Motamedian Saeed Reza, Mohaghegh Sadra, Niazmand Maral, Mohammad-Rahimi Hossein, Ahmadi Nima, Yaseri Mina, Boroojeni Helia Sadat Haeri
Department of Orthodontics, School of Dentistry, Shahid Beheshti University of Medical Sciences, Tehran, Iran.
Department of Oral and Maxillofacial Surgery, School of Dentistry, University of Nevada, Las Vegas, Nevada, USA.
Biomed Res Int. 2025 Jun 12;2025:8284581. doi: 10.1155/bmri/8284581. eCollection 2025.
This study was aimed at reviewing the application of different artificial intelligent algorithms used in different phases of orthognathic surgeries, which include diagnosis, treatment planning, soft tissue prediction, outcome evaluation, and complication assessment. This aimed to update clinicians on this technology to integrate it into their decision-making, in addition to being aware of its challenges and potential areas for further assessment. Electronic search was done in PubMed, Scopus, Embase, and Cochrane databases. Studies that reported the application of artificial intelligence (AI) in different aspects of orthognathic surgery were included. From 656 studies, a total of 29 articles met the inclusion criteria and were used to categorize the application of AI as follows: (1) Diagnosis in which studies showed the sensitivity of 75%-95.5% for specifying the need for orthognathic surgery; (2) treatment planning in which AI was used for osteotomy design and bony reference point determination with 3.99-4.73 mm of error; (3) soft tissue prediction in which AI models showed a success rate of 64.3%-100%; (4) outcome evaluation in which AI was used to assess the impact of surgery on asymmetry, facial attractiveness, and esthetic improvements quantitatively; and (5) complication assessment with an accuracy of 98.7% for predicting postsurgical systemic infection and 7.4 mL of error for blood loss. AI can be potentially considered as a proper alternative to conventional approaches to fasten the procedure related to orthognathic surgery and with comparable accuracy to conventional methods.
本研究旨在回顾不同人工智能算法在正颌外科手术不同阶段的应用，这些阶段包括诊断、治疗计划、软组织预测、结果评估和并发症评估。这旨在让临床医生了解这项技术，以便将其纳入决策过程，同时了解其面临的挑战和有待进一步评估的潜在领域。在PubMed、Scopus、Embase和Cochrane数据库中进行了电子检索。纳入了报告人工智能（AI）在正颌外科手术不同方面应用的研究。从656项研究中，共有29篇文章符合纳入标准，并用于将AI的应用分类如下：（1）诊断，其中研究显示确定正颌手术需求的敏感性为75%-95.5%；（2）治疗计划，其中AI用于截骨设计和骨参考点确定，误差为3.99-4.73毫米；（3）软组织预测，其中AI模型的成功率为64.3%-100%；（4）结果评估，其中AI用于定量评估手术对不对称、面部吸引力和美学改善的影响；（5）并发症评估，预测术后全身感染的准确率为98.7%，失血误差为7.4毫升。人工智能有可能被视为传统方法的合适替代方案，以加快与正颌外科手术相关的程序，并且与传统方法具有相当的准确性。
