# Application of Artificial Intelligence Models in Oral and Maxillofacial Prosthesis Restoration: A Systematic Review

## TL;DR

This review explores how artificial intelligence is being used to improve maxillofacial prosthetic restoration, highlighting its potential and areas needing further research.

## Contribution

The paper systematically reviews AI applications in image acquisition, data analysis, and coloring for maxillofacial prosthetics, identifying trends and gaps.

## Key findings

- AI applications in maxillofacial prosthetics are increasing, especially in image acquisition, data analysis, and pseudo-coloring.
- Current studies show high accuracy but lack focus on patient and doctor comfort and efficiency.
- More research is needed on personalized model design and accurate analysis of complex structures.

## Abstract

Artificial intelligence (AI) is increasingly favoured in the medical field, and the traditional maxillofacial prosthetic system has complex processes, low efficiency, and relatively insufficient model matching and aesthetic outcomes. With the rapid development of AI, it is expected that AI technology will address the shortcomings of these traditional methods.

To comprehensively analyse the research on AI in maxillofacial prosthetic image acquisition, data analysis, and colouring, and to elaborate on its application evaluation and future trends to improve the diagnosis and treatment experience of patients.

Based on the PICO principle, we searched the PubMed, Web of Science, and Cochrane databases, as well as conducted manual searches, and screened English articles from January 2016 to January 2024 according to specific inclusion and exclusion criteria. Quality assessment was performed by 2 investigators using the JBI Quasi-Experimental Study Critical Assessment Checklist.

A total of 17 related articles were identified, including 7 in the field of image acquisition, 8 in the field of data analysis, and 2 in the field of pseudo colouring. The results show that the application of AI in these 3 fields is on the rise, reaching a record high in 2024.

There is a lack of literature available for analysis, all studies have demonstrated high accuracy and clinical necessity. However, there is a relative lack of consideration when it comes to the comfort and efficiency of patients and doctors in practice, which is a key direction for future AI software development. In addition, we need more research on irregular bone or personalized model design, as well as more accurate analysis of complex structures that are vulnerable to damage.