The use of dental implants has grown as the main treatment option for partially and completely edentulous patients because of their high long-term survival rates, good esthetic outcomes, and the opportunity to restore masticatory function.[1] Today, implant dentistry has grown beyond an exclusively restorative field to include advanced imaging, computer-assisted treatment planning, and biomaterial science that can predict the clinical outcome.[2] Although implant survival of over 90–95% has been reported over 10 years, peri-implantitis, lack of osseointegration, implant instability, and bone loss remain major clinical issues.[3] Such complications are frequently related to the misjudgment of the bone quality and quantity in the preoperative period, the positioning of the implant, and the selection of the wrong regenerative biomaterial.[4]

‘Cone Beam Computed Tomography’ (CBCT) is a revolutionary technology that produces 3D images of the maxillofacial structures with a relatively low radiation dose compared to ‘traditional computed tomography’ (CT) imaging.[5] CBCT, in contrast to traditional two-dimensional radiography, allows for precise judgment of alveolar bone height and width, cortical thickness, trabecular architecture, the proximity of the various vital anatomical structures, and the morphology of possible implant sites.[6] This thorough analysis is helpful for precise implant positioning and reduces the risk of complications like inferior alveolar nerve damage, maxillary sinus perforation, and implant malalignment.[7] This has made CBCT an essential tool in the digital implant workflow and is recommended by the main professional bodies for complex implant cases.[6]

An important factor in implant success is the choice of biomaterials used for bone augmentation as well as guided bone regeneration.[8] Various biomaterials, including autogenous bone grafts, allografts, xenografts, alloplastic substitutes, collagen membranes, platelet concentrates, and bioactive regenerative materials, are commonly used to augment inadequate alveolar bone volume.[9] The success of these materials is, however, largely dependent on their suitability to the anatomical features of the implant site, involving bone density, cortical thickness, defect morphology, and vascularization.[8] Selecting biomaterials is largely a subjective clinical decision based on clinician experience and interpretation of the radiographic findings, which may lead to discrepancies among clinicians.[10] A lack of matching of biomaterials with local anatomical conditions could affect bone regeneration, result in delayed osseointegration, and lead to implant failure.[11]

However, artificial intelligence with CBCT imaging can help address these limitations by providing precise anatomical mapping.[12] Image analysis will be able to automatically quantify bone density, classify bone defect morphology, assess cortical and trabecular bone features, estimate the available bone volume, and identify critical anatomical landmarks, all with high reproducibility, using AI-enhanced image analysis.[13] Moreover, the use of AI-powered predictive models has been demonstrated to predict implant stability, potential for osseointegration, and treatment outcomes based on radiographic, anatomical, and clinical factors.[13]

While there are numerous studies specifically evaluating the diagnostic performance of CBCT, the clinical role of artificial intelligence-based anatomical mapping, and the utility of different regenerative biomaterials, few studies have investigated the interplay of AI-assisted CBCT anatomical mapping for precision selection of dental implant biomaterials in dental implant therapy. Advanced image analytics used in biomaterial decision-making is an emerging area that could improve surgical planning, decrease complications, improve treatment variation, and increase implant survival. Furthermore, standardized procedures using AI could enhance clinical workflow efficiency and help less experienced providers in making informed treatment choices. The recent study aimed to assess the value of the use of AI-assisted CBCT anatomical mapping in the precision selection of biomaterials for dental implant therapy.

This analytical cross-sectional study was conducted in the Department of Prosthodontic, over a period of six months, from July, 2025 to December 2025. The sample size was calculated using the OpenEpi version 3.01 sample size calculator for estimation of a population proportion. In dental imaging, a previous study found that AI models using CBCT showed an overall accuracy of 92% in anatomical landmark detection.[14] Considering a confidence level of 95% and an absolute precision of 5%, a sample size of 114 implant sites was calculated. A consecutive non-probability sampling technique was employed. The study included patients aged 18 years or older who needed one or more implants and had a CBCT diagnostic examination before implant treatment planning. Additionally, patients had sufficient CBCT image data quality for AI analysis and full demographic and clinical data. The following factors excluded patients from the study: severe motion artifacts or distorted CBCT scans, placement of an implant at the implant site before the study, presence of active maxillofacial malignancies, cystic or fibro-osseous lesions of the implant area, congenital craniofacial anomalies, recent facial trauma, incomplete clinical records, and refusal to participate. Demographic data such as age and gender were obtained after IRB approval and after each participant provided written informed consent. All participants had a standardized CBCT imaging protocol, with the same exposure parameters. The ‘Digital Imaging and Communications in Medicine’ (DICOM) data were loaded into a validated AI-assisted dental imaging software system that enables automatic anatomical segmentation and quantitative analysis. Anatomical variables such as alveolar bone height, alveolar bone width, cortical bone thickness, trabecular bone density, and bone volume, proximity to vital anatomical structures, defect morphology, and available implant space were automatically assessed by the AI software. Each of the two highly experienced implant clinicians independently assessed the same CBCT scans without reference to the AI-generated results. The clinicians decided on the most appropriate biomaterial for each implant site, as indicated: autogenous bone graft, allograft, xenograft, alloplastic graft, collagen membrane, platelet-rich fibrin (PRF), or no augmentation as appropriate. The biomaterial recommendation generated by the AI was then compared with the biomaterial selected by the clinician. There was an evaluation of agreement between the assessment that was made with the aid of AI and the conventional assessment. Additional implant planning variables were recorded, such as implant diameter, implant length, requirement for sinus augmentation, guided bone regeneration, ridge augmentation, and the anticipated implant stability. All gathered data were recorded on a standardized data collection proforma, and two independent investigators checked the data to ensure there were no transcription errors. SPSS version 27.0 was used for analyzing the data. All continuous variables such as age, bone height, bone width, cortical thickness, trabecular bone density, and bone volume were presented as mean ± SD after the normality test with the Shapiro–Wilk test. These categorical variables were described using frequencies and percentages: Gender, jaw location, implant location, bone quality classification, biomaterial selection, augmentation procedure, and AI recommendations. The level of agreement between AI-assisted biomaterial selection and clinician-based assessment was calculated using Cohen's kappa coefficient with 95% confidence intervals. The Chi-square test and Fisher's exact test were used to analyze the differences in categorical variables. Continuous variables were compared between two groups using an independent-samples t-test. Possible confounding factors such as age, gender, implant site, bone density, cortical thickness, and defect morphology were included as independent variables in a multivariable binary logistic regression model to identify the independent predictors of the selection of biomaterials recommended by AI. The predictive ability of the AI model was assessed through receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) was determined. A p-value of < 0.05 (two-tailed) was regarded as statistically significant.

A total of 114 subjects (mean age 46.8 ± 11.9 years) participated in the study. The majority of the participants were aged between 36 and 50 years (42.1%) and more than 50 years (36.8%). The number of males in the study population was 57.9%, and that of females was 42.1%. The molar region was the most frequently used implant site (43.8%), followed by the maxilla (54.4%) and the mandible (45.6%). There were no statistically significant differences between the distributions of demographic characteristics (p>0.05). (Table 1)

The mean alveolar bone height is 12.6 ± 2.8 mm, alveolar bone width (7.2 ± 1.5 mm), cortical thickness (2.08 ± 0.47 mm), alveolar bone density (785±142 gray values), and alveolar bone volume (1.94 ± 0.61 cm³) were obtained using the AI-enhanced CBCT analysis. Of the implant sites, 71.1% had adequate implant space, and 40.4% had bone defects, the most common being horizontal defects. In 23.7% of cases, the proximity to important anatomical structures was noted. (Table 2)

There was good agreement between conventional clinician assessment and AI-assisted CBCT analysis in recommendations for biomaterials. Both strategies most commonly recommended biomaterials were xenografts (31.6% and 33.3%, respectively), followed by alloplastic grafts and autogenous bone grafts. There was no statistically significant difference between the biomaterial used with the two methods (p=0.841), suggesting similar clinical decision-making. Moreover, the overall level of agreement between the recommendations of the AI and the clinicians was 90.4%, with a good Cohen's kappa coefficient of 0.86 (p<0.001). (Tables 3 and 4)

AI-suggested biomaterial augmentation sites exhibited significantly lower bone height, bone width, cortical thickness, bone density, and bone volume than sites that did not need biomaterial augmentation (all p<0.001). (Table 5)

Significant associations were observed between AI-recommended biomaterial selection and poor bone density, reduced cortical thickness, presence of alveolar bone defects, limited implant space, and proximity to vital anatomical structures (all p<0.01). (Table 6)

Bone defects (p<0.001), limited implant space (p=0.009), lower bone density (p<0.001), and thinner cortical bone (p=0.002) were all independent predictors of biomaterial selection recommended by AI, while age, gender, and implant location were not. The regression model was reasonably well calibrated with a non-significant Hosmer-Lemeshow test (0.682) and accounted for 58% of the variation in biomaterial selection. The AI prediction model also showed excellent diagnostic performance, an AUC of 0.93, sensitivity of 91.8%, specificity of 87.8%, and an overall diagnostic accuracy of 90.4% (p<0.001). (Tables 7 and 8)

Table 1. Demographic Characteristics of the Study Participants (n = 114)

Table 2. AI-Based CBCT Anatomical Mapping Findings

Table 3. Biomaterial Selection by Conventional Assessment and AI-Based Recommendation

Table 4. Agreement Between AI-Assisted and Conventional Biomaterial Selection

Table 5. Comparison of Anatomical Parameters According to AI Recommendation for Bone Augmentation

Table 6. Association of Anatomical Variables with AI-Recommended Biomaterial Selection

Table 7. Multivariable Binary Logistic Regression for Predictors of AI-Recommended Biomaterial Selection

Table 8. Diagnostic Performance of the AI Model for Precision Biomaterial Selection

The present study showed that the anatomical mapping done using AI-assisted CBCT images was in excellent agreement with the conventional clinician-based evaluation for the precision biomaterial selection, with an overall agreement rate of 90.4% and a Cohen's kappa coefficient of 0.86. The AI model also showed excellent diagnostic capabilities (AUC=0.93) and was successful in accurately determining the location of implants that need the augmentation of bone volume using objective anatomical features. The results underscore the increasing importance of artificial intelligence (AI) as an accurate decision-making tool in digital implant dentistry and indicate that using AI for CBCT analysis could enhance the accuracy and uniformity of biomaterial selection.

These results are similar to those of Bayrakdar et al. (2021), who assessed a deep learning system for implant planning in a CBCT database, obtaining an excellent agreement between the measurements obtained by the AI system and those obtained by the operator for the majority of implant regions for both bone height and bone thickness. They found that AI could reliably detect anatomical landmarks and considerably alleviate the clinical workload, in line with the high agreement recorded in the current study.[15]

The high diagnostic accuracy achieved in our AI model is also corroborated by Sarwar and Jabin (2023), who reviewed the application of AI in CBCT imaging and found that deep learning algorithms significantly enhance automated segmentation, anatomical landmark identification, bone assessment, and diagnostic consistency over traditional manual interpretation. Their findings are echoed in our review, in which they highlight that AI improves reproducibility and reduces the variation between observers.[16]

The main factors that affected the AI-based suggestions for biomaterial selection in this study were the decrease in bone density, a decrease in cortical thickness, lack of space for implants, and the presence of alveolar bone defects. In a related study, Roongruangsilp et al. (2025) compared the ‘Faster R-CNN and YOLOv7’ models for implant planning and found that the AI model could analyze CBCT-derived anatomical structures with high accuracy in various implant-planning software platforms. They found that objective radiographic measurements were helpful to enhance the consistency of treatment planning, which aligns with our results on the use of AI for biomaterial selection.[17]

These results align with the conclusions of the BMC Oral Health (2025) scoping review, which compiled evidence from implant planning and prognostic studies and found that using AI to combine CBCT data with digital workflows enhances implant site prediction, prosthetically driven implant planning and implant outcome prediction. The review highlighted that AI-assisted planning reduces variability among clinicians and facilitates personalized treatment planning, which is in agreement with the present study.[18]

The results of the recent systematic review conducted by Zaman et al. (2026) also support our findings, showing that AI-supported implant planning had diagnostic accuracy ranging from 92% to 99.7% and improved planning efficiency in some studies, although some also reported decreased implant success rates when compared to conventional planning. The authors concluded that AI-based CBCT analysis could offer significant value in optimizing implant treatment planning while noting the importance of larger prospective clinical trials, which resonates with our observations.[19]

The level of agreement between AI recommendations and clinician decisions seen in our study is similar to the registry-based cohort study by Dawood et al. (2026), where they found that AI use for implant planning in the esthetic maxilla compared favorably to experienced clinicians in terms of planning accuracy, required less time for planning, and had a more consistent workflow. The researchers' results further highlight the promise of AI to bring uniformity to implant planning in various clinical environments.[20]

Likewise, Elgarba, B. et al. (2026) found that the clinical validation study showed a high mean positional accuracy of 0.5 mm, a high mean implant dimension selection accuracy (0.2 mm), and high mean planning consistency (0.2 mm) between the AI-generated virtual implant placement and the expert-planned and clinically placed implants. These findings support the excellent agreement and predictive performance demonstrated by our AI model.[21]

The present study was important in determining that the bone defect, bone thickness, implant space, and bone density were independent factors that predicted the biomaterial selection recommended by the AI. This is biologically feasible because these anatomical features directly affect the primary stability of the implant and the requirement for regenerative procedures. These findings have also been documented in recent publications of studies on AI-assisted implant planning, where the quantitative measurements obtained from CBCT have been shown to deliver more objective and personalized treatment planning than the visual assessment.

The present study specifically investigated the use of AI-assisted anatomical mapping for precision biomaterial selection, which differed from previous research that was mainly about the position of the implants or automated landmark detection. This is a significant step further towards personalized implant dentistry, as it combines quantitative CBCT analysis with regeneration planning. The use of objective AI-derived anatomical parameters in biomaterial selection can lead to greater predictability in surgical outcomes, decreased inter-operator variability, and enhanced long-term outcomes of implant surgery. However, prospective, multicenter clinical trials with long-term follow-up will be necessary before biomaterial selection for implants can be routinely integrated into evidence-based implant practice with the support of AI.

There were some limitations in this study. First, it was tested at a single tertiary health care center with a relatively small number of subjects, which might cause the results to not be generalizable. Second, biomaterial recommendations were not based on long-term clinical outcomes like implant survival or osseointegration but on AI-assisted CBCT analysis and the clinician's evaluation. Third, only one AI software platform was analyzed, so results might not be directly applicable to other commercially available AI software platforms. Moreover, even though treatment planning was standardized, there could have been some variation due to the experience of the clinicians and the parameters used for acquiring the CBCT images. Further multicenter prospective studies with larger populations, a variety of AI platforms, and longer follow-up are warranted to confirm the clinical effectiveness of AI-guided biomaterial selection.

The accuracy of precision biomaterial selection in dental implant planning was excellent between AI-enhanced CBCT-based anatomical mapping and the conventional clinician assessment. An accurate identification of relevant anatomical features, such as bone density, cortical thickness, defects, and the available implant space, was a critical factor in choosing the appropriate biomaterial and was well captured by the AI model. AI's impressive diagnostic accuracy and predictive capability suggest it holds promise as an effective decision-support tool for customized implant treatment planning. The combination of AI and CBCT imaging technologies could lead to more consistent planning, efficient regenerative strategies, and more predictable dental implant therapy.

1. Sartoretto, S.C., et al., Comparing the long-term success rates of tooth preservation and dental implants: a critical review. Journal of Functional Biomaterials, 2023. 14(3): p. 142.
2. Neji, G., et al., AI-Powered Predictive Models in Implant Dentistry: Planning, Risk Assessment, and Outcomes. Journal of Clinical Medicine, 2025. 15(1): p. 228.
3. Roccuzzo, A., et al., Peri-implantitis: a clinical update on prevalence and surgical treatment outcomes. Journal of clinical medicine, 2021. 10(5): p. 1107.
4. Ramesh, V., et al., Hidden pitfalls: Understanding errors and complications in periodontal and implant therapy. World Academy of Sciences Journal, 2025. 7(6): p. 121.
5. Jain, S., et al., New evolution of cone-beam computed tomography in dentistry: Combining digital technologies. Imaging science in dentistry, 2019. 49(3): p. 179.
6. Sathish, S., Advances in CBCT for Dental Implantology: Trends in Assessment, Planning, and Monitoring, in Advanced Imaging and Diagnostic Trends in Oral and Maxillofacial Radiology. 2026, Springer. p. 717-759.
7. Chen, S.T., et al., Complications and treatment errors in implant positioning in the aesthetic zone: Diagnosis and possible solutions. Periodontology 2000, 2023. 92(1): p. 220-234.
8. Donos, N., et al., Bone regeneration in implant dentistry: Which are the factors affecting the clinical outcome? Periodontology 2000, 2023. 93(1): p. 26-55.
9. Ferraz, M.P., Bone grafts in dental medicine: an overview of autografts, allografts and synthetic materials. Materials, 2023. 16(11): p. 4117.
10. Murdoch, A.I.K., et al., Determinants of clinical decision making under uncertainty in dentistry: a scoping review. Diagnostics, 2023. 13(6): p. 1076.
11. Laubach, M., et al., The concept of scaffold-guided bone regeneration for the treatment of long bone defects: current clinical application and future perspective. Journal of functional biomaterials, 2023. 14(7): p. 341.
12. Neylon, J., et al., Proof‐of‐concept study of artificial intelligence‐assisted review of CBCT image guidance. Journal of Applied Clinical Medical Physics, 2023. 24(9): p. e14016.
13. Poorvitha, H. and B. Chandrakala. A Review of Osteoporosis Detection Based on Image Texture of Inner Trabecular Bone Microarchitecture Using Different Imaging Modalities. in International Conference on Next-Generation Communication and Computing. 2024. Springer.
14. Wu, Y., et al., Automatic identification of anatomical landmarks in three-dimensional computed tomography/cone-beam computed tomography: a scoping review. Frontiers in Dental Medicine, 2026. 7: p. 1847046.
15. Kurt Bayrakdar, S., et al., A deep learning approach for dental implant planning in cone-beam computed tomography images. BMC medical imaging, 2021. 21(1): p. 86.
16. Sarwar, S. and S. Jabin. AI techniques for cone beam computed tomography in dentistry: trends and practices. in 2023 International Conference on Recent Advances in Electrical, Electronics & Digital Healthcare Technologies (REEDCON). 2023. IEEE.
17. Roongruangsilp, P., W. Narkbuakaew, and P. Khongkhunthian, Performance of two different artificial intelligence models in dental implant planning among four different implant planning software: a comparative study. BMC Oral Health, 2025. 25(1): p. 984.
18. Khaohoen, A., et al., Can artificial intelligence optimize treatment planning and outcome prediction in fixed tooth-and implant-supported prosthodontics? A scoping review. BMC Oral Health, 2025.
19. Zaman, G., et al., The Use of Artificial Intelligence in Planning Dental Implant Procedures: A Systematic Review. Dentistry Journal, 2026. 14(5): p. 248.
20. Dawood, E.A., et al., Clinical Applicability of Artificial Intelligence–Driven Implant Planning and Surgical Guide Design in the Maxillary Esthetic Zone: A Registry‐Based Cohort Study. Clinical Oral Implants Research, 2026.
21. Elgarba, B.M., et al., Automating virtual dental implant planning: can artificial intelligence match clinical expertise? Dentomaxillofacial Radiology, 2026. 55(5): p. 421-430.