Clinical Case: Maxillofacial Surgery

Surgeons are under an increasing workload due to the fast-growing patient population of oral and maxillofacial surgery (OMS). Previously proposed methods have a limited benefit to surgeons’ manual work because of the significant individual diversity in OMS patients. Artificial intelligence (AI) offers a promising tool to mimic surgeons’ decision-making mechanisms to reduce the workload. In this chapter, our recent achievements in OMS by using a machine learning-based approach to assist surgical planning were introduced. We collected both preoperative and 1-year-later postoperative CT images of 56 patients to train a 12-layer cascaded deep neural network structure with two successive models. A virtual surgery planning approach was also presented to illustrate how the model-predicted results were further used to assist the surgeon’s planning work during Lefort I treatment. Current limitations and future trends of AI in OMS were also briefly discussed. AI demonstrates its feasibility in assisting the human surgeon and shows great potential for reducing the workload. Preparation of annotated medical data, high computing power hardware, efficient feature-extracting network, and ethical issues will be key challenges for a wide application of AI in OMS.