Towards Applications of the “Surgical GPS” on Spinal Procedures

Abstract

Surgical data science (SDS) and computer-assisted surgery (CAS) are becoming increasingly conventional as more physicians incorporate surgical navigation systems into their workflows. Consequently, the necessity for enhanced anatomical structure depictions during intraoperative surgical guidance is also increasing. Anatomically accurate musculoskeletal representations are particularly necessary for orthopedic operations, especially spinal procedures. Additionally, CAS system recognition of clinically relevant structures along surgical corridors is of equal importance; such recognition permits intraoperative guidance when procedures are supervised by surgical process models (SPMs). Therefore, this study outlines a comprehensive solution, termed the “Surgical GPS”, which merges patient-specific, ligamentoskeletal models and physician-designated landmarks and ontologies, to expand spinal anatomy representations and demonstrate SPM applications toward several spine procedures. Patient models are validated with Dice similarity coefficient and Hausdorff distance metrics. Anatomical landmarks are queried from SPMs using pertinent search engines and are highlighted on the patient models, potentially enhancing successful surgical outcomes through context-aware, “GPS” guided CAS.