Augmented reality-guided precision in mandibular reconstruction: Non-contact laser point cloud registration for simultaneous osteotomy and implantation

Abstract

Maxillofacial reconstruction often results in significant mandibular defects, necessitating precise surgical interventions. Traditional approaches suffer from lengthy preoperative preparations and substantial radiation exposure. This study introduces a novel augmented reality (AR) framework for simultaneous fibula flap mandibular reconstruction and implant placement, leveraging non-contact laser point cloud registration. Our method employs adaptive line laser scanning (ALLS) for surface reconstruction, enabling markerless registration to integrate preoperative virtual surgical plans with intraoperative reality. Virtual-real view calibration, enhanced by particle swarm optimization (PSO), ensures high alignment accuracy. Cadaver experiments demonstrated superior performance, with mandibular osteotomy achieving a 2.38 mm control point position error and fibular osteotomy an accuracy of 1.29 mm. AR-guided simultaneous implantation showed a position error of 1.28 mm, and overall mandibular reconstruction errors were approximately 1.77 mm, ensuring stable temporomandibular joint functionality. This framework offers a new technological paradigm for minimally invasive and intelligent precision maxillofacial reconstruction.