Towards a radiation-free clinical decision support system for intraoperative spinal alignment assessment.

Abstract

This paper introduces SpineAlign, a novel radiation-free clinical decision support system (CDSS) designed to address the challenge of intraoperative spinal alignment assessment during spinal deformity (SD) correction surgeries. SpineAlign aims to overcome the current limitations of existing systems by providing a quantitative assessment without radiation exposure in the operating room (OR), thus enhancing the safety and precision of computer-assisted spinal surgeries (CASS). The system focuses on spinal alignment calculation, leveraging Bézier curves and algorithm development to track vertebrae and estimate spinal curvature. Collaborative meetings with clinical experts identified challenges such as patient positioning complexities and limitations of minimal invasiveness. Thus, the method developed involves four algorithms: (1) tracking anatomical planes; (2) estimating the Bézier curve; (3) determining vertebrae positions; and (4) adjusting orientation. A proof of concept (PoC) using a porcine spinal segment validated SpineAlign's integrated algorithms and functionalities. The PoC demonstrated the system's accuracy and clinical applicability, successfully transitioning a spine without curvature to a lordotic spine. Quantitative evaluation of spinal alignment by the system showed high accuracy, with a maximum root mean squared error of 6 $_{}^{\circ }$ . The successful PoC marks an initial step towards developing a reliable CDSS for intraoperative spinal alignment assessment without medical image acquisition. Future steps will focus on enhancing system robustness and performing multi-surgeon serial studies to advance SpineAlign towards widespread clinical adoption.