LLM Performance in Endodontic Diagnostics

Evaluating ChatGPT-4o, Gemini and Claude 3.7 in Endodontic Diagnostics: A Prospective Clinical Study

The goal of this prospective observational study is to evaluate the ability of three large language models (ChatGPT-4o, Gemini Advanced, and Claude 3.7) to support diagnosis and treatment decision-making in adult patients presenting with common endodontic conditions.

The main questions the study aims to answer are:

Can LLMs accurately determine the endodontic diagnosis when provided with structured clinical information and periapical radiographs?

Can LLMs propose appropriate treatment plans comparable to decisions made by endodontic specialists?

To answer these questions, researchers will compare the diagnostic and treatment accuracy of three AI models using a consensus diagnosis from endodontic specialists as the reference standard.

Participants will:

Receive routine endodontic examination and periapical radiographs as part of standard clinical care.

Have their anonymized clinical histories and radiographs entered into the three AI models.

Not interact directly with any AI system; all evaluations will be performed by the research team.

This study aims to understand how large language models perform under real-world clinical conditions and whether these systems may play a supportive role in endodontic diagnostics in the future.

Study Overview

• Status: Completed
• Conditions: Endodontic Diagnosis, Endodontic Diseases, Endodontic Treatment, Endodontic Decision-making
• Intervention / Treatment: Diagnostic test: AI-Based Diagnostic Assessment

Detailed Description

This prospective observational study aims to evaluate the real-time diagnostic and treatment decision-making performance of three large language models-ChatGPT-4o, Gemini Advanced, and Claude 3.7-in an endodontic clinical setting. A total of 120 patients presenting to the endodontic clinic were examined, and detailed medical/dental histories, clinical findings, and periapical radiographs were collected. Each anonymized case was then presented to the three LLMs using a standardized prompt asking for the diagnosis and the appropriate treatment plan.

All models were used in their default multimodal configurations without enabling web-search functions, plug-ins, or external data retrieval. Each question was submitted only once in isolated chat sessions to prevent memory carry-over. Responses were saved verbatim and compared with the reference diagnoses and treatment plans established by a panel of endodontic specialists.

This study was designed to mimic real-world clinical conditions as closely as possible, providing a realistic assessment of how these systems might perform when used by clinicians in everyday practice. Understanding their capabilities and limitations in authentic clinical scenarios is essential, as LLMs are expected to play an increasingly vital role in future dental care particularly in decision support, triage, and patient education. By identifying where these models perform well and where they fall short, this research aims to inform safe and effective clinical integration as LLM technologies continue to advance.

• Study Type: Observational
• Enrollment (Actual): 120

Contacts and Locations

Study Locations

• Turkey (Türkiye)
  • Istanbul
    • Maltepe, Istanbul, Turkey (Türkiye), 34856
      • Faculty of Dentistry, Marmara University

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

The study population consisted of adult patients attending or referred to the endodontic clinic of Marmara University. All participants presented with common endodontic conditions such as pulpitis, necrosis, primary or secondary apical periodontitis, or the need for retreatment. After obtaining consent, each patient underwent a structured paper-based medical and dental history assessment and periapical radiographic examination. A total of 120 clinically verified endodontic cases were included.

Description

Inclusion Criteria:

• Adult patients (≥18 years old) presenting to or referred to the Endodontic Clinic.

Patients with a clinically verified endodontic condition requiring diagnosis and treatment planning.

Patients who agreed to participate and provided informed consent.

Patients for whom a complete paper-based medical/dental history and periapical radiograph were obtained during the clinical visit.

Exclusion Criteria:

• Exclusion Criteria

Patients who declined participation or did not provide informed consent.

Pediatric patients (<18 years old) referred to the Pediatric Dentistry Clinic.

Patients attending the clinic with non-endodontic complaints (e.g., post-extraction alveolitis, third-molar extraction problems).

Cases with incomplete clinical information or missing radiographs.

Patients unable to undergo standard endodontic examination procedures.

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort

Intervention / Treatment

Endodontic Patients Cohort; This cohort includes 120 consecutive patients presenting to the endodontic clinic with clinically verified endodontic conditions. Clinical history and periapical radiographs were collected, and diagnostic/treatment recommendations generated by AI models were compared with expert consensus.

Diagnostic test: AI-Based Diagnostic Assessment; Participants' anonymized clinical information, including structured patient history and periapical radiographs, was used as input for three large language models (ChatGPT-4o, Gemini Advanced, Claude 3.7). The models were asked to determine the endodontic diagnosis and propose an appropriate treatment plan. No treatment, device, or drug was administered to participants. The intervention consists solely of AI-based interpretation of pre-existing clinical data.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Clinician Diagnosis Accuracy Based on Paper-Based History and Periapical Radiograph	Assessment of the diagnostic decision made by endodontic clinicians after reviewing a paper-based patient history form and a standardized periapical radiograph. Accuracy is determined by comparing the clinician's diagnosis with the consensus diagnosis established by three independent endodontic specialists. Data will be collected for all 120 patients at the time of initial clinical evaluation.	7 july-5 august

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
LLM-Generated Diagnosis and Treatment Planning Performance	Evaluation of diagnostic and treatment recommendations generated by large language models (LLMs)-ChatGPT-4o, Gemini Advanced, and Claude 3.7-after receiving the same paper-based patient history and periapical radiograph provided to clinicians. LLM responses will be compared to the gold-standard specialist consensus for both diagnosis and treatment decisions.	august-september

Collaborators and Investigators

• Sponsor: Marmara University
• Investigators: Study Director: ayşe karadayı, asst. prof., Marmara University Faculty of Dentistry

Publications and helpful links

General Publications

• Abd-Alrazaq A, AlSaad R, Alhuwail D, Ahmed A, Healy PM, Latifi S, Aziz S, Damseh R, Alabed Alrazak S, Sheikh J. Large Language Models in Medical Education: Opportunities, Challenges, and Future Directions. JMIR Med Educ. 2023 Jun 1;9:e48291. doi: 10.2196/48291.
• Schwendicke F, Samek W, Krois J. Artificial Intelligence in Dentistry: Chances and Challenges. J Dent Res. 2020 Jul;99(7):769-774. doi: 10.1177/0022034520915714. Epub 2020 Apr 21.

Study record dates

Study Major Dates

• Study Start (Actual): July 7, 2025
• Primary Completion (Actual): August 5, 2025
• Study Completion (Actual): October 3, 2025

Study Registration Dates

• First Submitted: November 24, 2025
• First Submitted That Met QC Criteria: December 8, 2025
• First Posted (Actual): December 15, 2025

Study Record Updates

• Last Update Posted (Actual): December 15, 2025
• Last Update Submitted That Met QC Criteria: December 8, 2025
• Last Verified: December 1, 2025

More Information

Terms related to this study

• Keywords: access cavity cleaning, air abrasion, air polishing, CLSM, ethanol, push-out bond strength
• Other Study ID Numbers: 2025-38

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: UNDECIDED

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No