Artificial neural networks development in prosthodontics - a systematic mapping review

Area of Science:

• Dental research
• Artificial intelligence in dentistry
• Prosthodontics

Background:

• Artificial neural networks (ANNs) show promise in prosthodontics.
• A systematic overview of ANN applications in this field is needed.
• Understanding current trends and limitations is crucial for advancement.

Purpose of the Study:

• To systematically categorize ANN prediction models in prosthodontics.
• To provide a comprehensive overview of ANN aims, data, architectures, and evaluation metrics.
• To identify limitations in ANN development within prosthodontics.

Main Methods:

• Systematic literature review of articles published up to June 2024.
• Searches included terms like "prosthodontics," "neural networks," and "deep learning."
• Analysis of 70 selected articles focusing on ANN models in prosthodontics (2007-2024).

Main Results:

• ANN models are applied in both implant (29%) and non-implant (71%) prosthodontics.
• Diverse data types, architectures, and evaluation metrics are utilized.
• A significant portion of reviewed articles (33%) were published in 2023, indicating recent growth.

Conclusions:

• ANN models offer diverse applications in prosthodontics, with varying methodologies.
• Limitations exist, particularly concerning the data lifecycle in ANN development.
• Findings guide clinical integration, inform decision-making, and support AI advancement in dentistry.