Preclinical evaluation of a permanent magnetic tracking navigation system for dental implant placement: accuracy assessment and enhancement through robot-assisted coordinate calibration

Abstract Background The permanent magnetic tracking navigation system (PMTNS) is a novel approach designed to overcome the line-of-sight limitations of optical tracking systems. However, its accuracy and clinical applicability remain insufficiently validated. This study aimed to evaluate the accuracy of the PMTNS for dental implant placement and determine the extent of accuracy improvement following robotic-assisted coordinate calibration. Methods Ten implant fixtures were placed on a cuboid plate using the PMTNS. Preoperative planning was performed using cone beam computed tomography (CBCT), and implant placement was guided in real time with magnetic tracking sensors attached to the handpiece and jig. Postoperative CBCT scans were superimposed onto the preoperative plans using implant-planning software to quantify deviations. The coronal deviation, apical deviation, angular deviation, and signed depth were evaluated. For error correction, 200 random points were generated within a 100 × 100 × 100-mm space using a robotic arm. The PMTNS identified the actual positions of these points, and positional discrepancies were corrected through sequential rotation and distortion calibration. Results The implant registration accuracy demonstrated a mean coronal deviation of 1.99 ± 0.72 mm, an apical deviation of 2.05 ± 0.79 mm, a signed depth of 1.37 ± 0.59 mm, and an angular deviation of 2.04 ± 0.99°. Pre-calibration errors measured X: 3.44 ± 0.00 mm, Y: 5.47 ± 0.10 mm, and Z: 0.63 ± 0.06 mm. Following rotation calibration, errors were reduced to X: 1.02 ± 0.10 mm, Y: 0.46 ± 0.05 mm, and Z: 1.04 ± 0.36 mm. Subsequent rotation and distortion calibration further minimized the errors to X: 0.30 ± 0.04 mm, Y: 0.12 ± 0.01 mm, and Z: 0.12 ± 0.00 mm. Overall, robot-assisted calibration reduced the intrinsic spatial error of the system by approximately 90%. Conclusions The PMTNS showed potential for clinical application, with angular deviations remaining within clinically acceptable limits. Although the linear deviations slightly exceeded conventional thresholds, robot-assisted calibration significantly enhanced the intrinsic spatial accuracy of the system.