The results of this study showed that the use of a 3D printed template of the spine can improve the accuracy or pedicle screw placement and reduce complications when used for the treatment of pediatric patients with congenital scoliosis as compared to the freehand method. Clinical outcomes with respect to correction of Cobb angle and kyphosis were not different than when the freehand method was used.
Current screw placement methods include freehand, the use of a navigation system , and robot-assisted pedicle screw placement . The conventional freehand method is associated with low accuracy of screw placement [11, 12]. For patients with congenital scoliosis treated at an early age (3–5 years), hemi-vertebrae excision and short segment fusion may be performed when there is a single or 2-level congenital vertebral defect . Short segment fusion may reduce the number of segments that are fixed; however, this is often associated with instrumentation complications. Especially in pediatric scoliosis patients, pedicle screw misplacement is more likely to occur due to the presence of small pedicles and severely rotated vertebrae, leading to severe complication such as injury to the nerve roots, spinal cord, and major vessels as well as pedicle fracture [21–23]. An intraoperative pedicle fracture requires the need of an additional fused level, and an unidentified pedicle fracture may lead to loss of postoperative correction and neurological complications. Ledonio et al. . conducted a systematic review that included 13,536 pedicle screws placed in 1,353 pediatric patients with spinal deformities, and found that the overall accuracy of pedicle screw placement by the freehand technique was 94.5%; however, the accuracy decreased with more severe spinal deformities.
3D printing technology is increasingly being used for the production of guiding templates for screw placement in spinal operations [25, 26]. However, most studies have focused on adult patients; studies examining the use of 3D printed guiding templates in the treatment of congenital scoliosis are scarce. Recently, Vissarionov et al.  reported that the accuracy of screw placement using 3D printed guiding templates is more accurate than when placed using the freehand method for the treatment of congenital scoliosis (94.4% vs. 53.8%). Nevertheless, the study only compared the accuracy of screw placement; it did not examine the effectiveness of correction or postoperative complications.
In this study, we evaluated the accuracy and safety of 3D printed template-assisted screw placement in the treatment of congenital scoliosis. The results showed that the template-assisted group had significantly more screws placed and fused levels than the freehand group. The group that the 3D navigation templates were used with had more complex deformities and more severe scoliosis, which resulted in the use of more screws and a greater number of levels fused.
Our results showed no significant differences in the amount of intraoperative bleeding and surgical time between the 2 groups. This is mainly because corrective surgery for congenital scoliosis often requires hemivertebra resection, which requires a large amount of time and is associated with marked bleeding. We postulate that the reduction in surgical time and blood loss using the navigation templates were too small to affect the overall operation time and blood loss. On the other hand, a systematic review and meta-analysis by Yu et al.  that primarily included adult patients showed that use of a 3D printed drill guide template in the treatment of spine neurosurgery resulted in a shorter operation time and reduced intraoperative blood loss.
Although the good accuracy rate of pedicle screw placement was comparable between the 2 groups, the template-assisted group had a significantly higher excellent accuracy rate of pedicle screw placement as compared with the freehand group (96.10% vs. 88.64%, P = 0.007). As previously mentioned, Vissarionov et al.  also showed that using 3D printed navigation templates improved the accuracy of screw placement as compared to the freehand method (94.4% vs. 53.8%). Taken together, the results suggest that 3D printed navigation templates can significantly improve the accuracy of screw placement in the surgical treatment of congenital scoliosis. Congenital scoliosis is commonly treated with posterior lumbar fusion, and the large contact area between the posterior lamina and the navigation template assists in improving the accuracy of screw placement.
In this study, the Cobb angle of the main curve and kyphosis angle were significantly decreased after the operation in both groups; however, there were no significant differences between the 2 groups. This finding indicates that both methods are effective for the treatment of congenital scoliosis. However, there were no complications in the templated-assisted group, whereas there were 4 patients that experienced complications in the freehand group and all 4 patients required revision surgery.
There are limitations of this study that require consideration. This was a retrospective study, and the number of patients was relatively small. The primary outcome was the accuracy of screw placement as assessed by postoperative CT. However, the follow-up duration was short with respect to the condition treated and much longer follow-up is necessary to determine the long-term rate of complications and treatment effect.