Evaluation of Surgical Difficulty for Impacted Mandibular Third Molars Using a Machine Learning Model on Panoramic Radiographs

Status

Active, not recruiting

Phases

Unknown

Study type

Observational

Source

TCTR

Registry ID

TCTR20250922004

Enrollment

1600

Registered

2025-09-22

Start date

2025-09-30

Completion date

Unknown

Last updated

2026-08-03

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

To review and evaluate machine learning techniques applicable to dental radiograph analysis, and to develop a machine learning model for assessing the difficulty of impacted tooth surgery using panoramic radiographs. Dentomaxillofacial Radiology, Impacted Lower Third molar, Machine Learning, Deep Learning, State-of-the-art (SOTA)

Interventions

This results in a total of 1200 panoramic radiographs

Device Feasibility

Subjects

Eligibility

Sex/Gender

All

Inclusion criteria

Inclusion criteria: 1. Images will be acquired following standard institutional protocols, encompassing a wide range of dental pathologies and anomalies, including impacted teeth. Factors affecting image visibility will be either included or controlled. For example, radiographs from different manufacturers, models, and machines, with varying patient positioning, exposure settings, and image enhancement filters, will be collected and standardized. This variability is intended to improve the model's robustness in handling differences in imaging equipment commonly encountered in clinical practice. 2. Clear, well-illuminated panoramic photographs of the entire oral cavity will be obtained.

Exclusion criteria

Exclusion criteria: 1. Radiographs of non-diagnostic quality, including overexposed, underexposed, distorted, or blurred images. 2. Images in which conditions obscure the region of interest, such as pre-existing pathologies or a history of surgery or other physical treatments in the area. 3. As AI models can be developed as sub-models and subsequently integrated to achieve all research objectives, radiographs used in this study may be collected from different patients. This independent data collection approach offers the advantage of minimizing unnecessary radiation exposure for patients without a clinical indication for radiographic imaging.

Design outcomes

Primary

Measure	Time frame
Radiographs N/A N/A	—

Secondary

Measure	Time frame
N/A N/A N/A	—

Countries

Thailand

Contacts

Public ContactThantrira Porntaveetus

Faculty of Dentistry, Chulalongkorn University

thantrira.p@chula.ac.th0819999939

Outcome results

None listed

Source: TCTR (via WHO ICTRP) · Data processed: Aug 10, 2026