X23D—Intraoperative 3D Lumbar Spine Shape Reconstruction Based on Sparse Multi-View X-ray Data

Abstract

Visual assessment based on intraoperative 2D X-rays remains the predominant aid for intraoperative decision-making, surgical guidance, and error prevention. However, correctly assessing the 3D shape of complex anatomies, such as the spine, based on planar fluoroscopic images remains a challenge even for experienced surgeons. This work proposes a novel deep learning-based method to intraoperatively estimate the 3D shape of patients’ lumbar vertebrae directly from sparse, multi-view X-ray data. High-quality and accurate 3D reconstructions were achieved with a learned multi-view stereo machine approach capable of incorporating the X-ray calibration parameters in the neural network. This strategy allowed a priori knowledge of the spinal shape to be acquired while preserving patient specificity and achieving a higher accuracy compared to the state of the art. Our method was trained and evaluated on 17,420 fluoroscopy images that were digitally reconstructed from the public CTSpine1K dataset. As evaluated by unseen data, we achieved an 88% average F1 score and a 71% surface score. Furthermore, by utilizing the calibration parameters of the input X-rays, our method outperformed a counterpart method in the state of the art by 22% in terms of surface score. This increase in accuracy opens new possibilities for surgical navigation and intraoperative decision-making solely based on intraoperative data, especially in surgical applications where the acquisition of 3D image data is not part of the standard clinical workflow.

Current Trends in Artificial Intelligence-Assisted Spine Surgery: A Systematic Review

“…Spinal surgeries pose significant challenges because of their technical intricacy and proximity to critical organs such as the spinal cord, nerves, and aorta [9,39,40]. Previous studies have identified various surgeon-and patientspecific factors as potential causes of postoperative complications in these interventions, ranging from pedicle screw malplacement to cage malpositioning [41][42][43][44].…”

“…Previous studies have identified various surgeon-and patientspecific factors as potential causes of postoperative complications in these interventions, ranging from pedicle screw malplacement to cage malpositioning [41][42][43][44]. To address these challenges, CAS navigation techniques have been introduced [9]. These techniques facilitate and standardize spinal procedures by offering three-dimensional (3D) intraoperative spatial guidance.…”

“…The need for a registration process between preoperative and intraoperative data remains a significant bottleneck in the widespread adoption of CAS systems [9]. To overcome these drawbacks, recent research has explored the transformation of 3D images of the spine based on Xray images captured during surgery, rather than relying solely on preoperative images [9]. Jecklin et al [9] introduced an innovative DL approach designed to estimate the The CNN algorithm tested in this study provides an accurate means to automatically identify the vertebral body anatomy and provide measurements for implants and placement trajectories.…”

Translation of Medical AR Research into Clinical Practice

“…Medical AR is usually only one part of a complex system of different technologies, which together form the final medical device solution. For example, many medical AR systems are combined with AI and computer vision algorithms, or collaborate with additional hardware such as tracking systems, robots, or imaging devices [ 2 , 20 , 22 , 23 ]. The system complexity makes the validation of and implementation as a medical device challenging, expensive, and time-consuming.…”

Machine Learning in Spine Surgery: A Narrative Review

“…38 In addition to the 7D Surgical System referenced previously, 35 X23D is a ML-based system that transforms 2D fluoroscopic images into 3D reconstructions of the spine. 39 From these reconstructions, ML performs automatic segmentation, pedicle identification, and screw path suggestion, while remaining sensitive to changes in intraoperative conditions. 40 Physician hand-eye coordination and human inaccuracies and imprecisions, compared with a machine, can always be improved in the operating room (OR).…”