Does Pedicle Morphology Affect the Safety and Accuracy of Pedicle Screw Placement Using 3D-Printed Guides? A 5-Year, Single-Center Experience With 2210 Screws Placed for Adult Spinal Deformity Reconstruction

Freehand Technique

Pedicle screw placement accuracy using freehand technique with or without intraoperative fluoroscopy ranges from 84.6% in the thoracic spine to 90.9% in the lumbar spine.6 Castro et al reported a 71% accuracy rate in lumbar pedicle screw placement.7 Similarly, Weinstein et al reported a 79% accuracy rate for fluoroscopy-guided pedicle screws, with 92% of the breached screws having a breach of the medial pedicle wall.8 Suk et al retrospective study on 462 deformity patients with screws placed via freehand technique showed an accuracy of 89.6%, with 0.8% experiencing new neurological deficits.9 Certainly, in experienced hands, these accuracy rates can be higher, but there appears to be a non-negligible rate of pedicle screw misplacement when using a freehand technique but with a low rate of new neurological or vascular injuries.

Computer-Assisted and Robotic Navigation

The accuracy rates reported using navigation or robotics are generally higher than those reported for the freehand technique. Rajasekaran et al reported in a randomized controlled trial a 98% accuracy in thoracic pedicle screw placement for spinal deformity correction using navigation.10 Rivkin et al retrospective review of 266 patients undergoing thoracolumbar pedicle screw fixation using Stealth navigation (Medtronic, Memphis, TN) showed a 94.7% overall accuracy rate.11 Likewise, Van et al multinational, multicenter, prospective clinical study reported a 97.5% pedicle screw placement accuracy rate using Stealth navigation, with 1.8% classified as “unacceptable” and revised during the same procedure.12 Fan and colleagues compared screws placed using robotic-assisted navigation with screws placed using a freehand technique and found that in the robotic-guided group, 95.3% of screws were perfectly positioned, and 98.7% were deemed acceptable, compared with the freehand technique which resulted in 86.1% of screws being perfectly positioned and 93.5% deemed acceptable (P < 0.01).13 Yu et al compared robotic-assisted vs computer-assisted navigation for screw placement and found smaller screw angular deviations and a higher incidence of acceptable screws in the robot-assisted group suggesting enhanced accuracy using robotic technology.14

There are several downsides to the use of computer-assisted navigation and robotic-assisted navigation for the placement of spinal instrumentation for ASD surgery. First is the potential for shifting of the reference array during the often-long duration of surgery which can cause inaccuracies during navigation and risks misplacement of instrumentation. This can be somewhat mitigated with the use of multiple arrays, but this is cumbersome and alters the normal open ASD surgery workflow. Second is the effect of soft tissue pressure caused by retractors or instruments brought into the field, which can also cause shifting of the registration and inaccuracies in the navigation. Third, there is a tendency when using these technologies to look away from the open spine and at a monitor displaying the images. This removes the visual feedback of seeing the anatomy and can also result in grossly misplaced instrumentation if there is an error in the navigation, often worse than what might happen with an experienced freehand surgeon. Fourth, there can be a significant capital cost associated with acquiring navigation or robotic technology if it is not already present at a hospital. These limitations are significantly mitigated with the use of 3DPSG where there is level-by-level fit of the guides, no need for intraoperative registration, and minimal disruption to the typical workflow for ASD surgery. With regard to cost, 3DPSG are typically purchased on a case-by-case basis and do not require capital expenditure and thus may be easier to get into a hospital system in this era of cost-containment.

Here, we showed that with the use of 3DPSG, we had a 99.5% accuracy rate with pedicle screw placement in ASD surgery. This was in line with the results of an RCT comparing 3DPSG vs freehand placement, which showed an accuracy rate of 3DPSG of 90%.15 To our knowledge, there has not previously been a report examining the accuracy of pedicle screw placement using 3DPSG in adult spine deformity surgery with respect to pedicle morphology. In our study, the screw malposition rate was 0% in Watanabe types C and D pedicles, which are the most challenging pedicles in which to achieve solid and safe fixation.3 Given that these pedicle types are often present in the apex of scoliosis, achieving solid fixation in these pedicles can allow for more aggressive corrective forces to be applied to the spine and likely can improve overall curve correction.

Interestingly, the majority of misplaced pedicle screws were observed to have lateral breaches, all of which occurred on the concave side of the spinal deformity curve. It is important to note that the screw itself is ultimately placed freehand, and the guides only assist with initial cannulation and for expansion of the pedicle using a tap. Given that the lateral pedicle wall has less dense bone than the medial pedicle wall, we hypothesize that it is easier for the screw to inadvertently breach laterally through the soft lateral cortex, whereas misdirection of the screw medially through the hard cortical bone of the medial pedicle wall is more rare.

The implications of these findings carry significant clinical relevance, presenting potential advantages for both surgeons and patients undergoing ASD reconstruction. By facilitating more precise and accurate pedicle screw placement, 3DPSG holds the promise of improving surgical efficiency, reducing operative duration, and diminishing the necessity for intraoperative revisions. This aligns with the conclusions drawn by Lopez et al in their meta-analysis, which indicated that the utilization of 3D-printed pedicle screw templates may lead to enhanced accuracy and precision in screw placement, decreased operative time, and reduced perioperative complications and radiation exposure when compared to non–3D-printed (fluoroscopy or CT) guided procedures.2 Furthermore, the capability to achieve high implant density, even in scenarios involving challenging pedicle morphologies, may contribute to enhanced deformity correction and favorable long-term clinical outcomes for patients with ASD.

Despite the promising results observed in this study, a few limitations warrant consideration. First, as a retrospective cohort study, inherent biases and confounding factors may influence the interpretation of results. Also, the resolution of an O arm image is not equivalent to a formal CT scan; thus, mild breaches might have been difficult to detect especially with the placement of large screws into small pedicles. Given the resolution limitations with O arm imaging, the standardized Gertzbein and Robbins classification system for accurate pedicle screw positioning was not used. Notably, small breaches (Gertzbein-Robbins B or C) are typically clinically insignificant, and notably in our series, there were no vascular injuries and no new neurological deficits associated with screw placement. Additionally, the study’s single-center design and lack of a comparison group limits the generalizability of findings to broader surgical practice settings. Furthermore, our center has a very large experience utilizing 3DPSG, and less experienced surgeons may have more misplaced screws given the learning curve inherent in using the technique. The 4 surgeons in this study have used 3DPSG for a minimum of 5 years. It should also be noted that in this series, 5 of the misplaced screws occurred in 2019 to 2020, and 5 occurred in 2021 to 2023 suggesting no large difference in malposition rate over time. Moreover, while the absence of immediate complications related to screw placement is encouraging, multicenter prospective data are needed to assess the durability and sustainability of surgical outcomes achieved with 3DPSG.

Looking ahead, future research endeavors could focus on prospective, multicenter studies to validate the findings of this study in diverse patient populations and surgical settings and provide direct comparison with navigation or robotically placed screws. Additionally, prospective comparative studies evaluating the cost-effectiveness of 3DPSG vs conventional techniques could provide valuable insights into the economic implications of adopting this technology in ASD surgery.