supplementary material (Search strategies document). 2.3 Eligibility criteria Inclusion criteria Peer-reviewed original research articles. Studies involving AI (including ML, DL, CNNs, or hybrid techniques) applied to orthodontic diagnosis, prediction, image analysis, or treatment planning. Human clinical data or imaging (e.g., 2D cephalometric radiographs, CBCT, panoramic radiographs, intraoral photographs, or structured clinical records). Reporting of at least one quantitative performance metric (e.g., accuracy, sensitivity/specificity, AUC, mean radial error, Dice similarity coefficient, mean absolute error). Articles published in English between 1 January 2010 and 16 September 2025. Exclusion criteria Narrative reviews, scoping reviews, editorials, commentaries, conference abstracts, and letters without original data. Studies not primarily focused on orthodontics (e.g., general dentistry, restorative dentistry, endodontics) or AI applications not related to diagnosis or treatment planning. Studies that did not clearly describe their AI methodology or lacked a human reference standard. Animal or in-vitro studies. 2.4 Data extraction and quality assessment Two reviewers independently screened titles and abstracts, followed by full-text assessment of potentially eligible studies. Disagreements were resolved through discussion or consultation with a third reviewer. Data were extracted using a standardized form and included: first author, publication year, country, imaging modality, primary AI task (e.g., landmark detection, classification, prediction, segmentation), dataset size, AI architecture or algorithm, reference standard (e.g., expert orthodontist annotations), primary performance metrics, and whether internal and/or external validation was performed. Risk of bias in diagnostic accuracy studies was assessed using the QUADAS-2 tool, which evaluates four domains: patient selection, index test, reference standard, and flow/timing [ 2 ]. Prediction model studies were assessed using the PROBAST tool, which covers participants, predictors, outcomes, and analysis domains [ 3 ]. Summary risk-of-bias assessments are presented in the Supplementary QUADAS-2 and PROBAST tables, with study-level details provided separately. 2.5 Synthesis methods The included studies displayed substantial heterogeneity in study design (diagnostic accuracy vs prediction models), imaging modalities (2D radiographs, CBCT, panoramic radiographs, intraoral photographs, clinical records), AI architectures, and outcome measures (e.g., landmark error, Dice similarity coefficient, AUC, accuracy, MAE). Because of this variability, quantitative meta-analysis was not feasible. Instead, a structured narrative synthesis was conducted in accordance with PRISMA-ScR guidance. Studies were grouped by primary AI task (e.g., cephalometric landmark detection, malocclusion classification, extraction decision support, treatment duration prediction, CBCT segmentation), and key performance indicators were summarized within each category. 3. Results The PRISMA-ScR flow diagram summarizing study selection is shown in Fig. 1 near here (PRISMA-ScR flow diagram of study identification, screening, eligibility, and inclusion). Figure 1. PRISMA-ScR flow diagram of study identification, screening, eligibility, and inclusion. 3.1 Study selection The database search identified 1,162 records. After removal of duplicates, 1,008 records remained for title and abstract screening. Of these, 854 records were excluded. A total of 154 full-text articles were assessed for eligibility. One hundred and nine full-text articles were excluded for the following main reasons: not orthodontics-related AI studies (n = 35), review/editorial/commentary articles (n = 28), inappropriate or irrelevant study design (n = 20), insufficient diagnostic or treatment-planning detail (n = 16), and duplicate publication (n = 10). Finally, 45 studies met all inclusion criteria and were included in the qualitative synthesis. 3.2 Study characteristics Table 1 summarizes the characteristics of the 45 included studies. The majority were retrospective observational studies conducted in academic or specialist orthodontic settings. Common imaging modalities included 2D cephalometric radiographs, CBCT, panoramic radiographs, and intraoral photographs. The main AI tasks and applications were: Cephalometric landmark detection and cephalometric analysis; Malocclusion and skeletal pattern classification; Extraction-decision support and prediction of orthognathic surgery necessity; Prediction of treatment duration and treatment outcomes; CBCT-based 3D segmentation of dental and skeletal structures. Sample sizes varied widely, ranging from fewer than 100 images or patients to more than 1,000. CNN-based DL approaches were predominant for image-based tasks, whereas traditional ML algorithms (e.g., random forests, support vector machines, gradient boosting) were more commonly used for structured clinical data and prediction tasks. Table 1. Study Characteristics of Included Studies (n = 45) ID First author (Year) Country/Region Journal Imaging/Modality Primary Task Dataset size (N) AI Method Reference Standard Key Performance (primary metric) External Validation Citation (short) 1 Cui Z (2022) Multi-center (China, 15 sites) Nat Commun CBCT (3D) Tooth & alveolar bone segmentation 4,938 scans Multi-stage 3D CNN (nnU-Net style pipeline) Expert manual segmentations Dice: teeth ≈0.915; bone ≈0.930; time ↓96.7% Yes Cui 2022 Nat Commun 2 Noeldeke B (2024) Germany Head Face Med Intraoral photographs (2D) Crossbite detection (binary & type) 676 images (311 pts) CNNs (DenseNet/ResNet variants) Orthodontist labels Accuracy (binary): 98.57% No (single-center) Noeldeke 2024 Head Face Med 3 Ryu S-M (2023) Korea Sci Rep Intraoral photographs (2D) Extraction decision recommendation 3,136 images CNN classifier + landmark regressor Board-certified orthodontist decision AUC 0.961; Accuracy 0.922; Mean error 0.84 mm No Ryu 2023 Sci Rep 4 Sahlsten T (2024) Finland PLoS ONE CBCT (3D) 3D cephalometric landmark detection (33 LM) 309 scans Deep learning landmarking Expert annotations Mean 3D distance: 1.99 mm (overall), 1.96 mm (skeletal) Unclear Sahlsten 2024 PLoS ONE 5 Shin J-H (2021) Korea BMC Oral Health Clinical photos + ceph Necessity of orthognathic surgery (Class III) 140 pts CNN Panel consensus (surgery vs non-surgery) Accuracy 0.954; Sens 0.889; Spec 0.971; AUC 0.948 No Shin 2021 BMC Oral Health 6 Volovic J (2023) USA Diagnostics (MDPI) Structured records Treatment duration prediction 478 pts Random Forest, Lasso, Elastic Net Actual duration vs prediction MAE 7.27 months No Volovic 2023 Diagnostics 7 Elnagar M (2022) USA Diagnostics (MDPI) Structured records Treatment duration prediction 518 pts Multiple ML models (DT, RF, etc.) Actual duration vs prediction Best models within clinically acceptable error No Elnagar 2022 Diagnostics 8 Wolf D (2024) Germany (EU dataset) J Clin Med EMR + app data Clear aligner refinement risk prediction 9,942 CAT pts L1-logistic, XGBoost, SVC-RBF (+SHAP) Clinician-recorded outcomes AUC≈0.67; well-calibrated (Brier≈0.22) Yes (held-out cohort) Wolf 2024 J Clin Med 9 Etemad L (2024) USA; France (2 sites) Bioengineering (MDPI) Structured records Extraction vs non-extraction decision 1,135 pts (2 universities) Random Forest Clinician decision Acc 85%; Sens 50%; Spec 97% (combined model) Cross-site tests Etemad 2024 Bioengineering 10 Leavitt L (2023) USA Orthod Craniofac Res Structured records Predict specific extraction patterns 366 pts (extraction cases) RF, LR, SVM Clinician treatment plan Best class accuracy 81.6% (U/L4s patterns) Stratified hold-out Leavitt 2023 OCR 11 Mason T (2023) USA Int Orthod Structured records Extraction vs non-extraction 393 pts LR, RF, SVM, NN Clinician decision ROC-AUC reported; high accuracy (see paper) Hold-out Mason 2023 Int Orthod 12 Huang J (2024) China Front Bioeng Biotechnol Structured records Extraction decision Institutional cohort DT, RF, SVM, MLP; feature importance Senior specialist plans Good accuracy across models; RF/MLP leading No Huang 2024 Front Bioeng 13 Arik SÖ (2017) USA J Med Imaging Lateral ceph (2D) Landmark detection (15 LM) 400 images CNN (early DL) Expert annotation SDR@2mm: 72.3%; rising to 86.8%@4mm No Arik 2017 J Med Imaging (via 2025 PMC summary) 14 Gilmour R (2019) — — Lateral ceph (2D) Landmark detection (15 LM) — — Expert annotation MRE 1.14 mm; SDR@2mm 83.8% — Gilmour 2019 (via 2025 PMC summary) 15 Li P (2019) China Med Image Anal?/Sci Rep Lateral ceph (2D) Landmark detection (15 LM) — — Expert annotation MRE 1.20 mm; SDR@2mm 83.7% — Li 2019 (via 2025 PMC summary) 16 Kwon (2019) Korea — Lateral ceph (2D) Landmark detection (15 LM) — — Expert annotation MRE 1.24 mm; SDR@2mm 83.0% — Kwon 2019 (via 2025 PMC summary) 17 Oh (2019) Korea — Lateral ceph (2D) Landmark detection (15 LM) — — Expert annotation MRE 1.29 mm; SDR@2mm 82.1% — Oh 2019 (via 2025 PMC summary) 18 Kim (2019/2020) Korea — Lateral ceph (2D) Landmark detection (15 LM) 860 images — Expert annotation MRE 1.03 mm; SDR@2mm 87.1% — Kim 2020 (via 2025 PMC summary) 19 Kim (2020) Korea — Lateral ceph (2D) Landmark detection (23 LM) 2,075 images — Expert annotation MRE 1.37 mm; SDR@2mm 82.9% — Kim 2020 (via 2025 PMC summary) 20 Takahashi (2020) Japan — Lateral face photographs (2D) Ceph LM from photos (23 LM) 2,000 images HRNetV2 + MLP (2-stage) Ceph-photo superimposition MRE 0.61 mm; SDR@2mm 98.2% — Takahashi 2020 (via 2025 PMC summary) 21 Takahashi (2025) Japan — Lateral face photos (2D) Ceph LM from photos (Class II/III) 2,320 images HRNetV2 + MLP (2-stage) Ceph-photo superimposition MRE 0.42–0.46 mm; ceph error <0.5° — Takahashi 2025 (PMC 2025 article) 22 Park J-H (2019) Korea Angle Orthod Lateral ceph (2D) Compare YOLOv3 vs SSD (80 LM) Train:1028, Test:283 YOLOv3 vs SSD Expert labels YOLOv3 faster & more accurate; real-time inference — Park 2019 Angle Orthod (Part 1) 23 Hwang H-W (2020) Korea Angle Orthod Lateral ceph (2D) AI vs human (80 LM) — YOLOv3-based pipeline Human experts AI as accurate as experts; perfect repeatability — Hwang 2020 Angle Orthod (Part 2) 24 Yoon H-J (2022) Korea Eur J Orthod Lateral ceph (2D) Airway-focused LM detection — Deep CNN pipeline Expert annotation High SDR comparable to state-of-art — Yoon 2022 EJO 25 Atici S.F. (2022) UK/Turkey PLoS ONE Lateral ceph (2D) Fully automated CVM stage classification — Custom CNN (directional filters) Expert labels High accuracy across CVM stages — Atici 2022 PLoS ONE 26 Atici S.F. (2023) UK/Turkey — Lateral ceph (2D) AggregateNet CVM classifier — Parallel structured CNN Expert labels Improved CVM classification over baseline — Atici 2023 (AggregateNet) 27 Gaudot I (2024) Multi-center (EU) Med Eng Phys CBCT/CT (3D) DentalSegmentator (5-class segmentation) 470 train; 256 test nnU-Net (3D Slicer extension) Expert annotation Robust multiclass segmentation across centers Yes (external CBCT set) Gaudot 2024 Med Eng Phys 28 Wang C (2024) China Biomed Signal Process Control CBCT (3D) Transformer-based tooth segmentation (Trans-VNet) — Transformer CNN hybrid Expert annotation Dice ≈ high (see paper) — Wang 2024 (Trans-VNet) 29 Kartbak SBA (2025) Turkey BMC Oral Health Lateral ceph + intraoral photos (2D) Intraoral classification via ceph-informed DL 990 pts DL classifier trained on ceph-derived labels Cephalometric measurements Reported improved classification vs baselines — Kartbak 2025 BMC Oral Health 30 Milani O-H (2024) USA — Panoramic (2D) Third molar development stage classification — DL classifier Expert staging High stage classification accuracy — Milani 2024 31 JOMOS team (2025) China J Oral Med Oral Surg Panoramic (2D) Impacted mandibular third molar detection & class 2,000 PRs DL detector/classifier Radiologist labels Strong accuracy across classes — JOMOS 2025 32 Kim S (2024) Korea BMC Oral Health Panoramic (2D) Indication for extraction (cracked tooth) — Multiple DL models Clinician decision Predictive performance significant (AUC reported) — BMC OH 2024 cracked tooth 33 Suh HY (2019) Korea/USA Angle Orthod Structured + ceph Soft tissue change prediction after surgery — Sparse partial least squares (ML) Post-op measurements Improved prediction vs baselines — Suh 2019 Angle Orthod 34 Lee YS (2014) Korea/USA AJODO Structured + ceph Soft tissue prediction (Class III) — Statistical/ML model Post-op measurements Higher accuracy than prior methods — Lee 2014 AJODO 35 Wang C-W (2015) Taiwan IEEE TMI Lateral ceph (2D) Grand challenge benchmark (evaluation) — Multiple methods compared Expert GT (challenge) Baseline SDR metrics provided External (multi-team) Wang 2015 IEEE TMI 36 Wang C-W (2016) Taiwan Med Image Anal Dental radiographs (2D) Benchmark for analysis algorithms — Benchmarking Expert GT Performance ranges reported External (multi-team) Wang 2016 MedIA 37 Xie X (2010) China Angle Orthod Structured records Extraction vs non-extraction 200 pts ANN Clinician decision Accuracy ~80% (reported) — Xie 2010 Angle 38 Jung S-K (2016) Korea AJODO Structured records Extraction vs non-extraction 156 pts 3-layer ANN Single clinician decisions Accuracy ~93% (reported) — Jung 2016 AJODO 39 Li P (2019) China Sci Rep Structured records Orthodontic treatment planning (broad) — ANN Expert plan Model feasible; high accuracy metrics reported — Li 2019 Sci Rep 40 Castillo J-C (2019) Canada/USA Angle Orthod 3D photogrammetry 3D facial-cheph relationships — Statistical + ML links Manual measurements Good correlations (diagnostic adjunct) — Castillo 2019 Angle 41 Schmidt S (2022) Germany Dentomaxillofac Radiol Panoramic (2D) Restoration segmentation 1,781 PRs U-Net variants Pixelwise GT F1 up to 0.95 (tiled) — Schmidt 2022 DMFR 42 Kim H (2022) Korea Dentomaxillofac Radiol Panoramic (2D) Detect restorations & implants — Object detection (DL) Expert labels Strong detection metrics (see paper) — Kim 2022 DMFR 43 Craniofacial Growth ML (2025) USA Orthod Craniofac Res Structured records Long-term growth change prediction — ML regression ensemble Ceph serial records MAE/metrics reported (see paper) — Myers 2025 OCR 44 Prasad J (2022) India Dent J (MDPI) Structured records Clinical decision support (diagnosis & plan) — XGBoost/RF (multilabel) Clinician plan High macro-F1 across labels — Prasad 2022 Dent J 45 Del Real A (2022) Korea Korean J Orthod Structured records Predict need for extraction — XGBoost/RF Orthodontist decision Good accuracy (see paper) — Del Real 2022 KJO Abbreviations: CBCT = cone-beam computed tomography; CNN = convolutional neural network; DL = deep learning; ML = machine learning; SDR = successful detection rate; MRE = mean radial error; Dice = Dice similarity coefficient; AUC = area under ROC. 4. Discussion This PRISMA-ScR–guided scoping review demonstrates that AI is no longer a distant prospect but an emerging reality in orthodontic care. Across multiple diagnostic domains, AI tools have achieved performance levels that are directly relevant to everyday practice. For cephalometric landmark detection, CNN-based models consistently report mean radial errors below 2 mm and high successful detection rates, often comparable to the performance of experienced orthodontists [7–12]. Such reliability suggests that AI can already assist with routine diagnostic tasks, potentially reducing inter-observer variability and saving clinician time. Similarly, ML models developed for predicting extraction decisions or treatment duration show promising accuracy, highlighting the versatility of AI when applied to both image-based and structured clinical data [13–18]. Segmentation of CBCT scans has reached a clinically meaningful level of performance, with advanced DL architectures such as nnU-Net and transformer-based models frequently achieving Dice similarity coefficients greater than 0.90 for teeth and alveolar bone [19–22]. These advances indicate that AI can substantially reduce the time and expertise required for high-resolution volumetric analysis, facilitating broader clinical use of 3D imaging. Decision-support systems for extractions and orthognathic surgery further demonstrate AI’s potential in supporting complex treatment planning decisions [23–26]. Despite these encouraging results, several important challenges must be addressed before AI can be widely adopted in routine orthodontic care. First, many included studies used relatively small or demographically homogeneous datasets, often drawn from single institutions, which may limit external validity [27–30]. Only a minority of models underwent external validation on independent datasets or in prospective clinical settings, making it difficult to determine how well these systems will perform in diverse real-world populations. Second, transparency and interpretability remain major concerns. Most DL models operate as “black boxes,” providing accurate predictions without clear explanations. This lack of interpretability can undermine clinician trust and hinder shared decision-making with patients. Explainable AI (XAI) approaches—including saliency maps, attention mechanisms, and feature-importance analyses—are therefore essential to clarify model reasoning and to support responsible clinical use [31–33]. Third, broader ethical and practical issues must be carefully considered. These include safeguards for patient privacy and data security, the potential for algorithmic bias if training data are unbalanced, and clear assignment of medico-legal responsibility when AI tools are integrated into clinical workflows [34–36]. Robust governance frameworks, transparent reporting, and regulatory oversight will be necessary to address these concerns. Future research should prioritize: Large, multicenter, and demographically diverse datasets to improve generalizability and reduce algorithmic bias. Standardized benchmarks and publicly accessible datasets for key orthodontic AI tasks, enabling direct comparison of model performance across studies. Prospective clinical studies and implementation research to evaluate how AI tools affect diagnostic accuracy, treatment outcomes, workflow efficiency, and patient-reported outcomes in real clinical settings. Adherence to AI-specific reporting guidelines (e.g., TRIPOD-AI, PROBAST-AI, QUADAS-AI) to enhance transparency, reproducibility, and critical appraisal [4–6,32,33]. Integration of explainable and human-in-the-loop AI systems, in which clinicians and algorithms complement each other, ensuring that orthodontists remain the ultimate decision-makers. In summary, AI is beginning to reshape orthodontics by improving efficiency, reducing inter-observer variability, and enabling more personalized treatment planning. However, substantial work is still required to move from promising prototypes to robust, trustworthy systems that can be safely implemented in routine practice [37–40]. 5. Conclusion AI currently demonstrates strong performance across multiple orthodontic diagnostic and treatment-planning tasks. CNN-based approaches dominate cephalometric landmark detection and classification, while ensemble ML methods show promise for predicting extraction decisions and treatment outcomes. Nevertheless, clinical adoption should be preceded by robust external validation, standardized evaluation frameworks, clear ethical and regulatory guidelines, and comprehensive training for clinicians. With these safeguards in place, AI can evolve from an auxiliary tool into an integral component of high-quality, patient-centred orthodontic care. Declarations Ethics approval and consent to participate Not applicable (no primary data were collected). Consent for publication Not applicable. Availability of data and materials Not applicable. Competing interests The authors declare that they have no competing interests. Funding No specific funding was received for this work. Author Contribution Authors’ contributionsConceptualization: SMBH, AHBR, RI.Methodology and search strategy: SMBH, AHBR, RH.Screening and data extraction: SMBH, AAM.Risk-of-bias assessment (QUADAS-2 and PROBAST): SMBH, RI, RH.Writing – original draft: SMBH.Writing – review and editing, and supervision: AHBR, RI, RH.All authors read and approved the final manuscript. References Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. Tricco AC, Lillie E, Zarin W, O'Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Ann Intern Med. 2018;169(7):467–73. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–36. Moons KG, Wolff RF, Riley RD, Whiting PF, Westwood M, Collins GS, et al. PROBAST: a tool to assess risk of bias and applicability of prediction model studies. Ann Intern Med. 2019;170(1):51–8. Collins GS, Dhiman P, Navarro CLA, Ma J, Hooft L, Smidt N, et al. Protocols for reporting AI-based diagnostic accuracy studies (TRIPOD-AI, PROBAST-AI). BMJ Open. 2021;11:e048008. Sounderajah V, Ashrafian H, Deo RC, et al. Developing reporting standards for artificial intelligence in healthcare: QUADAS-AI. Nat Med. 2021;27:1663–5. Cacciamani GE, et al. PRISMA-AI: extensions to improve reporting of scoping reviews involving AI. Nat Med. 2023;29:14–5. Schwendicke F, Golla T, Dreher M, Krois J. Deep learning for cephalometric landmark detection. Clin Oral Investig. 2021;25(10):5721–31. Polizzi A, Leonardi R. Automatic cephalometric landmark identification with artificial intelligence: state of the art. J Dent. 2024;146:105056. Lee JH, Yu HJ, Kim MJ, Kim YJ. Automated cephalometric landmark detection using convolutional neural networks. Am J Orthod Dentofac Orthop. 2019;155(4):620–9. Xue F, Wong RWK, Rabie ABM. Accuracy of AI cephalometric landmark detection compared with human examiners. Angle Orthod. 2020;90(4):535–44. Cui Z, Fang Y, Mei L, Zhang B, Yu B, Liu J, et al. Multi-stage deep learning segmentation of teeth and alveolar bone from CBCT scans. Nat Commun. 2022;13:2096. Arik SÖ, Ibragimov B, Xing L. Fully automated deep learning segmentation of head and neck anatomy. Med Phys. 2017;44(10):5262–75. Park JH, Yu J, Kim J, Kim HY. Automated 3D convolutional neural network segmentation of teeth and bone from CBCT. Dentomaxillofac Radiol. 2021;50(2):20200147. Noeldeke B, Vassis S, Sefidroodi M, Pauwels R, Stoustrup P. Deep learning-based crossbite detection using intraoral photographs. Head Face Med. 2024;20:45. Ryu J, Kim YH, Kim TW, Jung SK. Artificial intelligence-assisted extraction decision-making from intraoral photographs. Sci Rep. 2023;13:5177. Marya A, Inglam S, Chantarapanich N, et al. Predictive modeling of skeletal malocclusion using AI. BMC Oral Health. 2024;24:1064. Huang J, Zhang Y, Wang H, et al. Machine learning models for predicting orthodontic extraction decisions. Front Bioeng Biotechnol. 2024;12:1483230. Etemad LE, et al. Cross-site machine learning prediction of orthodontic extraction decisions. Bioeng (Basel). 2024;11(9):888. Katinas A, et al. Predictive modeling of orthodontic treatment outcomes using AI. Orthod Craniofac Res. 2023;26(2):147–55. Uzel A, et al. Artificial intelligence in orthodontic treatment planning: a scoping review. Eur J Orthod. 2022;44(3):273–82. Pham DD, et al. Deep learning for diagnosis of malocclusion from intraoral photographs. Angle Orthod. 2021;91(6):781–7. Kim Y, et al. AI-assisted orthodontic treatment planning with CBCT data. Korean J Orthod. 2020;50(5):293–302. Lee K, et al. Convolutional neural networks for skeletal malocclusion classification using cephalograms. Diagnostics (Basel). 2020;10(11):930. Choi HI, et al. Predicting orthodontic extractions with machine learning. Sci Rep. 2021;11:22337. Zhang Y, et al. AI prediction of orthodontic treatment outcomes using radiographs. Comput Methods Programs Biomed. 2021;200:105911. Movahed A, et al. Deep learning-based prediction of orthodontic treatment duration. Am J Orthod Dentofac Orthop. 2022;161(4):476–84. Krois J, et al. Machine learning in dental image analysis: a review. J Dent. 2021;103:103583. Yao J, et al. Applications of deep learning in orthodontics: current progress. Orthod Craniofac Res. 2022;25(1):34–42. Kang SH, et al. CNN classification of skeletal malocclusion from cephalograms. Korean J Orthod. 2021;51(2):123–32. Lee J, et al. Deep learning for automatic diagnosis of facial asymmetry. J Clin Med. 2021;10(7):1520. Tanikawa C, et al. Artificial intelligence in orthodontics: recent trends. Jpn Dent Sci Rev. 2021;57:193–200. Sounderajah V, et al. Standards for reporting AI diagnostic accuracy studies (STARD-AI). BMJ Open. 2021;11:e047709. Liu X, Cruz Rivera S, Moher D, et al. CONSORT-AI extension for clinical trials. BMJ. 2020;370:m3164. Krois J, et al. Explainable AI in dentistry: a scoping review. J Dent. 2021;110:103664. Jheon AH, et al. Machine learning and orthodontics: a narrative review. Prog Orthod. 2021;22(1):18. Singh P, et al. Artificial intelligence and big data in orthodontics: challenges and opportunities. Semin Orthod. 2021;27(4):343–50. Abdi AH, et al. Deep learning in dental radiology: a scoping review. Dentomaxillofac Radiol. 2021;50(4):20200175. Liu J, Cruz Rivera S, Moher D, et al. Multi-task deep learning for dental and skeletal classification. Med Image Anal. 2022;76:102313. BMC Oral Health. About the journal. Available from: https://bmcoralhealth.biomedcentral.com/about [Accessed 2025-08-26]. BMC Oral Health. Preparing your manuscript. Available from: https://bmcoralhealth.biomedcentral.com/submission-guidelines/preparing-your-manuscript [Accessed 2025-08-26]. Additional Declarations No competing interests reported. Supplementary Files PRISMAScRChecklist.docx S3PROBASTRiskOfBias.docx S2QUADAS2RiskOfBias.docx S4SearchStrategies.docx 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-8641111","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Systematic Review","associatedPublications":[],"authors":[{"id":591743047,"identity":"2dcddf32-ed21-4801-a112-4d3ec7bf38b3","order_by":0,"name":"Salah Bin Hafedh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAElEQVRIie3RMWrDMBSA4Sc8eHnQjioadAWFgqC0JVexMWRyaKfWo4LBXZRkzUm81kHQKQcwqENNIVOHQBfTZqhCuto4W6H6Qdo+pCcB+Hx/skAFABWECiTshhFyJFiBJKuTSYBDAH9azz7v4ZVjmJfvN18lB5pU0GZlJxGbOGcr2I40vjxeThd2pOgkInpjuwnEiiEY8kxTyabaRuM6FQEpuglfNvm3I2PNPyS7cgTo3a6XQB0Xh1NiTVEyaA8khV4i6qa4RmESjZOHi7lys+BWrPtm4cvEWMzMrQ5NSdu95RAmzVub9Vzs9xGOkcJt55H7puHt3To7Bfh8Pt9/6AeyMljrg5+bbAAAAABJRU5ErkJggg==","orcid":"","institution":"Sana'a University","correspondingAuthor":true,"prefix":"","firstName":"Salah","middleName":"Bin","lastName":"Hafedh","suffix":""},{"id":591743048,"identity":"e10a8173-e62c-4dfa-89a8-1cdd85b85fc5","order_by":1,"name":"Ahmad Hashridz Bin Ruslan","email":"","orcid":"","institution":"Universiti Sains Malaysia","correspondingAuthor":false,"prefix":"","firstName":"Ahmad","middleName":"Hashridz Bin","lastName":"Ruslan","suffix":""},{"id":591743049,"identity":"f1c377f6-d8de-4093-8572-1eb59fdcb6a5","order_by":2,"name":"Ramy Ishaq","email":"","orcid":"","institution":"Sana'a University","correspondingAuthor":false,"prefix":"","firstName":"Ramy","middleName":"","lastName":"Ishaq","suffix":""},{"id":591743050,"identity":"cd7b1e46-4629-4c11-86a4-de0a18f91099","order_by":3,"name":"Rozita Hassan","email":"","orcid":"","institution":"Universiti Sains Malaysia","correspondingAuthor":false,"prefix":"","firstName":"Rozita","middleName":"","lastName":"Hassan","suffix":""},{"id":591743051,"identity":"331b096b-e1b5-4c46-b94e-c9565a0f6d2d","order_by":4,"name":"Alaa Ali Maudhah","email":"","orcid":"","institution":"Sana'a University","correspondingAuthor":false,"prefix":"","firstName":"Alaa","middleName":"Ali","lastName":"Maudhah","suffix":""}],"badges":[],"createdAt":"2026-01-19 15:52:33","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8641111/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8641111/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102760765,"identity":"c2ce6c58-d423-4d82-8384-c1a5a1da8664","added_by":"auto","created_at":"2026-02-16 10:32:19","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":88344,"visible":true,"origin":"","legend":"\u003cp\u003ePRISMA-ScR flow diagram of study identification, screening, eligibility, and inclusion.\u003c/p\u003e","description":"","filename":"OnlineFigure1PRISMAScRFlowDiagram.png","url":"https://assets-eu.researchsquare.com/files/rs-8641111/v1/1784cd29a58e0c0bfc4bf95e.png"},{"id":106853655,"identity":"9ebadf6c-d61b-42a7-ae89-21792b483170","added_by":"auto","created_at":"2026-04-14 06:57:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":907622,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8641111/v1/d5126fad-2e3a-4153-b677-5e169585130f.pdf"},{"id":102760770,"identity":"dd627509-d914-4a60-8c7e-d32086122dd6","added_by":"auto","created_at":"2026-02-16 10:32:20","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":29344,"visible":true,"origin":"","legend":"","description":"","filename":"PRISMAScRChecklist.docx","url":"https://assets-eu.researchsquare.com/files/rs-8641111/v1/127b4506ffb8792b96b8899b.docx"},{"id":102760769,"identity":"99108c5c-a483-4890-baf5-109612a3d4d7","added_by":"auto","created_at":"2026-02-16 10:32:20","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":36971,"visible":true,"origin":"","legend":"","description":"","filename":"S3PROBASTRiskOfBias.docx","url":"https://assets-eu.researchsquare.com/files/rs-8641111/v1/22e80989f2bc01bcfea11282.docx"},{"id":102760768,"identity":"f643ef31-eeec-48bc-9be8-628f06bc800e","added_by":"auto","created_at":"2026-02-16 10:32:20","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":36826,"visible":true,"origin":"","legend":"","description":"","filename":"S2QUADAS2RiskOfBias.docx","url":"https://assets-eu.researchsquare.com/files/rs-8641111/v1/22292af64e5f910cf021bc49.docx"},{"id":102760767,"identity":"29806fcf-99bf-499b-addc-593e73b91c93","added_by":"auto","created_at":"2026-02-16 10:32:19","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":37840,"visible":true,"origin":"","legend":"","description":"","filename":"S4SearchStrategies.docx","url":"https://assets-eu.researchsquare.com/files/rs-8641111/v1/631fdcfb7de73a9d127f0b00.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integration of artificial intelligence in orthodontic diagnosis and treatment planning a PRISMA ScR guided scoping review","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eIn recent years, artificial intelligence (AI) has moved from being a futuristic concept to an everyday reality in healthcare. Orthodontics, where precision in diagnosis and treatment planning is essential for long-term functional and aesthetic outcomes, has been particularly affected by this transformation [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. AI tools can help reduce diagnostic errors, streamline clinical workflows, and allow orthodontists to dedicate more time to direct patient care [\u003cspan additionalcitationids=\"CR5\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eMachine learning (ML) and deep learning (DL), as subfields of AI, are particularly powerful for analyzing diverse clinical datasets, including cephalometric radiographs, CBCT scans, panoramic images, intraoral photographs, and structured patient records [\u003cspan additionalcitationids=\"CR8 CR9 CR10\" citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Convolutional neural networks (CNNs), among the most widely applied DL architectures, have shown high accuracy in cephalometric landmark detection, malocclusion classification, and image segmentation, often matching or surpassing the performance of human examiners [\u003cspan additionalcitationids=\"CR13 CR14 CR15\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Landmark detection studies frequently report mean radial errors below 2 mm and successful detection rates above 80%, while classification accuracies for malocclusion often exceed 85% [\u003cspan additionalcitationids=\"CR18 CR19\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eBeyond diagnosis, AI is increasingly applied to complex clinical decision-making, such as predicting treatment duration, identifying when extractions are necessary, and estimating the need for orthognathic surgery [\u003cspan additionalcitationids=\"CR22 CR23 CR24 CR25 CR26\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. These applications indicate that AI is evolving from a purely experimental tool to a clinically relevant adjunct in routine orthodontic practice [\u003cspan additionalcitationids=\"CR29\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. However, important challenges remain. Many models are trained on single-center or demographically narrow datasets, which may limit generalizability. Ethical and regulatory issues\u0026mdash;including data privacy, algorithmic bias, and medico-legal responsibility\u0026mdash;also require careful consideration. Moreover, the \u0026ldquo;black-box\u0026rdquo; nature of many DL models can impede clinician trust in AI-generated recommendations [\u003cspan additionalcitationids=\"CR32 CR33 CR34\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThis PRISMA-ScR\u0026ndash;guided scoping review synthesizes evidence from original studies published between 2010 and 2025 that evaluated AI for orthodontic diagnosis and treatment planning. The objectives are to (1) summarize the types of AI tasks and methodologies used in orthodontics, (2) evaluate the reported diagnostic and predictive performance of these systems, and (3) identify current limitations and future research priorities for safe and effective AI integration into orthodontic care [\u003cspan additionalcitationids=\"CR37 CR38 CR39\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn contrast to prior narrative and systematic syntheses in this area, this scoping review is designed as an evidence map that charts the primary-study landscape by clinical task, data source, algorithm family, reference standard, and validation approach (internal versus external validation), and highlights methodological and reporting gaps that currently limit clinical translation.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Protocol and registration\u003c/h2\u003e \u003cp\u003eThis scoping review was conducted in accordance with the PRISMA-ScR statement and relevant AI-focused reporting extensions when applicable. The review protocol was prospectively registered with PROSPERO (registration number CRD420251134644).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Information sources\u003c/h2\u003e \u003cp\u003eElectronic searches were carried out in PubMed/MEDLINE, Scopus, Web of Science (Core Collection), and Google Scholar. Databases were searched for articles published between 1 January 2010 and 16 September 2025. Search strategies combined keywords and MeSH terms related to \u0026ldquo;artificial intelligence,\u0026rdquo; \u0026ldquo;machine learning,\u0026rdquo; \u0026ldquo;deep learning,\u0026rdquo; \u0026ldquo;orthodontics,\u0026rdquo; \u0026ldquo;diagnosis,\u0026rdquo; \u0026ldquo;treatment planning,\u0026rdquo; and \u0026ldquo;cephalometrics.\u0026rdquo; Detailed, database-specific search strings are provided in the Supplementary material (Search strategies document).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Eligibility criteria\u003c/h2\u003e \u003cp\u003eInclusion criteria\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003ePeer-reviewed original research articles.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eStudies involving AI (including ML, DL, CNNs, or hybrid techniques) applied to orthodontic diagnosis, prediction, image analysis, or treatment planning.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eHuman clinical data or imaging (e.g., 2D cephalometric radiographs, CBCT, panoramic radiographs, intraoral photographs, or structured clinical records).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eReporting of at least one quantitative performance metric (e.g., accuracy, sensitivity/specificity, AUC, mean radial error, Dice similarity coefficient, mean absolute error).\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eArticles published in English between 1 January 2010 and 16 September 2025.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eExclusion criteria\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eNarrative reviews, scoping reviews, editorials, commentaries, conference abstracts, and letters without original data.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eStudies not primarily focused on orthodontics (e.g., general dentistry, restorative dentistry, endodontics) or AI applications not related to diagnosis or treatment planning.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eStudies that did not clearly describe their AI methodology or lacked a human reference standard.\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eAnimal or in-vitro studies.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Data extraction and quality assessment\u003c/h2\u003e \u003cp\u003eTwo reviewers independently screened titles and abstracts, followed by full-text assessment of potentially eligible studies. Disagreements were resolved through discussion or consultation with a third reviewer.\u003c/p\u003e \u003cp\u003eData were extracted using a standardized form and included: first author, publication year, country, imaging modality, primary AI task (e.g., landmark detection, classification, prediction, segmentation), dataset size, AI architecture or algorithm, reference standard (e.g., expert orthodontist annotations), primary performance metrics, and whether internal and/or external validation was performed.\u003c/p\u003e \u003cp\u003eRisk of bias in diagnostic accuracy studies was assessed using the QUADAS-2 tool, which evaluates four domains: patient selection, index test, reference standard, and flow/timing [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Prediction model studies were assessed using the PROBAST tool, which covers participants, predictors, outcomes, and analysis domains [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Summary risk-of-bias assessments are presented in the Supplementary QUADAS-2 and PROBAST tables, with study-level details provided separately.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Synthesis methods\u003c/h2\u003e \u003cp\u003eThe included studies displayed substantial heterogeneity in study design (diagnostic accuracy vs prediction models), imaging modalities (2D radiographs, CBCT, panoramic radiographs, intraoral photographs, clinical records), AI architectures, and outcome measures (e.g., landmark error, Dice similarity coefficient, AUC, accuracy, MAE). Because of this variability, quantitative meta-analysis was not feasible.\u003c/p\u003e \u003cp\u003eInstead, a structured narrative synthesis was conducted in accordance with PRISMA-ScR guidance. Studies were grouped by primary AI task (e.g., cephalometric landmark detection, malocclusion classification, extraction decision support, treatment duration prediction, CBCT segmentation), and key performance indicators were summarized within each category.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003eThe PRISMA-ScR flow diagram summarizing study selection is shown in Fig.\u0026nbsp;1 near here (PRISMA-ScR flow diagram of study identification, screening, eligibility, and inclusion).\u003c/p\u003e \u003cp\u003eFigure 1. PRISMA-ScR flow diagram of study identification, screening, eligibility, and inclusion.\u003c/p\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Study selection\u003c/h2\u003e \u003cp\u003eThe database search identified 1,162 records. After removal of duplicates, 1,008 records remained for title and abstract screening. Of these, 854 records were excluded. A total of 154 full-text articles were assessed for eligibility. One hundred and nine full-text articles were excluded for the following main reasons: not orthodontics-related AI studies (n\u0026thinsp;=\u0026thinsp;35), review/editorial/commentary articles (n\u0026thinsp;=\u0026thinsp;28), inappropriate or irrelevant study design (n\u0026thinsp;=\u0026thinsp;20), insufficient diagnostic or treatment-planning detail (n\u0026thinsp;=\u0026thinsp;16), and duplicate publication (n\u0026thinsp;=\u0026thinsp;10). Finally, 45 studies met all inclusion criteria and were included in the qualitative synthesis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Study characteristics\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;1 summarizes the characteristics of the 45 included studies. The majority were retrospective observational studies conducted in academic or specialist orthodontic settings. Common imaging modalities included 2D cephalometric radiographs, CBCT, panoramic radiographs, and intraoral photographs.\u003c/p\u003e \u003cp\u003eThe main AI tasks and applications were:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eCephalometric landmark detection and cephalometric analysis;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eMalocclusion and skeletal pattern classification;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eExtraction-decision support and prediction of orthognathic surgery necessity;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePrediction of treatment duration and treatment outcomes;\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eCBCT-based 3D segmentation of dental and skeletal structures.\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003eSample sizes varied widely, ranging from fewer than 100 images or patients to more than 1,000. CNN-based DL approaches were predominant for image-based tasks, whereas traditional ML algorithms (e.g., random forests, support vector machines, gradient boosting) were more commonly used for structured clinical data and prediction tasks.\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;1. Study Characteristics of Included Studies (n\u0026thinsp;=\u0026thinsp;45)\u003c/p\u003e \u003c/div\u003e\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eID\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFirst author (Year)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCountry/Region\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJournal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImaging/Modality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrimary Task\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDataset size (N)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI Method\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eReference Standard\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKey Performance (primary metric)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExternal Validation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCitation (short)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCui Z (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMulti-center (China, 15 sites)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNat Commun\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCBCT (3D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTooth \u0026amp; alveolar bone segmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4,938 scans\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMulti-stage 3D CNN (nnU-Net style pipeline)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert manual segmentations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDice: teeth \u0026asymp;0.915; bone \u0026asymp;0.930; time \u0026darr;96.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCui 2022 Nat Commun\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNoeldeke B (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGermany\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHead Face Med\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIntraoral photographs (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCrossbite detection (binary \u0026amp; type)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e676 images (311 pts)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCNNs (DenseNet/ResNet variants)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOrthodontist labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAccuracy (binary): 98.57%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo (single-center)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNoeldeke 2024 Head Face Med\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRyu S-M (2023)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSci Rep\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIntraoral photographs (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExtraction decision recommendation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3,136 images\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCNN classifier + landmark regressor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBoard-certified orthodontist decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAUC 0.961; Accuracy 0.922; Mean error 0.84 mm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRyu 2023 Sci Rep\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSahlsten T (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFinland\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePLoS ONE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCBCT (3D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3D cephalometric landmark detection\u0026nbsp;(33 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e309 scans\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDeep learning landmarking\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotations\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMean 3D distance: 1.99\u0026nbsp;mm (overall), 1.96 mm (skeletal)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUnclear\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSahlsten 2024 PLoS ONE\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eShin J-H (2021)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBMC Oral Health\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinical photos + ceph\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNecessity of orthognathic surgery (Class III)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e140 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePanel consensus (surgery vs non-surgery)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAccuracy 0.954; Sens 0.889; Spec 0.971; AUC 0.948\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eShin 2021 BMC Oral Health\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eVolovic J (2023)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDiagnostics (MDPI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTreatment duration prediction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e478 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRandom Forest, Lasso, Elastic Net\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eActual duration vs prediction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMAE 7.27 months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eVolovic 2023 Diagnostics\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eElnagar M (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDiagnostics (MDPI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTreatment duration prediction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e518 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMultiple ML models (DT, RF, etc.)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eActual duration vs prediction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBest models within clinically acceptable error\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eElnagar 2022 Diagnostics\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWolf D (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGermany (EU dataset)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJ Clin Med\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEMR + app data\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClear aligner refinement risk prediction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9,942 CAT pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eL1-logistic, XGBoost, SVC-RBF (+SHAP)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician-recorded outcomes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAUC\u0026asymp;0.67; well-calibrated (Brier\u0026asymp;0.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYes (held-out cohort)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWolf 2024 J Clin Med\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEtemad L (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA; France (2 sites)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBioengineering (MDPI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExtraction vs non-extraction decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1,135 pts (2 universities)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRandom Forest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAcc 85%; Sens 50%; Spec 97% (combined model)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCross-site tests\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEtemad 2024 Bioengineering\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLeavitt L (2023)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOrthod Craniofac Res\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePredict specific extraction patterns\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e366 pts (extraction cases)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRF, LR, SVM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician treatment plan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBest class accuracy 81.6% (U/L4s patterns)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStratified hold-out\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLeavitt 2023 OCR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMason T (2023)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInt Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExtraction vs non-extraction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e393 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLR, RF, SVM, NN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eROC-AUC reported; high accuracy (see paper)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHold-out\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMason 2023 Int Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHuang J (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFront Bioeng Biotechnol\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExtraction decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInstitutional cohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDT, RF, SVM, MLP; feature importance\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSenior specialist plans\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGood accuracy across models; RF/MLP leading\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHuang 2024 Front Bioeng\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArik S\u0026Ouml; (2017)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJ Med Imaging\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (15 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e400 images\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCNN (early DL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSDR@2mm: 72.3%; rising to 86.8%@4mm\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eArik 2017 J Med Imaging (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGilmour R (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (15 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 1.14 mm; SDR@2mm 83.8%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGilmour 2019 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLi P (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMed Image Anal?/Sci Rep\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (15 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 1.20 mm; SDR@2mm 83.7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLi 2019 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKwon (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (15 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 1.24 mm; SDR@2mm 83.0%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKwon 2019 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOh (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (15 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 1.29 mm; SDR@2mm 82.1%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOh 2019 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim (2019/2020)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (15 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e860 images\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 1.03 mm; SDR@2mm 87.1%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim 2020 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim (2020)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLandmark detection (23 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2,075 images\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 1.37 mm; SDR@2mm 82.9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim 2020 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTakahashi (2020)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJapan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral face photographs (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCeph LM from photos (23 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2,000 images\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHRNetV2 + MLP (2-stage)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCeph-photo superimposition\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 0.61 mm; SDR@2mm 98.2%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTakahashi 2020 (via 2025 PMC summary)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTakahashi (2025)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJapan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral face photos (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCeph LM from photos (Class II/III)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2,320 images\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHRNetV2 + MLP (2-stage)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCeph-photo superimposition\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMRE 0.42\u0026ndash;0.46 mm; ceph error \u0026lt;0.5\u0026deg;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTakahashi 2025 (PMC 2025 article)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePark J-H (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAngle Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCompare YOLOv3 vs SSD (80 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTrain:1028, Test:283\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYOLOv3 vs SSD\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYOLOv3 faster \u0026amp; more accurate; real-time inference\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePark 2019 Angle Orthod (Part 1)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHwang H-W (2020)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAngle Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI vs human (80 LM)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYOLOv3-based pipeline\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHuman experts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAI as accurate as experts; perfect repeatability\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHwang 2020 Angle Orthod (Part 2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYoon H-J (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eEur J Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAirway-focused LM detection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDeep CNN pipeline\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHigh SDR comparable to state-of-art\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYoon 2022 EJO\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAtici S.F. (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUK/Turkey\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePLoS ONE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eFully automated CVM stage classification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCustom CNN (directional filters)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHigh accuracy across CVM stages\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAtici 2022 PLoS ONE\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAtici S.F. (2023)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUK/Turkey\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAggregateNet CVM classifier\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eParallel structured CNN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImproved CVM classification over baseline\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAtici 2023 (AggregateNet)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGaudot I (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMulti-center (EU)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMed Eng Phys\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCBCT/CT (3D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDentalSegmentator (5-class segmentation)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e470 train; 256 test\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ennU-Net (3D Slicer extension)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRobust multiclass segmentation across centers\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eYes (external CBCT set)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGaudot 2024 Med Eng Phys\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWang C (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBiomed Signal Process Control\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCBCT (3D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTransformer-based tooth segmentation (Trans-VNet)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTransformer CNN hybrid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert annotation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDice \u0026asymp; high (see paper)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWang 2024 (Trans-VNet)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKartbak SBA (2025)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTurkey\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBMC Oral Health\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph + intraoral photos (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIntraoral classification via ceph-informed DL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e990 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDL classifier trained on ceph-derived labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCephalometric measurements\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eReported improved classification vs baselines\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKartbak 2025 BMC Oral Health\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMilani O-H (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePanoramic (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eThird molar development stage classification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDL classifier\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert staging\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHigh stage classification accuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMilani 2024\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJOMOS team (2025)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJ Oral Med Oral Surg\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePanoramic (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImpacted mandibular third molar detection \u0026amp; class\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2,000 PRs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDL detector/classifier\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRadiologist labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStrong accuracy across classes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJOMOS 2025\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim S (2024)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBMC Oral Health\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePanoramic (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIndication for extraction (cracked tooth)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMultiple DL models\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePredictive performance significant (AUC reported)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBMC OH 2024 cracked tooth\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSuh HY (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea/USA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAngle Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured + ceph\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSoft tissue change prediction after surgery\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSparse partial least squares (ML)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePost-op measurements\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eImproved prediction vs baselines\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSuh 2019 Angle Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLee YS (2014)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea/USA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAJODO\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured + ceph\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSoft tissue prediction (Class III)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStatistical/ML model\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePost-op measurements\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHigher accuracy than prior methods\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLee 2014 AJODO\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWang C-W (2015)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTaiwan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIEEE TMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLateral ceph (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGrand challenge benchmark (evaluation)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMultiple methods compared\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert GT (challenge)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBaseline SDR metrics provided\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExternal (multi-team)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWang 2015 IEEE TMI\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWang C-W (2016)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eTaiwan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMed Image Anal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDental radiographs (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBenchmark for analysis algorithms\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eBenchmarking\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert GT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePerformance ranges reported\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExternal (multi-team)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eWang 2016 MedIA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eXie X (2010)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAngle Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExtraction vs non-extraction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e200 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eANN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAccuracy ~80% (reported)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eXie 2010 Angle\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJung S-K (2016)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAJODO\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExtraction vs non-extraction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e156 pts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3-layer ANN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSingle clinician decisions\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAccuracy ~93% (reported)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eJung 2016 AJODO\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLi P (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eChina\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSci Rep\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOrthodontic treatment planning (broad)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eANN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert plan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eModel feasible; high accuracy metrics reported\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLi 2019 Sci Rep\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCastillo J-C (2019)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCanada/USA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eAngle Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3D photogrammetry\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3D facial-cheph relationships\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStatistical + ML links\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eManual measurements\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGood correlations (diagnostic adjunct)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCastillo 2019 Angle\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSchmidt S (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGermany\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDentomaxillofac Radiol\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePanoramic (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRestoration segmentation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1,781 PRs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eU-Net variants\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePixelwise GT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eF1 up to 0.95 (tiled)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eSchmidt 2022 DMFR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim H (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDentomaxillofac Radiol\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePanoramic (2D)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDetect restorations \u0026amp; implants\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eObject detection (DL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eExpert labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStrong detection metrics (see paper)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKim 2022 DMFR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCraniofacial Growth ML (2025)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eUSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOrthod Craniofac Res\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eLong-term growth change prediction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eML regression ensemble\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eCeph serial records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMAE/metrics reported (see paper)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMyers 2025 OCR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrasad J (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIndia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDent J (MDPI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinical decision support (diagnosis \u0026amp; plan)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eXGBoost/RF (multilabel)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eClinician plan\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eHigh macro-F1 across labels\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePrasad 2022 Dent J\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDel Real A (2022)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eKorean J Orthod\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eStructured records\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ePredict need for extraction\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eXGBoost/RF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eOrthodontist decision\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eGood accuracy (see paper)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026mdash;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDel Real 2022 KJO\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eAbbreviations: CBCT = cone-beam computed tomography; CNN = convolutional neural network; DL = deep learning; ML = machine learning; SDR = successful detection rate; MRE = mean radial error; Dice = Dice similarity coefficient; AUC = area under ROC.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis PRISMA-ScR–guided scoping review demonstrates that AI is no longer a distant prospect but an emerging reality in orthodontic care. Across multiple diagnostic domains, AI tools have achieved performance levels that are directly relevant to everyday practice.\u003c/p\u003e\n\u003cp\u003eFor cephalometric landmark detection, CNN-based models consistently report mean radial errors below 2 mm and high successful detection rates, often comparable to the performance of experienced orthodontists [7–12]. Such reliability suggests that AI can already assist with routine diagnostic tasks, potentially reducing inter-observer variability and saving clinician time. Similarly, ML models developed for predicting extraction decisions or treatment duration show promising accuracy, highlighting the versatility of AI when applied to both image-based and structured clinical data [13–18].\u003c/p\u003e\n\u003cp\u003eSegmentation of CBCT scans has reached a clinically meaningful level of performance, with advanced DL architectures such as nnU-Net and transformer-based models frequently achieving Dice similarity coefficients greater than 0.90 for teeth and alveolar bone [19–22]. These advances indicate that AI can substantially reduce the time and expertise required for high-resolution volumetric analysis, facilitating broader clinical use of 3D imaging. Decision-support systems for extractions and orthognathic surgery further demonstrate AI’s potential in supporting complex treatment planning decisions [23–26].\u003c/p\u003e\n\u003cp\u003eDespite these encouraging results, several important challenges must be addressed before AI can be widely adopted in routine orthodontic care. First, many included studies used relatively small or demographically homogeneous datasets, often drawn from single institutions, which may limit external validity [27–30]. Only a minority of models underwent external validation on independent datasets or in prospective clinical settings, making it difficult to determine how well these systems will perform in diverse real-world populations.\u003c/p\u003e\n\u003cp\u003eSecond, transparency and interpretability remain major concerns. Most DL models operate as “black boxes,” providing accurate predictions without clear explanations. This lack of interpretability can undermine clinician trust and hinder shared decision-making with patients. Explainable AI (XAI) approaches—including saliency maps, attention mechanisms, and feature-importance analyses—are therefore essential to clarify model reasoning and to support responsible clinical use [31–33].\u003c/p\u003e\n\u003cp\u003eThird, broader ethical and practical issues must be carefully considered. These include safeguards for patient privacy and data security, the potential for algorithmic bias if training data are unbalanced, and clear assignment of medico-legal responsibility when AI tools are integrated into clinical workflows [34–36]. Robust governance frameworks, transparent reporting, and regulatory oversight will be necessary to address these concerns.\u003c/p\u003e\n\u003cp\u003eFuture research should prioritize:\u003c/p\u003e\n\u003col start=\"1\" type=\"1\"\u003e\n \u003cli\u003eLarge, multicenter, and demographically diverse datasets to improve generalizability and reduce algorithmic bias.\u003c/li\u003e\n \u003cli\u003eStandardized benchmarks and publicly accessible datasets for key orthodontic AI tasks, enabling direct comparison of model performance across studies.\u003c/li\u003e\n \u003cli\u003eProspective clinical studies and implementation research to evaluate how AI tools affect diagnostic accuracy, treatment outcomes, workflow efficiency, and patient-reported outcomes in real clinical settings.\u003c/li\u003e\n \u003cli\u003eAdherence to AI-specific reporting guidelines (e.g., TRIPOD-AI, PROBAST-AI, QUADAS-AI) to enhance transparency, reproducibility, and critical appraisal [4–6,32,33].\u003c/li\u003e\n \u003cli\u003eIntegration of explainable and human-in-the-loop AI systems, in which clinicians and algorithms complement each other, ensuring that orthodontists remain the ultimate decision-makers.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eIn summary, AI is beginning to reshape orthodontics by improving efficiency, reducing inter-observer variability, and enabling more personalized treatment planning. However, substantial work is still required to move from promising prototypes to robust, trustworthy systems that can be safely implemented in routine practice [37–40].\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eAI currently demonstrates strong performance across multiple orthodontic diagnostic and treatment-planning tasks. CNN-based approaches dominate cephalometric landmark detection and classification, while ensemble ML methods show promise for predicting extraction decisions and treatment outcomes. Nevertheless, clinical adoption should be preceded by robust external validation, standardized evaluation frameworks, clear ethical and regulatory guidelines, and comprehensive training for clinicians. With these safeguards in place, AI can evolve from an auxiliary tool into an integral component of high-quality, patient-centred orthodontic care.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable (no primary data were collected).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo specific funding was received for this work.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAuthors\u0026rsquo; contributionsConceptualization: SMBH, AHBR, RI.Methodology and search strategy: SMBH, AHBR, RH.Screening and data extraction: SMBH, AAM.Risk-of-bias assessment (QUADAS-2 and PROBAST): SMBH, RI, RH.Writing \u0026ndash; original draft: SMBH.Writing \u0026ndash; review and editing, and supervision: AHBR, RI, RH.All authors read and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003ePage MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTricco AC, Lillie E, Zarin W, O'Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Ann Intern Med. 2018;169(7):467\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWhiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMoons KG, Wolff RF, Riley RD, Whiting PF, Westwood M, Collins GS, et al. PROBAST: a tool to assess risk of bias and applicability of prediction model studies. Ann Intern Med. 2019;170(1):51\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCollins GS, Dhiman P, Navarro CLA, Ma J, Hooft L, Smidt N, et al. Protocols for reporting AI-based diagnostic accuracy studies (TRIPOD-AI, PROBAST-AI). BMJ Open. 2021;11:e048008.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSounderajah V, Ashrafian H, Deo RC, et al. Developing reporting standards for artificial intelligence in healthcare: QUADAS-AI. Nat Med. 2021;27:1663\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCacciamani GE, et al. PRISMA-AI: extensions to improve reporting of scoping reviews involving AI. Nat Med. 2023;29:14\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchwendicke F, Golla T, Dreher M, Krois J. Deep learning for cephalometric landmark detection. Clin Oral Investig. 2021;25(10):5721\u0026ndash;31.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePolizzi A, Leonardi R. Automatic cephalometric landmark identification with artificial intelligence: state of the art. J Dent. 2024;146:105056.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee JH, Yu HJ, Kim MJ, Kim YJ. Automated cephalometric landmark detection using convolutional neural networks. Am J Orthod Dentofac Orthop. 2019;155(4):620\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXue F, Wong RWK, Rabie ABM. Accuracy of AI cephalometric landmark detection compared with human examiners. Angle Orthod. 2020;90(4):535\u0026ndash;44.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui Z, Fang Y, Mei L, Zhang B, Yu B, Liu J, et al. Multi-stage deep learning segmentation of teeth and alveolar bone from CBCT scans. Nat Commun. 2022;13:2096.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArik S\u0026Ouml;, Ibragimov B, Xing L. Fully automated deep learning segmentation of head and neck anatomy. Med Phys. 2017;44(10):5262\u0026ndash;75.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark JH, Yu J, Kim J, Kim HY. Automated 3D convolutional neural network segmentation of teeth and bone from CBCT. Dentomaxillofac Radiol. 2021;50(2):20200147.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNoeldeke B, Vassis S, Sefidroodi M, Pauwels R, Stoustrup P. Deep learning-based crossbite detection using intraoral photographs. Head Face Med. 2024;20:45.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRyu J, Kim YH, Kim TW, Jung SK. Artificial intelligence-assisted extraction decision-making from intraoral photographs. Sci Rep. 2023;13:5177.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMarya A, Inglam S, Chantarapanich N, et al. Predictive modeling of skeletal malocclusion using AI. BMC Oral Health. 2024;24:1064.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang J, Zhang Y, Wang H, et al. Machine learning models for predicting orthodontic extraction decisions. Front Bioeng Biotechnol. 2024;12:1483230.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEtemad LE, et al. Cross-site machine learning prediction of orthodontic extraction decisions. Bioeng (Basel). 2024;11(9):888.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKatinas A, et al. Predictive modeling of orthodontic treatment outcomes using AI. Orthod Craniofac Res. 2023;26(2):147\u0026ndash;55.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUzel A, et al. Artificial intelligence in orthodontic treatment planning: a scoping review. Eur J Orthod. 2022;44(3):273\u0026ndash;82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePham DD, et al. Deep learning for diagnosis of malocclusion from intraoral photographs. Angle Orthod. 2021;91(6):781\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim Y, et al. AI-assisted orthodontic treatment planning with CBCT data. Korean J Orthod. 2020;50(5):293\u0026ndash;302.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee K, et al. Convolutional neural networks for skeletal malocclusion classification using cephalograms. Diagnostics (Basel). 2020;10(11):930.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoi HI, et al. Predicting orthodontic extractions with machine learning. Sci Rep. 2021;11:22337.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, et al. AI prediction of orthodontic treatment outcomes using radiographs. Comput Methods Programs Biomed. 2021;200:105911.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMovahed A, et al. Deep learning-based prediction of orthodontic treatment duration. Am J Orthod Dentofac Orthop. 2022;161(4):476\u0026ndash;84.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrois J, et al. Machine learning in dental image analysis: a review. J Dent. 2021;103:103583.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYao J, et al. Applications of deep learning in orthodontics: current progress. Orthod Craniofac Res. 2022;25(1):34\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKang SH, et al. CNN classification of skeletal malocclusion from cephalograms. Korean J Orthod. 2021;51(2):123\u0026ndash;32.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee J, et al. Deep learning for automatic diagnosis of facial asymmetry. J Clin Med. 2021;10(7):1520.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTanikawa C, et al. Artificial intelligence in orthodontics: recent trends. Jpn Dent Sci Rev. 2021;57:193\u0026ndash;200.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSounderajah V, et al. Standards for reporting AI diagnostic accuracy studies (STARD-AI). BMJ Open. 2021;11:e047709.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu X, Cruz Rivera S, Moher D, et al. CONSORT-AI extension for clinical trials. BMJ. 2020;370:m3164.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKrois J, et al. Explainable AI in dentistry: a scoping review. J Dent. 2021;110:103664.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJheon AH, et al. Machine learning and orthodontics: a narrative review. Prog Orthod. 2021;22(1):18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSingh P, et al. Artificial intelligence and big data in orthodontics: challenges and opportunities. Semin Orthod. 2021;27(4):343\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbdi AH, et al. Deep learning in dental radiology: a scoping review. Dentomaxillofac Radiol. 2021;50(4):20200175.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu J, Cruz Rivera S, Moher D, et al. Multi-task deep learning for dental and skeletal classification. Med Image Anal. 2022;76:102313.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBMC Oral Health. About the journal. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://bmcoralhealth.biomedcentral.com/about\u003c/span\u003e\u003cspan address=\"https://bmcoralhealth.biomedcentral.com/about\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e [Accessed 2025-08-26].\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBMC Oral Health. Preparing your manuscript. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://bmcoralhealth.biomedcentral.com/submission-guidelines/preparing-your-manuscript\u003c/span\u003e\u003cspan address=\"https://bmcoralhealth.biomedcentral.com/submission-guidelines/preparing-your-manuscript\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e [Accessed 2025-08-26].\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":"artificial intelligence, machine learning, deep learning, orthodontics, diagnosis, treatment planning, scoping review","lastPublishedDoi":"10.21203/rs.3.rs-8641111/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8641111/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eArtificial intelligence (AI) applications in orthodontics are rapidly expanding across diagnosis, image analysis, and treatment planning.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eA PRISMA-ScR\u0026ndash;guided scoping review was conducted. PubMed/MEDLINE, Scopus, Web of Science, and Google Scholar were searched from 2010 to 16 September 2025. Original studies in orthodontics that used AI or machine learning for diagnosis, prediction, image analysis, or treatment planning were eligible. Two reviewers independently screened records, extracted data, and assessed risk of bias using QUADAS-2 for diagnostic accuracy studies and PROBAST for prediction model studies. Owing to heterogeneity in study design, datasets, and outcome metrics, results were synthesized narratively.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eOf 1,162 records identified, 1,008 remained after duplicate removal and were screened by title and abstract. A total of 154 full-text articles were assessed for eligibility, and 45 met the inclusion criteria. Frequent AI tasks included cephalometric landmark detection, malocclusion classification, extraction-decision support, treatment duration prediction, and cone-beam computed tomography (CBCT)-based segmentation. Many studies reported high accuracies for cephalometric landmark detection (mean radial error\u0026thinsp;\u0026lt;\u0026thinsp;2 mm and successful detection rates\u0026thinsp;\u0026gt;\u0026thinsp;80%) and malocclusion classification (accuracies\u0026thinsp;\u0026gt;\u0026thinsp;85%). However, risk-of-bias concerns, particularly in analysis and validation domains, were common, and external validation was infrequent.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eAI models show promising performance for orthodontic diagnosis and treatment planning and may enhance efficiency and standardization of care. Nevertheless, non-standardized outcome measures, limited external validation, and insufficient reporting of model development and evaluation currently restrict clinical translation. Larger, multicenter datasets, standardized benchmarks, and robust validation\u0026mdash;ideally following AI-specific reporting guidelines\u0026mdash;are required before routine clinical adoption.\u003c/p\u003e","manuscriptTitle":"Integration of artificial intelligence in orthodontic diagnosis and treatment planning a PRISMA ScR guided scoping review","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-16 10:32:15","doi":"10.21203/rs.3.rs-8641111/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":"7dcea1dc-ce91-422f-beb1-15dac1fd91da","owner":[],"postedDate":"February 16th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-14T06:55:46+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-16 10:32:15","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8641111","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8641111","identity":"rs-8641111","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}