Development of a Support Vector Machine-based Automated Facial Le Fort I Osteotomy Robot

The development of automatically-controlled surgical robots (ACSRs) is hindered by limitations, such as their rigidity, which presents a significant challenge in safeguarding delicate structures like blood vessels and nerves. We aimed to overcome these limitations and pioneer the successful practical surgical application of an ACSR. While performing Le Fort I osteotomy in maxillofacial surgery, the ACSR should avoid damaging the descending palatine artery, as such damage can cause fatal bleeding and necrosis of the osteotomized maxilla. The bone structure in front of the greater palatine canal through which the descending palatine artery passes in the human maxilla is the anterior cortical bone–cancellous bone–posterior cortical bone. The proposed algorithm uses a support vector machine to sequentially determine the bone type using the osteotomy reaction force obtained by the reaction force estimation observer, and automatically stop the osteotomy just before penetrating the posterior cortical bone before reaching the descending palatine artery. Commercially available porcine ribs, which resemble the bone structure in front of the greater palatine canal, were used as the experimental material. Autostop and bone type determination were 100% successful without obtaining numerical information of the model in advance.