Emine Nur Kahraman
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Emine Nur Kahraman has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Busra Yilmaz
School of Dental Medicine, Istinye University
3 shared publications - Ayşe Gülşahı
Başkent
2 shared publications - Ipek Necla Guldiken
Department of Oral and Maxillofacial Surgery, Faculty of Dentistry, İstinye University, İstanbul, Türkiye
2 shared publications - Isabella Lima Arrais Ribeiro
Postgraduate Program in Dentistry, Federal University of Paraíba, Campus I, João Pessoa 58051-900, Brazil
2 shared publications - Roberto Cameriere
Department of Medicine and Health Sciences, University of Molise, 86100 Campobasso, Italy
2 shared publications - Şehrazat Evirgen
İzmir Tınaztepe University
2 shared publications
Latest Publications
- Open Access
- 11 September 2026
- Published by Anatolian Current Medical Journal in Anatolian Current Medical Journal
- Vol. 8 (5) , 1029-1039
Factors associated with maxillary sinus alveolar crest pneumatization: age, septa, nasal septum deviation, and root-sinus relationships
Aims: To identify factors independently associated with extensive maxillary sinus alveolar crest pneumatization on computed tomography (CT) in a cross-sectional sample.Methods: We retrospectively analyzed 800 maxillary quadrants from 400 patients. Pneumatization was classified as normal or extensive, and its associations with demographic, dental, and sinonasal variables were tested by multivariable logistic regression with cluster-robust standard errors. Intra- and inter-observer agreement was quantified on a random subsample of 82 quadrants. Generalized estimating equations and penalized-likelihood (Firth) regression were used as sensitivity analyses.Results: Extensive pneumatization was present in 603 quadrants (75.4%) and was independently associated with age [odds ratio (OR) 0.932 per year; 95% confidence interval (CI) 0.908-0.957; p
- 11 September 2026
- Published by Anatolian Current Medical Journal in Anatolian Current Medical Journal
- Vol. 8 (5) , 1029-1039
Open AccessFactors associated with maxillary sinus alveolar crest pneumatization: age, septa, nasal septum deviation, and root-sinus relationships
Aims: To identify factors independently associated with extensive maxillary sinus alveolar crest pneumatization on computed tomography (CT) in a cross-sectional sample.Methods: We retrospectively analyzed 800 maxillary quadrants from 400 patients. Pneumatization was classified as normal or extensive, and its associations with demographic, dental, and sinonasal variables were tested by multivariable logistic regression with cluster-robust standard errors. Intra- and inter-observer agreement was quantified on a random subsample of 82 quadrants. Generalized estimating equations and penalized-likelihood (Firth) regression were used as sensitivity analyses.Results: Extensive pneumatization was present in 603 quadrants (75.4%) and was independently associated with age [odds ratio (OR) 0.932 per year; 95% confidence interval (CI) 0.908-0.957; p
- 29 June 2026
- Published by MediHealth Academy in Journal of Radiology in Medicine
- Vol. 3 (2) , 33-41
Panoramic radiographic age estimation in Turkish children and adolescents: tooth-and sex-specific modelling using Cameriere’s open apex method
Aims: To evaluate panoramic radiographic measurements of individual mandibular teeth as predictors of chronological age in Turkish children and adolescents, using Cameriere's open apex method, and to assess whether sex-stratified modelling improves predictive accuracy. Methods: This retrospective cross-sectional study analysed 199 digital panoramic radiographs (PRs) from Turkish individuals aged 5-14 years (111 females, 88 males). Normalized open apex measurements were obtained for the seven left mandibular permanent teeth. Univariate screening followed by backward elimination was used to construct multiple linear regression models for the overall cohort and stratified by sex. Internal validation was performed using 10-fold cross-validation and bootstrapping with 1000 resamples. Results: In the overall model, teeth 34 and 37 were retained as significant predictors (R²=77.0%; standard error of estimate [SEE]=1.06 years; cross-validated R²=70,4%). The female-specific model retained teeth 34 and 35 (R²=78.9%; SEE=1.00 year; cross-validated R²=69.6%). The male-specific model retained teeth 33, 35 and 37 (R²=84.0%; SEE=0.92 years; cross-validated R²=77.8%). Age-stratified analyses revealed systematic overestimation in the 12-13-year age range across all models. Conclusion: Individual teeth differed substantially in their predictive contributions, and sex-stratified models showed higher cross-validated performance than the pooled model. The moderate sample size (n=88 for the male subgroup), single-centre design and absence of external validation or head-to-head comparison with the classical composite framework limit the conclusions that can be drawn. The findings are hypothesis-generating and require replication before clinical or forensic application.
- 29 June 2026
- Published by MediHealth Academy in Journal of Radiology in Medicine
- Vol. 3 (2) , 33-41
Panoramic radiographic age estimation in Turkish children and adolescents: tooth-and sex-specific modelling using Cameriere’s open apex method
Aims: To evaluate panoramic radiographic measurements of individual mandibular teeth as predictors of chronological age in Turkish children and adolescents, using Cameriere's open apex method, and to assess whether sex-stratified modelling improves predictive accuracy. Methods: This retrospective cross-sectional study analysed 199 digital panoramic radiographs (PRs) from Turkish individuals aged 5-14 years (111 females, 88 males). Normalized open apex measurements were obtained for the seven left mandibular permanent teeth. Univariate screening followed by backward elimination was used to construct multiple linear regression models for the overall cohort and stratified by sex. Internal validation was performed using 10-fold cross-validation and bootstrapping with 1000 resamples. Results: In the overall model, teeth 34 and 37 were retained as significant predictors (R²=77.0%; standard error of estimate [SEE]=1.06 years; cross-validated R²=70,4%). The female-specific model retained teeth 34 and 35 (R²=78.9%; SEE=1.00 year; cross-validated R²=69.6%). The male-specific model retained teeth 33, 35 and 37 (R²=84.0%; SEE=0.92 years; cross-validated R²=77.8%). Age-stratified analyses revealed systematic overestimation in the 12-13-year age range across all models. Conclusion: Individual teeth differed substantially in their predictive contributions, and sex-stratified models showed higher cross-validated performance than the pooled model. The moderate sample size (n=88 for the male subgroup), single-centre design and absence of external validation or head-to-head comparison with the classical composite framework limit the conclusions that can be drawn. The findings are hypothesis-generating and require replication before clinical or forensic application.
- 19 May 2026
- Published by Journal of Health Sciences and Medicine in Journal of Health Sciences and Medicine
- Vol. 9 (3) , 621-626
Sex-specific linear regression models for age estimation using Cameriere’s third Molar Maturity Index (I3M): a preliminary report from Turkiye
Aims: Cameriere’s Third Molar Maturity Index (I3M) is well established for age-threshold determination, yet its application in regression-based age estimation has not been explored in the Turkish population. This study aimed to derive sex-specific linear regression models for estimating chronological age from I3M in Turkish adolescents and young adults.Methods: This retrospective cross-sectional study evaluated 200 digital panoramic radiographs of individuals aged 14-22 years. I3M was quantified on left mandibular third molars using standardized ImageJ-based measurements by a single blinded examiner (ICC=0.90). Linear regression models were constructed for the pooled sample and separately by sex.Results: I3M showed a statistically significant inverse association with chronological age in all models (p
- 19 May 2026
- Published by Journal of Health Sciences and Medicine in Journal of Health Sciences and Medicine
- Vol. 9 (3) , 621-626
Sex-specific linear regression models for age estimation using Cameriere’s third Molar Maturity Index (I3M): a preliminary report from Turkiye
Aims: Cameriere’s Third Molar Maturity Index (I3M) is well established for age-threshold determination, yet its application in regression-based age estimation has not been explored in the Turkish population. This study aimed to derive sex-specific linear regression models for estimating chronological age from I3M in Turkish adolescents and young adults.Methods: This retrospective cross-sectional study evaluated 200 digital panoramic radiographs of individuals aged 14-22 years. I3M was quantified on left mandibular third molars using standardized ImageJ-based measurements by a single blinded examiner (ICC=0.90). Linear regression models were constructed for the pooled sample and separately by sex.Results: I3M showed a statistically significant inverse association with chronological age in all models (p
- 16 April 2026
- Published by AVES YAYINCILIK A.Ş. in Essentials of Dentistry
- Vol. 5, 1-8
Radiographic Prevalence and Association of Hypercementosis and Apical Root Resorption in Mandibular Molars Requiring Retreatment: A Cross-Sectional Study
Background: This study aimed to investigate the prevalence of apical root resorption and hypercementosis in mandibular first molars requiring endodontic retreatment due to persistent periapical lesions and to evaluate the potential association between these 2 conditions.Methods: A total of 487 digital panoramic radiographs were retrospectively analyzed, comprising 591 mandibular first molars with a periapical index score of ≥3. Apical root resorption was classified into 3 grades (0: none, 1: mild, 2: severe) based on the extent of root structure loss, and hypercementosis was categorized as diffuse, focal, or circular according to its morphological characteristics. Statistical analyses were conducted using chi-square and Fisher’s exact tests, with P < .05 considered statistically significant.Results: Apical root resorption was observed in 43.5% of the teeth, and hypercementosis in 41.1%. Both conditions coexisted in 14.7% of cases. A statistically significant negative association was found between apical resorption and hypercementosis (P=.012), indicating that hypercementosis tends to decrease as the severity of resorption increases. Diffuse hypercementosis was the most common type, followed by focal and circular forms. Hypercementosis was significantly more prevalent in patients aged 50 years and older (58.2%) compared to those under 50 years (33.9%) (P=.032).Conclusion: Apical root resorption and hypercementosis are frequently encountered in mandibular molars indicated for retreatment. Their co-occurrence may reflect a complex interplay between chronic inflammatory destruction and compensatory reparative responses. While hypercementosis might represent a protective adaptation, it can also pose clinical challenges during endodontic procedures and should be carefully evaluated during treatment planning.Cite this article as: Tosun S, Kahraman EN, Yılmaz B. Radiographic prevalence and association of hypercementosis and apical root resorption in mandibular molars requiring retreatment: a cross-sectional study. Essent Dent. 2026, 5, 0072, doi:10.5152/EssentDent.2026.25072.
- 16 April 2026
- Published by AVES YAYINCILIK A.Ş. in Essentials of Dentistry
- Vol. 5, 1-8
Radiographic Prevalence and Association of Hypercementosis and Apical Root Resorption in Mandibular Molars Requiring Retreatment: A Cross-Sectional Study
Background: This study aimed to investigate the prevalence of apical root resorption and hypercementosis in mandibular first molars requiring endodontic retreatment due to persistent periapical lesions and to evaluate the potential association between these 2 conditions.Methods: A total of 487 digital panoramic radiographs were retrospectively analyzed, comprising 591 mandibular first molars with a periapical index score of ≥3. Apical root resorption was classified into 3 grades (0: none, 1: mild, 2: severe) based on the extent of root structure loss, and hypercementosis was categorized as diffuse, focal, or circular according to its morphological characteristics. Statistical analyses were conducted using chi-square and Fisher’s exact tests, with P < .05 considered statistically significant.Results: Apical root resorption was observed in 43.5% of the teeth, and hypercementosis in 41.1%. Both conditions coexisted in 14.7% of cases. A statistically significant negative association was found between apical resorption and hypercementosis (P=.012), indicating that hypercementosis tends to decrease as the severity of resorption increases. Diffuse hypercementosis was the most common type, followed by focal and circular forms. Hypercementosis was significantly more prevalent in patients aged 50 years and older (58.2%) compared to those under 50 years (33.9%) (P=.032).Conclusion: Apical root resorption and hypercementosis are frequently encountered in mandibular molars indicated for retreatment. Their co-occurrence may reflect a complex interplay between chronic inflammatory destruction and compensatory reparative responses. While hypercementosis might represent a protective adaptation, it can also pose clinical challenges during endodontic procedures and should be carefully evaluated during treatment planning.Cite this article as: Tosun S, Kahraman EN, Yılmaz B. Radiographic prevalence and association of hypercementosis and apical root resorption in mandibular molars requiring retreatment: a cross-sectional study. Essent Dent. 2026, 5, 0072, doi:10.5152/EssentDent.2026.25072.
- Open Access
- 28 March 2026
- Published by Wiley in European Journal of Dental Education
- Vol. 30 (4) , 1951-1963
Comparative Assessment of Forensic Odontology Awareness Among Dental and Medical Students in Turkey
Introduction: This study aimed to assess and compare the awareness and knowledge of forensic odontology among dental and medical students in Turkey with a focus on identifying educational gaps and evaluating interdisciplinary readiness. Materials and Methods: A cross‐sectional web‐based survey was conducted with 234 students (144 dental and 90 medical students) in their clinical years. The questionnaire covered five domains: demographics, exposure to forensic education, conceptual familiarity, terminology recognition and professional responsibilities. Associations between survey responses and demographic variables were analysed using Pearson's chi‐square test and Cramér's V. Results: Dental students reported higher frequencies of forensic exposure, with 67.4% completing coursework and 86.8% attending workshops, compared to 28.9% and 50.0% of medical students, respectively. Conversely, medical students demonstrated higher recognition rates in specific domains, such as lip print terminology (24.4% vs. 4.9%) and the utility of teeth as a DNA source (33.3% vs. 12.5%). Senior students and female participants exhibited significantly higher indices of awareness and ethical sensitivity. Notably, a marked dissonance was observed between participants' high self‐rated sufficiency (> 80%) and their limited technical familiarity with specific forensic parameters. Conclusions: Despite having greater curricular exposure, dental students did not consistently demonstrate superior forensic knowledge compared to medical students, highlighting a disconnect between educational quantity and professional proficiency. Widespread deficits in critical medicolegal areas suggest that current instructional models are insufficient.
- 28 March 2026
- Published by Wiley in European Journal of Dental Education
- Vol. 30 (4) , 1951-1963
Open AccessComparative Assessment of Forensic Odontology Awareness Among Dental and Medical Students in Turkey
Introduction: This study aimed to assess and compare the awareness and knowledge of forensic odontology among dental and medical students in Turkey with a focus on identifying educational gaps and evaluating interdisciplinary readiness. Materials and Methods: A cross‐sectional web‐based survey was conducted with 234 students (144 dental and 90 medical students) in their clinical years. The questionnaire covered five domains: demographics, exposure to forensic education, conceptual familiarity, terminology recognition and professional responsibilities. Associations between survey responses and demographic variables were analysed using Pearson's chi‐square test and Cramér's V. Results: Dental students reported higher frequencies of forensic exposure, with 67.4% completing coursework and 86.8% attending workshops, compared to 28.9% and 50.0% of medical students, respectively. Conversely, medical students demonstrated higher recognition rates in specific domains, such as lip print terminology (24.4% vs. 4.9%) and the utility of teeth as a DNA source (33.3% vs. 12.5%). Senior students and female participants exhibited significantly higher indices of awareness and ethical sensitivity. Notably, a marked dissonance was observed between participants' high self‐rated sufficiency (> 80%) and their limited technical familiarity with specific forensic parameters. Conclusions: Despite having greater curricular exposure, dental students did not consistently demonstrate superior forensic knowledge compared to medical students, highlighting a disconnect between educational quantity and professional proficiency. Widespread deficits in critical medicolegal areas suggest that current instructional models are insufficient.
- 27 March 2026
- Published by Cumhuriyet University in Cumhuriyet Dental Journal
- Vol. 29 (1) , 158-167
Determinants of Early and Late Complications After Lateral Window Sinus Elevation: A Two-Year Retrospective Cohort Study
Objectives: To evaluate early postoperative complications and 1- and 2-year implant-related outcomes following lateral window maxillary sinus floor augmentation by correlating CBCT-based anatomical parameters with surgical and prosthetic variables.Materials and Methods: This retrospective single-center cohort study included 134 patients who underwent lateral window sinus augmentation between 2020 and 2023. Preoperative CBCT imaging was used to assess residual bone height, lateral wall thickness, palatal–nasal recess angle, ostium diameter, Schneiderian membrane thickness (2 mm). Early postoperative complications were recorded within the first month. Implant-related outcomes at 1 and 2 years were classified as implant failure or peri-implant crestal bone resorption (CBR). Statistical analyses included univariate tests and multivariate logistic regression (p2 mm, with bleeding/hematoma occurring exclusively in this group (p=0.006). Other anatomical parameters, including ostium diameter, showed no significant association with early complications. At 1 year, implant outcomes were significantly associated with prosthetic superstructure type, residual crestal bone height, intraoperative complications—particularly Schneiderian membrane perforation—and postoperative complication management strategy. These associations were primarily driven by increased CBR rather than implant failure. At 2 years, prosthetic superstructure type remained significantly associated with outcomes, while Schneiderian membrane thickness ≥3 mm was linked to higher CBR rates. Implant failure remained infrequent throughout the follow-up period.Conclusions: Lateral window sinus augmentation yielded predictable short- and mid-term outcomes when guided by CBCT-based presurgical assessment and structured postoperative management. Vascular anatomy, particularly AAA diameter, was the main determinant of early bleeding-related complications, whereas prosthetic design was the most consistent factor influencing long-term crestal bone changes.
- 27 March 2026
- Published by Cumhuriyet University in Cumhuriyet Dental Journal
- Vol. 29 (1) , 158-167
Determinants of Early and Late Complications After Lateral Window Sinus Elevation: A Two-Year Retrospective Cohort Study
Objectives: To evaluate early postoperative complications and 1- and 2-year implant-related outcomes following lateral window maxillary sinus floor augmentation by correlating CBCT-based anatomical parameters with surgical and prosthetic variables.Materials and Methods: This retrospective single-center cohort study included 134 patients who underwent lateral window sinus augmentation between 2020 and 2023. Preoperative CBCT imaging was used to assess residual bone height, lateral wall thickness, palatal–nasal recess angle, ostium diameter, Schneiderian membrane thickness (2 mm). Early postoperative complications were recorded within the first month. Implant-related outcomes at 1 and 2 years were classified as implant failure or peri-implant crestal bone resorption (CBR). Statistical analyses included univariate tests and multivariate logistic regression (p2 mm, with bleeding/hematoma occurring exclusively in this group (p=0.006). Other anatomical parameters, including ostium diameter, showed no significant association with early complications. At 1 year, implant outcomes were significantly associated with prosthetic superstructure type, residual crestal bone height, intraoperative complications—particularly Schneiderian membrane perforation—and postoperative complication management strategy. These associations were primarily driven by increased CBR rather than implant failure. At 2 years, prosthetic superstructure type remained significantly associated with outcomes, while Schneiderian membrane thickness ≥3 mm was linked to higher CBR rates. Implant failure remained infrequent throughout the follow-up period.Conclusions: Lateral window sinus augmentation yielded predictable short- and mid-term outcomes when guided by CBCT-based presurgical assessment and structured postoperative management. Vascular anatomy, particularly AAA diameter, was the main determinant of early bleeding-related complications, whereas prosthetic design was the most consistent factor influencing long-term crestal bone changes.
- 11 March 2026
- Published by Taylor & Francis in International Journal of Occupational Safety and Ergonomics
- p. 1-7
Radiation protection knowledge and training needs among dental healthcare workers in Turkey
Objectives. This study evaluated radiation protection knowledge among oral and dental healthcare professionals in Turkey, focusing on medical imaging technicians (MITs), dentists and dental assistants (DAs). Methods. A cross-sectional online survey was conducted with 236 participants using the validated Turkish version of the healthcare professional knowledge of radiation protection (Tr-HPKRP) scale. Knowledge was assessed across three domains: radiation physics and principles; radiation protection; and safe use of ionizing radiation. Mann-Whitney U tests and Kruskal-Wallis tests were applied for group comparisons. Results. MITs achieved significantly higher scores than dentists and DAs across all domains (p < 0.01). Structured education through academic programs or formal in-service training was positively associated with knowledge (p < 0.01), while reliance on informal sources such as media or colleagues showed no effect (p > 0.05). Professional experience was not correlated with knowledge (p > 0.05). Conclusions. Substantial knowledge gaps exist among dentists and DAs compared with MITs. Integrating standardized and regularly updated radiation safety training into undergraduate curricula and continuing professional development is essential to improve compliance with best practices and ensure patient and practitioner safety in dental diagnostic imaging.
- 11 March 2026
- Published by Taylor & Francis in International Journal of Occupational Safety and Ergonomics
- p. 1-7
Radiation protection knowledge and training needs among dental healthcare workers in Turkey
Objectives. This study evaluated radiation protection knowledge among oral and dental healthcare professionals in Turkey, focusing on medical imaging technicians (MITs), dentists and dental assistants (DAs). Methods. A cross-sectional online survey was conducted with 236 participants using the validated Turkish version of the healthcare professional knowledge of radiation protection (Tr-HPKRP) scale. Knowledge was assessed across three domains: radiation physics and principles; radiation protection; and safe use of ionizing radiation. Mann-Whitney U tests and Kruskal-Wallis tests were applied for group comparisons. Results. MITs achieved significantly higher scores than dentists and DAs across all domains (p < 0.01). Structured education through academic programs or formal in-service training was positively associated with knowledge (p < 0.01), while reliance on informal sources such as media or colleagues showed no effect (p > 0.05). Professional experience was not correlated with knowledge (p > 0.05). Conclusions. Substantial knowledge gaps exist among dentists and DAs compared with MITs. Integrating standardized and regularly updated radiation safety training into undergraduate curricula and continuing professional development is essential to improve compliance with best practices and ensure patient and practitioner safety in dental diagnostic imaging.
- Open Access
- 5 January 2026
- Published by Springer Nature in BMC Medical Imaging
- Vol. 26 (1)
Buccal or palatal? AI-based localization of impacted maxillary canines using panoramic radiographs
Third-dimensional binary localization, determining whether impacted maxillary canines (IMCs) requiring surgical exposure lie buccally or palatally, is amon
- 5 January 2026
- Published by Springer Nature in BMC Medical Imaging
- Vol. 26 (1)
Open AccessBuccal or palatal? AI-based localization of impacted maxillary canines using panoramic radiographs
Third-dimensional binary localization, determining whether impacted maxillary canines (IMCs) requiring surgical exposure lie buccally or palatally, is amon
- Open Access
- 29 October 2025
- Published by MDPI AG in Bioengineering
- Vol. 12 (11) , 1176
Prognostic Evaluation of Lower Third Molar Eruption Status from Panoramic Radiographs Using Artificial Intelligence-Supported Machine and Deep Learning Models
The prophylactic extraction of third molars is highly dependent on the surgeon’s experience as the common practices and guidelines contradict. The purpose of this study was to evaluate the eruption status of impacted third molars using deep learning-based artificial intelligence (AI) and to develop a model that predicts their final positions at an early stage to aid clinical decisions. In this retrospective study, 1102 panoramic radiographs (PANs) were annotated by three expert dentists to classify eruption status as either initial or definitive. A dataset was created and two deep learning architectures, InceptionV3 and ResNet50, were tested through a three-phase protocol: hyperparameter tuning, model evaluation, and assessment of preprocessing effects. Accuracy, recall, precision, and F1 score were used as performance metrics. Classical machine learning (ML) algorithms (SVM, KNN, and logistic regression) were also applied to features extracted from the deep models. ResNet50 with preprocessing achieved the best performance (F1 score: 0.829). Models performed better with definitive cases than with initial ones, where performance dropped (F1 score: 0.705). Clinically, the model predicted full eruption or impaction with 83% and 75% accuracy, respectively, but showed lower accuracy for partial impactions. These results suggest that AI can support early prediction of third molar eruption status and enhance clinical decision-making. Deep learning models (particularly ResNet50) demonstrated promising results in predicting third molar eruption outcomes. With larger datasets and improved optimization, AI tools may achieve greater accuracy and support routine clinical applications.
- 29 October 2025
- Published by MDPI AG in Bioengineering
- Vol. 12 (11) , 1176
Open AccessPrognostic Evaluation of Lower Third Molar Eruption Status from Panoramic Radiographs Using Artificial Intelligence-Supported Machine and Deep Learning Models
The prophylactic extraction of third molars is highly dependent on the surgeon’s experience as the common practices and guidelines contradict. The purpose of this study was to evaluate the eruption status of impacted third molars using deep learning-based artificial intelligence (AI) and to develop a model that predicts their final positions at an early stage to aid clinical decisions. In this retrospective study, 1102 panoramic radiographs (PANs) were annotated by three expert dentists to classify eruption status as either initial or definitive. A dataset was created and two deep learning architectures, InceptionV3 and ResNet50, were tested through a three-phase protocol: hyperparameter tuning, model evaluation, and assessment of preprocessing effects. Accuracy, recall, precision, and F1 score were used as performance metrics. Classical machine learning (ML) algorithms (SVM, KNN, and logistic regression) were also applied to features extracted from the deep models. ResNet50 with preprocessing achieved the best performance (F1 score: 0.829). Models performed better with definitive cases than with initial ones, where performance dropped (F1 score: 0.705). Clinically, the model predicted full eruption or impaction with 83% and 75% accuracy, respectively, but showed lower accuracy for partial impactions. These results suggest that AI can support early prediction of third molar eruption status and enhance clinical decision-making. Deep learning models (particularly ResNet50) demonstrated promising results in predicting third molar eruption outcomes. With larger datasets and improved optimization, AI tools may achieve greater accuracy and support routine clinical applications.
- 3 October 2025
- Published by Wiley in Oral Diseases
- Vol. 32 (2) , 422-430
Accuracy of ChatGPT‐4 Plus in Providing Information on Oral Cancer Management
Objective: Artificial intelligence (AI)‐driven large language models, such as Chat Generative Pre‐Trained Transformer (ChatGPT)‐4 Plus, are increasingly used for patient education and clinical decision support in oral oncology, although their accuracy in oral cancer (OC) management remains uncertain. This study evaluates the accuracy of ChatGPT‐4 Plus responses to clinically relevant questions regarding OC diagnosis, treatment, recovery, and prevention. Methods: A cross‐sectional study assessed 65 clinically relevant OC‐related questions using a paid ChatGPT‐4 Plus subscription without modifications. Three oral medicine specialists and one radiation oncologist rated accuracy on a four‐point scoring system. Interrater reliability was measured with the intraclass correlation coefficient (ICC), and chi‐square tests were used for comparisons. Results: Among 65 questions, 63% of responses were Score 1, with none rated as Score 4. Score 1 was most frequent in Recovery (72%), followed by Treatment (62%), Prevention (60%), and Diagnosis (55%). Scores 2 and 3 responses were highest in Diagnosis (45%). Recovery had significantly higher Score 1 responses than Diagnosis ( p < 0.05), while other comparisons were not significant. ICC ranged from 0.85 to 0.93. Conclusions: ChatGPT‐4 Plus provided accurate responses to clinically relevant OC‐related questions, particularly regarding recovery. However, diagnostic inconsistencies highlight the need for clinician oversight before integrating AI into practice.
- 3 October 2025
- Published by Wiley in Oral Diseases
- Vol. 32 (2) , 422-430
Accuracy of ChatGPT‐4 Plus in Providing Information on Oral Cancer Management
Objective: Artificial intelligence (AI)‐driven large language models, such as Chat Generative Pre‐Trained Transformer (ChatGPT)‐4 Plus, are increasingly used for patient education and clinical decision support in oral oncology, although their accuracy in oral cancer (OC) management remains uncertain. This study evaluates the accuracy of ChatGPT‐4 Plus responses to clinically relevant questions regarding OC diagnosis, treatment, recovery, and prevention. Methods: A cross‐sectional study assessed 65 clinically relevant OC‐related questions using a paid ChatGPT‐4 Plus subscription without modifications. Three oral medicine specialists and one radiation oncologist rated accuracy on a four‐point scoring system. Interrater reliability was measured with the intraclass correlation coefficient (ICC), and chi‐square tests were used for comparisons. Results: Among 65 questions, 63% of responses were Score 1, with none rated as Score 4. Score 1 was most frequent in Recovery (72%), followed by Treatment (62%), Prevention (60%), and Diagnosis (55%). Scores 2 and 3 responses were highest in Diagnosis (45%). Recovery had significantly higher Score 1 responses than Diagnosis ( p < 0.05), while other comparisons were not significant. ICC ranged from 0.85 to 0.93. Conclusions: ChatGPT‐4 Plus provided accurate responses to clinically relevant OC‐related questions, particularly regarding recovery. However, diagnostic inconsistencies highlight the need for clinician oversight before integrating AI into practice.