Clinical Scorecard: Evaluating the Precision of a Novel 3D-2D Registration Technique Utilizing a Non-Invasive Skin Marker Model for Navigated Spinal Surgery

At a Glance

Key Highlights

• The novel 3D2D registration algorithm enables accurate vertebral registration without bone- or table-mounted instruments.
• Non-invasive hybrid skin markers containing radiopaque spheres facilitate registration and pose estimation without additional fluoroscopic device calibration.
• The algorithm supports low-cost, easy-to-use 3D navigation integrated into routine fluoroscopic-guided minimally invasive spine surgery workflows.

Guideline-Based Recommendations

Diagnosis

• Use preoperative CT scans for 3D anatomical data acquisition.
• Acquire intraoperative 2D fluoroscopic images with spine protocol for optimal vertebral visualization.

Management

• Apply non-invasive hybrid skin markers with radiopaque spheres on patient skin for registration.
• Perform manual indication of vertebral levels on CT and fluoroscopic images to guide segmentation and registration.
• Utilize the 3D2D registration algorithm offline to match vertebrae positions between CT and fluoroscopy.

Monitoring & Follow-up

• Assess registration accuracy by comparing digitally reconstructed radiographs with fluoroscopic images using gradient-based similarity measures.
• Monitor vertebral pose and inter-vertebral deformation during registration to ensure precise navigation.

Risks

• Potential inaccuracies due to spinal curvature changes or vertebral shifting between preoperative CT and intraoperative imaging.
• Dependence on manual identification of vertebral levels may introduce user variability.

Patient & Prescribing Data

Patients undergoing minimally invasive spinal surgery without prior spinal surgery history

The technique facilitates accurate pedicle screw placement with potential for reduced operation time and improved outcomes by integrating 3D navigation into standard fluoroscopic workflows without additional invasive markers or equipment.

Clinical Best Practices

• Ensure proper application and visibility of hybrid skin markers on the patient’s skin before imaging.
• Perform manual vertebral level identification carefully to optimize segmentation and registration accuracy.
• Use two fluoroscopic images with known relative pose for per vertebrae registration to correct for spinal curvature and vertebral shifts.
• Leverage the pose-estimation algorithm to avoid the need for fluoroscopic device calibration or tracking.

References

• Minimally invasive spine surgery outcomes and costs
• Advantages of intraoperative 3D navigation in spine surgery
• Challenges of current 3D imaging and navigation equipment
• Development of the 3D2D registration algorithm
• Hybrid skin marker model for registration
• Gradient-based registration and pose estimation algorithms