Automatic correction of registration errors in surgical navigation systems.

Area of Science:

• Medical imaging
• Surgical navigation systems
• Image-guided surgery

Background:

• Surgical navigation systems are crucial in modern medicine, particularly in ENT and maxillofacial surgery.
• Intraoperative registration, mapping image to patient coordinates, is fundamental for image-guided surgery.
• Current registration methods, despite advancements, have inherent inaccuracies leading to instrument positioning errors.

Purpose of the Study:

• To develop and evaluate an automated method for real-time improvement of intraoperative registration accuracy in surgical navigation.
• To address and correct deviations between navigated instrument positions and their representation in patient image data.

Main Methods:

• The proposed method automatically inspects and validates the projection of navigated instruments into image data.
• Errors in image-to-patient registration are identified by calculating intersections between virtual instrument axes and hard bone surfaces from patient imaging.
• Registration accuracy is improved by adding correction points at detected error locations, integrated into a paired points registration system.

Main Results:

• The developed method was successfully integrated into an existing surgical navigation system.
• Experiments demonstrated that the automatic error correction successfully rectified deliberately misplaced registration points.
• Significant improvements in registration quality were observed in all tested cases.

Conclusions:

• The proposed automated method effectively enhances the accuracy of surgical navigation systems post-initial registration.
• This technique offers a continuous, real-time solution to mitigate registration errors, improving intraoperative guidance.
• The findings suggest a valuable advancement for image-guided surgery, particularly in procedures requiring high precision.