Merve Koseoglu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Merve Koseoglu has a total of 72 co-authors. Here are the top co-authors by number of shared publications:
- Cortino Sukotjo
Department of Prosthodontics, School of Dental Medicine University of Pittsburgh Pittsburgh Pennsylvania USA
7 shared publications Funda Bayindir
Professor, Department of Prosthodontics, Faculty of Dentistry, University of Atatürk, Erzurum, Turkey
4 shared publicationsJudy Chia-Chun Yuan
Professor and Associate Dean for Clinical Affairs, Department of Restorative Dentistry, College of Dentistry, University of Illinois Chicago, Chicago, Ill
3 shared publications- Selma Altindi̇ş
Sakarya Üniversitesi
3 shared publications - Hakan Akin
Department of Prosthodontics, Faculty of Dentistry, Sakarya University, Sakarya, Turkey
2 shared publications - Hatice Özdemir
Professor, Atatürk University, Department of Prosthodontics, Faculty of Dentistry, Erzurum, Turkey
2 shared publications
Latest Publications
- Open Access
- 29 June 2026
- Published by Türkiye Çocuk Cerrahisi Dernegi in Turkish Journal of Pediatric Surgery
- Vol. 40 (2) , 53-64
Differentiating acute appendicitis in children in the emergency department: A decision tree-based model
Objectives: This study aims to develop a clinically applicable and interpretable decision tree-based model using routine clinical and laboratory parameters to differentiate acute appendicitis (AA) from nonspecific abdominal pain (NAP) in pediatric patients. Patients and methods: This retrospective study included 708 pediatric patients presenting with abdominal pain between January 2019 and December 2023. Demographic characteristics, physical examination findings, laboratory parameters, and Alvarado scores (AS) were analyzed. The dataset was divided into training and held-out internal validation sets using stratified sampling to preserve class distribution. Model performance was evaluated using stratified five-fold cross-validation on the training set, while the held-out internal validation set was kept completely unseen during model development. Class imbalance was addressed using the Synthetic Minority Over-sampling Technique applied exclusively within training folds. The dataset was partitioned using stratified sampling, and a decision tree model based on the classification and regression tree algorithm with an entropy splitting criterion was constructed. Model performance was evaluated on the heldout internal validation set using area under the curve (AUC), precision-recall (PR)-AUC, and F1-score. Results: A total of 708 patients categorized into two groups: NAP group (296 males, 307 females; mean age: 10.20 ± 3.81 years, range, 4 to 17 years) and AA group (68 males, 37 females; mean age: 11.21 ± 3.51 years, range, 5 to 17 years) presenting with abdominal pain were recruited. Patients with AA had significantly higher white blood cell counts, neutrophil percentages, neutrophil-tolymphocyte ratios, C-reactive protein levels, and AS than those in the NAP group (all p < 0.001). Right lower quadrant (RLQ) tenderness emerged as the most important discriminative variable, followed by the AS, left shift, age, and anorexia. The decision tree model achieved an F1-score of 0.884, an AUC of 0.946, and a PR-AUC of 0.8545. Importantly, AA was associated with a very low likelihood in patients without RLQ tenderness and with low AS, defining a clinically meaningful low-risk subgroup. Conclusion: Decision tree-based models may serve as valuable adjuncts to clinical judgment in evaluating suspected pediatric AA, particularly by identifying low-risk patients for whom further diagnostic testing or imaging is unlikely to alter clinical management.
- 29 June 2026
- Published by Türkiye Çocuk Cerrahisi Dernegi in Turkish Journal of Pediatric Surgery
- Vol. 40 (2) , 53-64
Open AccessDifferentiating acute appendicitis in children in the emergency department: A decision tree-based model
Objectives: This study aims to develop a clinically applicable and interpretable decision tree-based model using routine clinical and laboratory parameters to differentiate acute appendicitis (AA) from nonspecific abdominal pain (NAP) in pediatric patients. Patients and methods: This retrospective study included 708 pediatric patients presenting with abdominal pain between January 2019 and December 2023. Demographic characteristics, physical examination findings, laboratory parameters, and Alvarado scores (AS) were analyzed. The dataset was divided into training and held-out internal validation sets using stratified sampling to preserve class distribution. Model performance was evaluated using stratified five-fold cross-validation on the training set, while the held-out internal validation set was kept completely unseen during model development. Class imbalance was addressed using the Synthetic Minority Over-sampling Technique applied exclusively within training folds. The dataset was partitioned using stratified sampling, and a decision tree model based on the classification and regression tree algorithm with an entropy splitting criterion was constructed. Model performance was evaluated on the heldout internal validation set using area under the curve (AUC), precision-recall (PR)-AUC, and F1-score. Results: A total of 708 patients categorized into two groups: NAP group (296 males, 307 females; mean age: 10.20 ± 3.81 years, range, 4 to 17 years) and AA group (68 males, 37 females; mean age: 11.21 ± 3.51 years, range, 5 to 17 years) presenting with abdominal pain were recruited. Patients with AA had significantly higher white blood cell counts, neutrophil percentages, neutrophil-tolymphocyte ratios, C-reactive protein levels, and AS than those in the NAP group (all p < 0.001). Right lower quadrant (RLQ) tenderness emerged as the most important discriminative variable, followed by the AS, left shift, age, and anorexia. The decision tree model achieved an F1-score of 0.884, an AUC of 0.946, and a PR-AUC of 0.8545. Importantly, AA was associated with a very low likelihood in patients without RLQ tenderness and with low AS, defining a clinically meaningful low-risk subgroup. Conclusion: Decision tree-based models may serve as valuable adjuncts to clinical judgment in evaluating suspected pediatric AA, particularly by identifying low-risk patients for whom further diagnostic testing or imaging is unlikely to alter clinical management.
- Open Access
- 14 April 2026
- Published by Wiley in Journal of Prosthodontics
Assessing the ability of large language models to summarize and generate maxillofacial prosthetic treatment options
Purpose: The aim was to evaluate the ability of four large language models (LLMs) (OpenAI's ChatGPT‐3.5, Microsoft 365 Copilot, DeepSeek‐R1, and Google Gemini 2.5 Pro) to develop treatment options when presented with clinical cases published in the maxillofacial prosthodontics literature. Materials and Methods: Six maxillofacial case reports were fed to the LLMs following a prompt that requested prosthodontic treatment options from the perspective of a prosthodontist. Expert evaluators scored the relevance, clarity, depth, focus, and coherence of the responses. Statistical analyses, including descriptive statistics, two‐way analysis of variance (ANOVA), post hoc Tukey tests, Pearson correlation analyses, and intraclass correlation coefficients (ICCs), were performed ( α < 0.05). Results: There were significant differences among the total mean relevance ( p = 0.003), clarity ( p = 0.006), depth ( p < 0.001), focus ( p < 0.001), and coherence ( p < 0.001) scores of chatbots. Copilot consistently scored the lowest, and Gemini or DeepSeek scored the highest for all five factors. Depth ( p = 0.006), focus ( p = 0.024), and coherence ( p = 0.013) scores of senior prosthodontists were slightly higher than those of junior prosthodontists. Pearson correlation analysis revealed positive correlations between the total mean scores for all five factors ( p < 0.001). Conclusions: The study demonstrates the ability of LLMs to develop maxillofacial prosthetic treatment plans tailored to specific clinical scenarios. There were significant differences between the abilities of the LLMs evaluated in this study. Copilot scored the lowest for all factors evaluated, and Gemini and/or DeepSeek scored the highest.
- 14 April 2026
- Published by Wiley in Journal of Prosthodontics
Open AccessAssessing the ability of large language models to summarize and generate maxillofacial prosthetic treatment options
Purpose: The aim was to evaluate the ability of four large language models (LLMs) (OpenAI's ChatGPT‐3.5, Microsoft 365 Copilot, DeepSeek‐R1, and Google Gemini 2.5 Pro) to develop treatment options when presented with clinical cases published in the maxillofacial prosthodontics literature. Materials and Methods: Six maxillofacial case reports were fed to the LLMs following a prompt that requested prosthodontic treatment options from the perspective of a prosthodontist. Expert evaluators scored the relevance, clarity, depth, focus, and coherence of the responses. Statistical analyses, including descriptive statistics, two‐way analysis of variance (ANOVA), post hoc Tukey tests, Pearson correlation analyses, and intraclass correlation coefficients (ICCs), were performed ( α < 0.05). Results: There were significant differences among the total mean relevance ( p = 0.003), clarity ( p = 0.006), depth ( p < 0.001), focus ( p < 0.001), and coherence ( p < 0.001) scores of chatbots. Copilot consistently scored the lowest, and Gemini or DeepSeek scored the highest for all five factors. Depth ( p = 0.006), focus ( p = 0.024), and coherence ( p = 0.013) scores of senior prosthodontists were slightly higher than those of junior prosthodontists. Pearson correlation analysis revealed positive correlations between the total mean scores for all five factors ( p < 0.001). Conclusions: The study demonstrates the ability of LLMs to develop maxillofacial prosthetic treatment plans tailored to specific clinical scenarios. There were significant differences between the abilities of the LLMs evaluated in this study. Copilot scored the lowest for all factors evaluated, and Gemini and/or DeepSeek scored the highest.
- Open Access
- 1 January 2026
- Published by Wiley in International Journal of Dentistry
- Vol. 2026 (1) , 5468404
Machine Learning Methods to Predict the Academic Performance of Pre‐Clinical Dental Students Based on Pre‐University Information
Background: Information regarding the prediction of dentistry students’ academic performances using machine learning (ML) techniques is lacking. The objective of this study was to predict the academic performance of pre‐clinical dentistry students using ML techniques based on their pre‐university information. Methods: A total of 96 data points, including graduation years, genders, high school grade point average (GPA), step 1 university exam score (Step 1 UES), Step 2 UES, total UES, university exam ratings (UER), high school types (HSTypes), hometowns, parent educations (PEs), and parent incomes (PIs) from dentistry grads, were collected. Various ML regression techniques were validated to predict the academic performance of dentistry students in this cross‐sectional study. Results: Preliminary findings from the dataset indicated that UESs or rankings were more predictive than other demographic features, such as HSType, gender, graduation year, and hometown. The selection of target variables has resulted in two major studies. The initial study employed overall preclinical scores as the target variable, whereas the subsequent study utilized letter grades from specific preclinical courses, including dental morphology manipulation, dental materials I, prosthodontics I, and prosthodontics II, for ML models. The performance measures of these models demonstrated the capability of ML models to evaluate student performance in relation to demographics, high school performance, and admission scores. Regression models exhibiting an R 2 value of 0.74 or greater, alongside a classification accuracy beyond 70% for a balanced dataset of preclinical subject scores, demonstrated that ML models can predict student achievement in both overall and subject‐specific grades. Conclusion: The present study revealed the utilization of ML techniques to predict the academic performance of dentistry students successfully.
- 1 January 2026
- Published by Wiley in International Journal of Dentistry
- Vol. 2026 (1) , 5468404
Open AccessMachine Learning Methods to Predict the Academic Performance of Pre‐Clinical Dental Students Based on Pre‐University Information
Background: Information regarding the prediction of dentistry students’ academic performances using machine learning (ML) techniques is lacking. The objective of this study was to predict the academic performance of pre‐clinical dentistry students using ML techniques based on their pre‐university information. Methods: A total of 96 data points, including graduation years, genders, high school grade point average (GPA), step 1 university exam score (Step 1 UES), Step 2 UES, total UES, university exam ratings (UER), high school types (HSTypes), hometowns, parent educations (PEs), and parent incomes (PIs) from dentistry grads, were collected. Various ML regression techniques were validated to predict the academic performance of dentistry students in this cross‐sectional study. Results: Preliminary findings from the dataset indicated that UESs or rankings were more predictive than other demographic features, such as HSType, gender, graduation year, and hometown. The selection of target variables has resulted in two major studies. The initial study employed overall preclinical scores as the target variable, whereas the subsequent study utilized letter grades from specific preclinical courses, including dental morphology manipulation, dental materials I, prosthodontics I, and prosthodontics II, for ML models. The performance measures of these models demonstrated the capability of ML models to evaluate student performance in relation to demographics, high school performance, and admission scores. Regression models exhibiting an R 2 value of 0.74 or greater, alongside a classification accuracy beyond 70% for a balanced dataset of preclinical subject scores, demonstrated that ML models can predict student achievement in both overall and subject‐specific grades. Conclusion: The present study revealed the utilization of ML techniques to predict the academic performance of dentistry students successfully.
- 31 December 2025
- Published by Elsevier in The Journal of Prosthetic Dentistry
- Vol. 135 (6) , 1147-1148
Response to Letter to the Editor regarding, “Exploring the role of Instagram Reels in predoctoral implant education, a randomized crossover study”
- 31 December 2025
- Published by Elsevier in The Journal of Prosthetic Dentistry
- Vol. 135 (6) , 1147-1148
Response to Letter to the Editor regarding, “Exploring the role of Instagram Reels in predoctoral implant education, a randomized crossover study”
- Open Access
- 29 December 2025
- Published by Selcuk Dental Journal in Selcuk Dental Journal
- Vol. 12 (3) , 499-505
The Clinical Use of Virtual Articulators in the Digital Era
Virtual articulators simulate inter-jaw relationships in a computerized environment. Virtual and augmented reality are used as digital tools to support dental clinical practice and education. Many authors have published studies investigating the digital techniques available for the use of a virtual articulator. The aim of the present study was to analyze the existing literature on virtual articulators. Within the scope of the study, 41 articles published in different databases were included. In current technology, there are 2 different types of virtual articulators: fully adjustable and mathematically simulated. The workflow of virtual articulators consists of 4 main steps: digitizing the dental arch by scanning it with an intraoral scanner or 3D laser scanner, occlusal registration, virtual facebow, and virtual mounting. The advantages of virtual articulators are that they provide three-dimensional visualization, improve communication between dentist, dental technician, and patient, increase efficiency, facilitate patient education, integrate with computer-aided design and manufacturing systems, evaluate both static and dynamic occlusion, and analyze occlusal interferences. The disadvantages of virtual articulators are that mounting the casts to the articulator and scanning these takes time, increases the cost as it requires special software, scanners, and technical knowledge and skills. In conclusion, it is understood that virtual reality has changed the practice of dentistry and will continue to improve dental procedures in the near future. Virtual articulators simulate occlusion in a virtual environment with the potential to be translated into clinical practice. In the future, these systems need to be augmented by 4-dimensional technology in dynamic simulations.
- 29 December 2025
- Published by Selcuk Dental Journal in Selcuk Dental Journal
- Vol. 12 (3) , 499-505
Open AccessThe Clinical Use of Virtual Articulators in the Digital Era
Virtual articulators simulate inter-jaw relationships in a computerized environment. Virtual and augmented reality are used as digital tools to support dental clinical practice and education. Many authors have published studies investigating the digital techniques available for the use of a virtual articulator. The aim of the present study was to analyze the existing literature on virtual articulators. Within the scope of the study, 41 articles published in different databases were included. In current technology, there are 2 different types of virtual articulators: fully adjustable and mathematically simulated. The workflow of virtual articulators consists of 4 main steps: digitizing the dental arch by scanning it with an intraoral scanner or 3D laser scanner, occlusal registration, virtual facebow, and virtual mounting. The advantages of virtual articulators are that they provide three-dimensional visualization, improve communication between dentist, dental technician, and patient, increase efficiency, facilitate patient education, integrate with computer-aided design and manufacturing systems, evaluate both static and dynamic occlusion, and analyze occlusal interferences. The disadvantages of virtual articulators are that mounting the casts to the articulator and scanning these takes time, increases the cost as it requires special software, scanners, and technical knowledge and skills. In conclusion, it is understood that virtual reality has changed the practice of dentistry and will continue to improve dental procedures in the near future. Virtual articulators simulate occlusion in a virtual environment with the potential to be translated into clinical practice. In the future, these systems need to be augmented by 4-dimensional technology in dynamic simulations.
- Open Access
- 22 August 2025
- Published by Wiley in Journal of Prosthodontics
Assessing readability and accuracy of content produced by the American College of Prosthodontists and large language models for patient education in prosthodontics
Purpose: This study aims to evaluate the readability and accuracy of content produced by ChatGPT, Copilot, Gemini, and the American College of Prosthodontists (ACP) for patient education in prosthodontics.Materials and Methods: A series of 26 questions were selected from the ACP's list of questions (GoToAPro.org FAQs) and their published answers. Answers to the same questions were generated from ChatGPT‐3.5, Copilot, and Gemini. The word counts of responses from chatbots and the ACP were recorded. The readability was calculated using the Flesch Reading Ease Scale and Flesch–Kincaid Grade Level. The responses were also evaluated for accuracy, completeness, and overall quality. Descriptive statistics were used to calculate mean and standard deviations (SD). One‐way analysis of variance was performed, followed by the Tukey multiple comparisons to test differences across chatbots, ACP, and various selected topics. The Pearson correlation coefficient was used to examine the relationship between each variable. Significance was set at α < 0.05.Results: ChatGPT had a higher word count, while ACP had a lower word count (p < 0.001). The cumulative scores of the prosthodontist topic had the lowest Flesch Reading Ease Scale score, while brushing and flossing topics displayed the highest score (p < 0.001). Brushing and flossing topics also had the lowest Flesch–Kincaid Grade Level score, whereas the prosthodontist topic had the highest score (p < 0.001). Accuracy for denture topics was the lowest across the chatbots and ACP, and it was the highest for brushing and flossing topics (p = 0.006).Conclusions: This study highlights the potential for large language models to enhance patient's prosthodontic education. However, the variability in readability and accuracy across platforms underscores the need for dental professionals to critically evaluate the content generated by these tools before recommending them to patients.
- 22 August 2025
- Published by Wiley in Journal of Prosthodontics
Open AccessAssessing readability and accuracy of content produced by the American College of Prosthodontists and large language models for patient education in prosthodontics
Purpose: This study aims to evaluate the readability and accuracy of content produced by ChatGPT, Copilot, Gemini, and the American College of Prosthodontists (ACP) for patient education in prosthodontics.Materials and Methods: A series of 26 questions were selected from the ACP's list of questions (GoToAPro.org FAQs) and their published answers. Answers to the same questions were generated from ChatGPT‐3.5, Copilot, and Gemini. The word counts of responses from chatbots and the ACP were recorded. The readability was calculated using the Flesch Reading Ease Scale and Flesch–Kincaid Grade Level. The responses were also evaluated for accuracy, completeness, and overall quality. Descriptive statistics were used to calculate mean and standard deviations (SD). One‐way analysis of variance was performed, followed by the Tukey multiple comparisons to test differences across chatbots, ACP, and various selected topics. The Pearson correlation coefficient was used to examine the relationship between each variable. Significance was set at α < 0.05.Results: ChatGPT had a higher word count, while ACP had a lower word count (p < 0.001). The cumulative scores of the prosthodontist topic had the lowest Flesch Reading Ease Scale score, while brushing and flossing topics displayed the highest score (p < 0.001). Brushing and flossing topics also had the lowest Flesch–Kincaid Grade Level score, whereas the prosthodontist topic had the highest score (p < 0.001). Accuracy for denture topics was the lowest across the chatbots and ACP, and it was the highest for brushing and flossing topics (p = 0.006).Conclusions: This study highlights the potential for large language models to enhance patient's prosthodontic education. However, the variability in readability and accuracy across platforms underscores the need for dental professionals to critically evaluate the content generated by these tools before recommending them to patients.
- Open Access
- 21 April 2025
- Published by Selcuk Dental Journal in Selcuk Dental Journal
- Vol. 12 (1) , 42-47
Complication Rates and Patient Satisfaction Evaluation of Implant-Supported Prostheses
Background: The aim of the study was to assess implant and prosthetic failure rates and patients’ evaluations following 2-year observation. Methods: The study was conducted among 110 patients (60 male, 50 female). During recall visits during the first and second year of prosthesis insertion, researchers conducted the clinical examination. A questionnaire was used to evaluate the patient’s satisfaction with the prostheses'. The obtained data were analyzed by descriptive statistics, Chi-Square, Mann-Witney, and Kruskal-Wallis analysis (α
- 21 April 2025
- Published by Selcuk Dental Journal in Selcuk Dental Journal
- Vol. 12 (1) , 42-47
Open AccessComplication Rates and Patient Satisfaction Evaluation of Implant-Supported Prostheses
Background: The aim of the study was to assess implant and prosthetic failure rates and patients’ evaluations following 2-year observation. Methods: The study was conducted among 110 patients (60 male, 50 female). During recall visits during the first and second year of prosthesis insertion, researchers conducted the clinical examination. A questionnaire was used to evaluate the patient’s satisfaction with the prostheses'. The obtained data were analyzed by descriptive statistics, Chi-Square, Mann-Witney, and Kruskal-Wallis analysis (α
- Open Access
- 4 January 2025
- Published by Elsevier in The Journal of Prosthetic Dentistry
- Vol. 133 (6) , 1604-1613
Exploring the potential of machine learning models to predict nasal measurements through facial landmarks
- 4 January 2025
- Published by Elsevier in The Journal of Prosthetic Dentistry
- Vol. 133 (6) , 1604-1613
Open AccessExploring the potential of machine learning models to predict nasal measurements through facial landmarks
- 12 February 2024
- Published by Wiley in Journal of Dental Education
- Vol. 88 (6) , 786-797
Assessing methodological quality in dental education research using MERSQI: Analysis of publications from two journals
Purpose: The Medical Education Research Study Quality Instrument (MERSQI) has been used frequently to assess the methodological quality of medical education but not for dental education. The present study aimed to assess the methodological quality using MERSQI scores of articles published in the Journal of Dental Education (JDE) and the European Journal of Dental Education (EJDE).Methods: A cross‐sectional assessment of the quality of manuscripts published in 2012, 2017, and 2022 JDE and EJDE was conducted. MERSQI data, numbers of authors, first and corresponding author degrees, geographic origins, and funding information were also extracted for each included study. Descriptive and analytical statistics were conducted, and significance level was set at α < 0.05.Results: Four hundred ninety‐five articles met the inclusion criteria. The most common study design was a single‐group cross‐sectional or single‐group posttest and conducted in one institution for all studied years. In all journals and years, studies were assessed mainly by participants. The study outcome was mostly satisfaction, attitudes, perceptions, opinions, and general facts. The total mean MERSQI score for each journal and year varied. Year and geographic origin significantly affected the total MERSQI score. Papers originating from Asia had the highest score, followed by South America, Europe, North America, Oceania, and Africa.Conclusion: MERSQI score is applicable to the assessment of the methodological quality of dental educational research. The MERSQI score for most of the domains was similar for both journals. The MERSQI score was affected by publication years and geographic origins.
- 12 February 2024
- Published by Wiley in Journal of Dental Education
- Vol. 88 (6) , 786-797
Assessing methodological quality in dental education research using MERSQI: Analysis of publications from two journals
Purpose: The Medical Education Research Study Quality Instrument (MERSQI) has been used frequently to assess the methodological quality of medical education but not for dental education. The present study aimed to assess the methodological quality using MERSQI scores of articles published in the Journal of Dental Education (JDE) and the European Journal of Dental Education (EJDE).Methods: A cross‐sectional assessment of the quality of manuscripts published in 2012, 2017, and 2022 JDE and EJDE was conducted. MERSQI data, numbers of authors, first and corresponding author degrees, geographic origins, and funding information were also extracted for each included study. Descriptive and analytical statistics were conducted, and significance level was set at α < 0.05.Results: Four hundred ninety‐five articles met the inclusion criteria. The most common study design was a single‐group cross‐sectional or single‐group posttest and conducted in one institution for all studied years. In all journals and years, studies were assessed mainly by participants. The study outcome was mostly satisfaction, attitudes, perceptions, opinions, and general facts. The total mean MERSQI score for each journal and year varied. Year and geographic origin significantly affected the total MERSQI score. Papers originating from Asia had the highest score, followed by South America, Europe, North America, Oceania, and Africa.Conclusion: MERSQI score is applicable to the assessment of the methodological quality of dental educational research. The MERSQI score for most of the domains was similar for both journals. The MERSQI score was affected by publication years and geographic origins.