3D-printed guides versus computer navigation for pedicle screw placement in the surgical treatment of congenital scoliosis deformities

Abstract

Background

To compare the safety and clinical outcomes of 3D-printed guides versus computer navigation for pedicle screw placement in the correction of congenital scoliosis deformities.

Methods

The study was a single-centre retrospective controlled study and was approved by the hospital ethics committee for the analysis all patients under the age of 18 years with at least 2 years of follow-up. Sixty-three patients who underwent surgical correction for congenital scoliosis deformities in our hospital from January 2015 to December 2020 were divided into two groups based on the decision following preoperative doctor‒patient communication. Among them, 43 patients had pedicle screws placed with 3D-printed guider plates, while the remaining 20 patients had screws inserted with the assistance of computer navigation. The perioperative period, follow-up results and imaging data were compared between the groups.

Results

The operation was completed successfully for patients in both groups. The 3D-printed guide-assisted screw placement technique proved to be significantly superior to the computer navigation technique in terms of operation time, screw placement time, and intraoperative blood loss (p < .05), although the former had more frequent intraoperative fluoroscopies than the latter (p < .05). The mean follow-up time was 41.4 months, and the SRS-22 scores significantly improved in both groups over time postoperatively (p < .05). The 3D-printing group had better SRS-22 scores than the navigation group 6 months after surgery and at the last follow-up (p < .05). Compared with preoperative values, the coronal Cobb angle, local kyphotic Cobb angle, C7-S1 coronal deviation (C7PL-CSVL), and sagittal deviation (SVA) were significantly improved in both groups after surgery (p < .05).

Conclusion

Both techniques achieve the purpose of precise screw placement and proper correction of the deformities. In contrast, the 3D-printed guide-assisted screw placement technique showed advantages in terms of operation time, screw placement time, intraoperative blood loss and patient satisfaction with outcomes.

Keywords

3D-printed guider computer navigation congenital spinal deformity pedicle screw fixation IVa

Introduction

In 1959, Boucher was the first to use pedicle structures to drive screws, and in the 1990s, Professor Suk¹ in Korea pioneered the use of pedicle screws for scoliosis, which became the predominant fixation method for scoliosis implant fusion with good correction, a low incidence of pseudoarthrosis, and secure three-column fixation.^2–4 However, in congenital scoliosis, it is difficult to place pedicle screws due to anatomical changes such as the narrowing and rotation of the pedicle.⁵ Poorly positioned pedicle screws can lead to several complications, such as nerve and vascular injury.⁶ The precise placement of pedicle screws becomes the key to the success or failure of this procedure, and for this reason, various assisted screwing techniques and equipment have emerged, including 2D/3D navigation-assisted screwing, portable electronic navigation openers for screwing, guide plate-assisted screwing, and fluoroscopy-assisted screwing. The high accuracy, safety and reliability of CT navigation-assisted stapling have been reported in the literature.^7,8 However, the CT navigation technique is associated with disadvantages such as a complicated operation, expensive equipment and long operation time. With the rapid development of digital orthopaedics in recent years, the 3D-printed guide-assisted screw placement technique has been increasingly used in orthopaedic surgery for the treatment of spinal deformities because of its simplicity and because no special equipment is required.⁹

A total of 63 patients with congenital scoliosis were treated with 3D-printed guide-assisted screwing and CT navigation-assisted screwing in our hospital over a period of 6 years. In this paper, the safety and clinical efficacy of the two techniques are compared in detail, and their respective advantages and disadvantages are summarized and reported as follows.

Materials and methods

Inclusion and exclusion criteria

Inclusion criteria were as follows: (1) age ≤18 years; (2) a diagnosis of congenital scoliosis; and (3) complete case information.

Exclusion criteria were as follows: (1) the presence of congenital heart disease, renal insufficiency, etc.; (2) the presence of haematologic disorders resulting in abnormal coagulation function, etc.; and (3) less than 2 years of follow-up;

General information

In this retrospective study, patients with congenital scoliosis surgically treated in our paediatric orthopaedic department from January 2015 to December 2020 were analysed, and a total of 63 patients meeting the above criteria were included in this study. The patients were divided into two groups based on the results of preoperative doctor‒patient communication: 43 patients in the 3D group and 20 patients in the CT group. The study was approved by the hospital ethics committee, and all patients were informed about the design of the 3D-printed guide and the use of CT navigation and signed an informed consent form before surgery. The general data of the two groups are shown in Table 1. The differences in age, sex, body mass index (BMI), fusion segment and major deformity segment between the groups were not statistically significant (p > .05).

Table 1.

Comparison of preoperative general data between the groups.

Indicators	3D group (n = 43)	CT group (n = 20)	p value
Age (years, x ± s)	10.22 ± 2.16	9.61 ± 2.70	0.962
Sex (case, male/female)	15/28	6/14	0.702
BMI (kg/m², x ± s)	20.07 ± 4.09	20.54 ± 5.35	0.730
Major segments (case, thoracic/thoracolumbar/lumbar)	10/22/11	4/12/4	0.802
Range of fusion (segment)	7.7 ± 2.5	8.4 ± 2.9	0.330

Surgical methods

After the installation of the spinal cord monitoring device, routine disinfection and towel laying, and according to the preoperative surgical design, surgeons were assisted by a 3D-printed guide plate and CT computer navigation for spinal pedicle screw placement.

Guide plate-assisted method: A thin layer CT scan was performed on the patient’s spine, and the original data obtained from the CT scan were imported into the 3D reconstruction software Mimics 10.01 (Materialise, Belgium) in the DICOM format. First, the original mask of the corresponding segment of the spine was obtained by using the threshold selection technique. Second, the structural gap holes resulting from the data conversion were repaired by using the filling technique, and a new mask was obtained. Then, the 3D reconstructed images of the selected structural region were obtained in the 3D reconstruction option. In the MedCAD module, a 4 mm diameter cylinder was used instead of the screw to design the pedicle screw insertion trajectory, and the roller was slowly dragged to observe the relationship between the position of the cylinder through the corresponding bony structures of the pedicle in the 3D interface, axial plane, and sagittal plane. Then, the preliminary design of the screw insertion trajectory was completed, and the model data were saved in STL format. The saved data in UG Imageware software (EDS, USA) were opened, and whether the screw path penetrated the bone cortex from all angles was observed, with fine-tuning if necessary to ensure the safety and accuracy of the screw path. According to the axis line of the design track, the diameter of the screw path was reset to 2.8 mm, the outer diameter of the guide plate screw path was set to 7 mm, and the height was set to 30 mm to complete the design of the directional tube of the guide template. Then, the anatomical data corresponding to the posterior bony surface of the vertebral plate were extracted and reverse thickened by 2.5 mm to create a reverse substrate with the same morphology, the screw tract data were imported, and the two were combined to reconstruct the prototype of the guide plate (the external screw tract was not penetrated) to form an individualized navigation template for pedicle screws with bilateral positioning tubes. In addition, the directional duct was removed, leaving only the substrate part and the 2.8 mm diameter directional hole to complete the design of the fixed-point guide plate. Using photosensitive resin material, the individualized physical guide plate and 3D solid model of the spine were produced. The process of screw placement assisted by 3D guide plate is shown in Figure 1.

Navigation operation: After exposure, the reference frame was fixed on the spines of the adjacent segments of the operation, and the X-ray machine was used for positioning and confirmation. The point matching method was used to select at least six points on the posterior spine surface with a well-defined bony structure. A guide rod was used to contact the articular protrusion so that it overlapped with the anatomical point of the vertebral arch. on the anterolateral 3D images of the spine. Guided by the navigation system, the screw entry position, direction and depth were determined to complete the screw path.

After screw placement, the vertebral spine was repositioned, and the spinal cord was closely monitored. The corresponding hemivertebral body resection or osteotomy was performed. Ponte osteotomy and bending rod orthosis were performed at the necessary segments according to the specific situation, bone grafting and fusion were performed at the corresponding segments, and the wound was closed layer by layer.

Evaluation indicators

Perioperative data were recorded, including operative time, screw placement time, bleeding volume, intraoperation transmission times, incision length, postoperative drainage, and length of hospital stay. Clinical outcomes were assessed using the Scoliosis Research Society (SRS-22) score. Image examination was performed to determine the accuracy of screw placement according to Gertzbein ∼ Robbins grading criteria.¹⁰ Grades A and B were considered accurate screw placement. Scoliosis and local kyphosis angles, C7-S1 coronal offset (i.e., C7PL-CSVL, horizontal distance between the C7 plumb line and the midsacral perpendicular line), sagittal offset (SVA; i.e., the horizontal distance between the C7 plumb line and the posterior upper corner of the endplate of the upper sacral body), etc., were measured.

Statistical processing

Statistical software SPSS19.0 was used to analyse the data. The measurement data were expressed as (x ± s), and when the data were normally distributed, an independent sample t test was used for comparisons between two groups, and a paired t test was used for comparisons between two time points within groups; when the data were not normally distributed, a rank sum test was used. Count data were tested using the χ² test or Fisher’s exact test. Grade data were compared between the groups using the Mann‒Whitney U test. Differences for which the p value was <0.05 were considered statistically significant differences.

Results

Perioperative conditions

A total of 814 pedicle screws were placed in both groups, including 551 in the 3D group and 263 in the CT group, with 22 screws at the C, D, and E levels in the 3D group and eight in the CT group, the details of which are shown in Table 2. The operation time, screw placement time and intraoperative blood loss in the 3D group were significantly less than those in the CT group (p < .001). The number of fluoroscopies in the 3D group was significantly higher than that in the CT group (p < .05). There were no significant differences in incision length, postoperative drainage volume, or length of hospital stay between the groups (p > .05). There were two patients with damaged pleura, which were sutured a in timely manner. In 3 patients, there was an 80% decrease in spinal cord monitoring motor evoked potentials, of which in two patients, spinal canal exploration and decompression were performed at the site of hemivertebra resection, and in one patient, motor evoked potentials basically returned to normal after proper recovery following partial restoration of the deformity. There was one patient with a loss of sensation and muscle strength in both lower limbs 3 days after surgery, and an emergency investigation was performed, which found that the bone graft had prolapsed into the vertebral canal and compressed the spinal cord. The bone graft was removed for decompression and returned to normal after 2 weeks. Regarding incision healing, among 43 patients in the 3D group, the incisions of 40 patients healed at grade A, three at grade B, and 0 at grade C. Among 20 patients in the navigation group, the incisions of 18 patients healed at grade A, two at grade B, and 0 at grade C. The difference between the groups was not statistically significant (p = .649). In 5 patients with grade B healing, two patients underwent rewound suturing under local anaesthesia. Detailed perioperative data are shown in Table 3. No complications, such as deep incision infection and symptomatic thrombosis, occurred in the two groups. There was no surgical revision due to pedicle screw placement in either group. Typical cases are shown in Figures 2–3.

Table 2.

Malpositioned screws by level and direction.

Level	3D (22)				CT (8)
	Left		Right		Left		Right
	Lateral	Medial	Lateral	Medial	Lateral	Medial	Lateral	Medial
T1		1
T2				1			1
T3	1
T4	1			1	1
T5				1
T6		1
T7	1	1
T8				1				1
T9			1	1
T10				1
T11			2
T12		1
L1		2		1		1
L2			1	1			2
L3		1				1		1
L4
L5

Table 3.

Perioperative data (x ± s) and comparison between the groups of patients.

Indicators	3D group (n = 43)	CT group (n = 20)	p value
Operation time (min)	274.61 ± 94.90	456.39 ± 123.71	<0.001
Screw placement time (min)	53.40 ± 29.15	136.83 ± 37.55	<0.001
Incision length (cm)	17.73 ± 5.22	19.06 ± 6.33	0.383
Intraoperative blood loss (ml)	761.69 ± 204.18	1108.69 ± 457.48	<0.001
Postoperative drainage (ml)	557.62 ± 114.36	524.51 ± 91.77	0.261
Length of hospital stay (d)	13.79 ± 1.86	13.41 ± 2.09	0.471
Incisional healing (A/B/C)	40/3/0	18/2/0	0.649
Number of fluoroscopy (times)	7.61 ± 1.58	4.50 ± 0.83	<0.001

Follow-up results

All 63 patients were followed up with an average follow-up time of 41.4 ± 15.8 months. None of the 43 patients in the 3D group had a loosening or breakage of the internal fixation. Among the 20 patients in the navigation group, one patient had a broken pedicle screw at the caudal end of the surgical fusion segment 2 years after surgery that formed junctional kyphosis, and the deformity was corrected after the revision and extension of the fixed segment. In the remaining 19 patients, there was no loosening or breakage of the internal fixation. The follow-up data of the two groups are shown in Table 4. There was no significant difference in the recovery time to full weight-bearing activities between the groups (p > .05). The SRS-22 score improved significantly in both groups over time postoperatively (p < .05). There was no significant difference in SRS-22 scores between the groups at 1 month postoperatively. The SRS-22 score at 6 months postoperatively and at the final follow-up were better in the 3D group than in the CT group, with statistically significant differences.

Table 4.

Follow-up information (x ± s) and comparison between the groups of patients.

Indicators	Time point	3D group (n = 43)	CT group (n = 20)	p value
SRS-22 score (points)	1 month after surgery	20.39 ± 2.44	19.15 ± 3.28	0.098
	6 months after surgery	21.69 ± 1.30	20.88 ± 1.81	0.047
	Last follow up	22.34 ± 1.41	21.29 ± 1.43	0.008
	p value*	0.003/0.285	0.064/1.000

p value*: Comparison between 6 months and 1 month after surgery/comparison between the last follow-up and 6 months after surgery.

Imaging evaluation

Postoperative CT 3D reconstruction showed that the precise screw placement rate in the 3D group was 96.04 ± 2.06% and 97.02 ± 1.83% in the navigation group, with no significant difference (p > .05). Compared with the preoperative period, the coronal Cobb angle, local kyphosis Cobb angle, C7-S1 coronal deviation (C7PL-CSVL), and sagittal deviation (SVA) were significantly improved in both groups immediately after surgery (p < .05). Compared with the immediate postoperative period, the correction of the lateral bending Cobb angle, local kyphosis Cobb angle, C7PL-CSVL, and SVA were lost in both groups at the last follow-up, but the differences in each index between the time points were not statistically significant (p > .05). From the immediate postoperative period to the last follow-up, there were no statistically significant differences between the groups in the lateral bending Cobb angle, local kyphosis Cobb angle, C7PL-CSVL, or SVA (p > .05) (Table 5).

Table 5.

Image measurement results (x ± s) and comparison between the groups.

Indicators	Time point	3D group (n = 43)	CT group (n = 20)	p value
Precise screw placement rate (%)	Immediate postoperative period	96.04 ± 2.06	97.02 ± 1.83	0.074
Correction rage (%)	Immediate postoperative period	67.26 ± 8.13	65.30 ± 7.33	0.362
Coronal Cobb angle (°)	Preoperative period	77.67 ± 20.24	72.60 ± 18.07	0.343
	Immediate postoperative period	26.56 ± 10.29	23.96 ± 8.51	0.329
	Last follow up	28.26 ± 11.06	27.03 ± 9.13	0.667
	p value*	<0.001/1.000	<0.001/1.000
Local kyphosis Cobb angle (°)	Preoperative period	64.25 ± 23.93	61.38 ± 21.44	0.649
	Immediate postoperative period	19.35 ± 8.20	20.57 ± 10.48	0.649
	Last follow up	16.59 ± 7.46	17.04 ± 8.00	0.828
	p value*	<0.001/1.000	<0.001/1.000
\|C₇PL-CSVL\| (mm)	Preoperative period	33.72 ± 8.81	32.07 ± 9.06	0.495
	Immediate postoperative period	11.52 ± 3.47	10.38 ± 4.20	0.261
	Last follow up	9.81 ± 2.25	9.42 ± 2.96	0.605
	p value*	<0.001/0.482	<0.001/1.000
\|SVA\| (smm)	Preoperative period	22.17 ± 10.72	21.63 ± 11.55	0.856
	Immediate postoperative period	15.32 ± 7.88	16.86 ± 8.67	0.487
	Last follow up	11.31 ± 4.96	10.79 ± 6.35	0.749
	p value*	0.001/0.075	0.309/0.118

|C7PL-CSVL|: Horizontal distance between the central plumb line of the C7 vertebra and the central plumb line of the sacrum, shown with absolute value. |SVA |: Horizontal distance between the central plumb line of the C7 vertebra and the posterior upper corner of the upper endplate of the sacral one vertebra, shown with the absolute value. p value: immediate postoperative compared with preoperative/last follow-up compared with immediate postoperative.

Discussion

Accurate screw placement using the pedicle screw technique for scoliosis remains difficult, especially in the presence of abnormal pedicle development. In this retrospective study, 3D-printed guide-assisted screw placement was compared with intraoperative CT-guided screw placement. The accuracy of 3D-printed guide-assisted screw placement was slightly lower than that of CT navigation-assisted screw placement, but the difference was not statistically significant. 3D-printed guide-assisted screw placement was effective in reducing the operative time in screw placement.

Clearly, there are some shortcomings of this study. First, this study was a retrospective study, and the selection of surgical methods was inconsistent. Especially in the later period of the study, most patients chose 3D-printed guide plate-assisted screw placement for surgery, resulting in an uneven distribution of cases between the groups in this study. Second, 3D spinal models for patients in the CT navigation group were not created to help surgeons understand the specific conditions of the spinal deformity. Third, the sample size was small, especially for the CT navigation group. Last, in this study, only 3D-printed guide plate-assisted screw placement and CT navigation-assisted screw placement were compared, with no evaluation of techniques such as freehand screw placement, although there have been many similar studies.

A total of 814 pedicle screws were implanted in this group, including 30 screws with grade C, D and E placement, 22 in the 3D group and eight in the CT group. Common features of these poorly positioned screws were the poor development of the pedicle and the thin or severe rotation of the vertebral body, which are the main causes of poorly placed pedicle screws. Another possible reason for the poor position of the screws in the 3D group was that the pedicle screws were not placed in the screw path prepared by the guide plate, which should be confirmed and improved in our future work. Moreover, other reasons for the poor positioning of the screws in the CT group was that only four to six screws could be placed in one CT scan, these screws were preferentially placed in the most deformed areas, and not all the pedicle screws were placed by the CT guide. In addition, the accuracy of screw placement is related to the location of screws. It is generally believed that the accuracy of screw placement in thoracic vertebrae is lower than that in lumbar vertebrae, and inaccurately placed screws are more likely to cause vascular injury.^11,12 Other factors include osteoporosis, although this is less common in children. Fortunately, there was no cases of reoperation due to poor screw positioning.

In this study, there were 20 patients who underwent CT navigational screw placement, with 263 screws placed, with an accuracy rate of 97.02%, which was similar to relevant literature reports.^13–15 However, in the process of use, we found that CT navigational screw placement has the following deficiencies. First, computer navigation is complex and has a long learning curve, requiring professional assistance and cooperation and good communication between the surgeon and the relevant operators. Second, navigation equipment is relatively expensive and needs to be used in conjunction with intraoperative CT and special surgical beds, which requires a large economic investment and thus a certain economic burden for medical institutions. Third, navigation instruments and intraoperative CT machines occupy a large area, and a large space is needed to facilitate operation, which requires a certain size of the operating room. Fourth, during the intraoperative CT scan, the length of each scan was relatively short, approximately 4-6 segments. If the scan range is large, two or more scans are needed. Fifth, complicated intraoperative navigation operations, CT scanning and data reconstruction take a certain amount of time, thus prolonging the operation time and increasing the risk of bleeding and infection. Sixth, the intraoperative CT scan radiation dose is relatively large, which will cause certain radiation damage to patients and operators, especially paediatric patients, and some scholars have similar views.^16,17

3D-printed guide screw placement greatly overcomes the disadvantages of computer-guided screw placement and leads to a high accuracy rate of screw placement. In this group, there were 43 patients who underwent guide plate-assisted screw placement, and 551 screws were placed, with an average screw placement accuracy rate of 96.04%. Ming Luo et al.¹⁸ reported that 244 screws were inserted in 15 patients with lateral curvature with Cobb angles greater than 90° by using 3D-printed guide plate fixation. The accuracy rate was 93%, which was significantly better than that of manual fixation and could shorten the operative time of screw placement. There are two advantages. First, the operation is simple, the learning curve is short, and even less experienced surgeons with proper training can achieve accurate screw placement. Second, the cost is low, and no other auxiliary equipment, such as intraoperative CT and carbon fibre surgical beds, is needed. However, there was an increase in the number of fluoroscopies in the 3D-printing group relative to the navigation group.

This study showed that 3D-printed guide screw placement was effective in reducing screw placement time and saving operative time. The average screw placement time was 83.4 min less than that in the navigation group, and even the 3D-printed guide plate screw placement time was less than half of the time used in the CT navigation group. Although all procedures in this group were performed by one team, this longer time for the CT navigation group may also be related to the long learning curve and lack of proficiency in CT navigation techniques. In a multicentre study including 4588 patients, the duration of surgery was found to be an independent risk factor for postoperative complications in spine surgery.¹⁹ Therefore, a shorter duration of surgery is beneficial in reducing the risk of infection and postoperative complications. The imaging of the patients in this study showed a correction rate of 67.26% ± 8.13% in the 3D-printing group and 65.30% ± 7.33% in the navigation group. The literature reports that the correction rate of scoliosis with posterior surgery alone is generally 40%–70%.^20–22 In contrast, Lawrence et al. treated seven children with severe spinal deformities using individualized 3D-printed models and guides and achieved an impressive correction rate of 83%.²³ Patients in both groups were satisfied with the correction of the deformity and achieved the expected results, and we believe the main reason for this is that good pedicle screw positioning provides better holding power and orthopaedic ability. Second, good screw positioning can effectively avoid nerve and vascular damage, and the surgeon can perform deformity correction surgery with more confidence and peace of mind.

The experiences in using guide plate-assisted screw placement are as follows. First, according to the surgeon’s screw placement habit and patient anatomical structure, it is necessary to fully communicate with engineering staff before making the guide plate so that the intraoperative use of the guide plate can be more convenient. Second, a more detailed and complete preoperative design and plan should be created, including osteotomy position, osteotomy mode, spinal shape after osteotomy, the predicted fixation rod shape and the specific type of each screw. In particular, the right screw should be prepared before the surgery to be used directly during the surgery so that the operation time and blood loss can be reduced. Third, appropriate lengthening of the incision, especially the caudal end of the incision, is needed. The blockage of the muscle wall requires appropriate extension of the incision to ensure adequate apposition between the guide plate and the vertebral plate, articular eminence, etc. Fourth, the guide plate and the model should be used in combination. During the intraoperative operation, the guide plate should be compared with the general position on the model and then placed in the body for use so that the position of the guide plate can be better determined. Fifth, Kirschner wire with a diameter of 2.5 mm is generally used during intraoperative operation, while the diameter of the guide hole is designed to be 2.8 mm, which makes the operation easier.

This study has the following limitations. First, this study was a retrospective study, and the sample size of navigation group was small. Second, 3D printing guide plate requires data collection, modeling, design, printing and disinfection, which need time and cost. Third, there was no control group and hand screw.

Conclusion

In this retrospective study, 3D-printed guide plate-assisted screw placement and CT navigation-assisted screw placement can both achieve precise screw placement and proper correction of deformities. In contrast, the 3D-printed guide-assisted screw placement technique has advantages in terms of operation time, screw placement time and intraoperative blood loss, and no special equipment is needed, making it easier for wider application. However, further confirmation is needed in prospective studies.

Footnotes

Author contributions

HLY, XHF, and LC contributed to the study design. XHF and LC contributed to the data analysis and drafted the manuscript. HLY directed data collection and provided administrative support for the project. All authors contributed to interpretation of the data, commented on the manuscript, revised the manuscript, revised the manuscript, and approved the final version for publication.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Natural Science Foundation of China [Grant number: 81171735], the Provincial Natural Science Foundation of Shaanxi [Grant number: 2017JC2-04] and Air Force Military Medical University.

Ethical statement

Informed consent

Informed consent was obtained from all individual participants and their legal guardians included in the study.

ORCID iD

Lu-Yu Huang

Data Availability Statement

The data are available from the corresponding author upon reasonable request and with permission from the Xijing Hospital, Air Force Military Medical University.

Appendix

References

Suk

Lee

Kim

, et al. Segmental pedicle screw fixation in the treatment of thoracic idiopathic scoliosis. Spine 1995; 20(12): 1399–1405.

Kawakami

Goto

, et al. Comparison of vertebral rotation corrected by different techniques and anchors in surgical treatment of adolescent thoracic idiopathic scoliosis. J Spinal Disord Tech 2009; 22(3): 182–189.

Lenke

Kuklo

Ondra

, et al. Rationale behind the current state-of-the-art treatment of scoliosis. Spine 2008; 33(10): 1051–1054.

Kuklo

Potter

Lenke

, et al. Surgical revision rates of hooks versus hybrid versus screws versus combined anteroposterior spinal fusion for adolescent idiopathic scoliosis. Spine 2007; 32(20): 2258–2264.

Sucato

Duchene

. The position of the aorta relative to the spine: a comparison of patients with and without idiopathic scoliosis. J Bone Joint Surg Am 2003; 85(8): 1461–1469.

Hicks

Singla

Shen

, et al. Complications of pedicle screw fixation in scoliosis surgery: a systematic review. Spine 2010; 35(11): E465–E470.

Gelalis

Paschos

Pakos

, et al. Accuracy of pedicle screw placement: a systematic review of prospective in vivo studies comparing free hand, fluoroscopy guidance and navigation techniques. Eur Spine J 2012; 21(2): 247–255.

Baky

Milbrandt

Echternacht

, et al. Intraoperative computed tomography-guided navigation for pediatric spine patients reduced return to operating room for screw malposition compared with freehand/fluoroscopic techniques. Spine Deform 2019; 7(4): 577–581.

Senkoylu

Daldal

Cetinkaya

. 3D printing and spine surgery. J Orthop Surg 2020; 28(2): 2309499020927081. DOI: 10.1177/2309499020927081

10.

Gertzbein

Robbins

. Accuracy of pedicular screw placement in vivo. Spine 1990; 15(1): 11–14.

11.

Parker

McGirt

Farber

, et al. Accuracy of free-hand pedicle screws in the thoracic and lumbar spine: analysis of 6816 consecutive screws. Neurosurgery 2011; 68(1): 170–178.

12.

Parker

Amin

Santiago-Dieppa

, et al. Incidence and clinical significance of vascular encroachment resulting from freehand placement of pedicle screws in the thoracic and lumbar spine: analysis of 6816 consecutive screws. Spine 2014; 39(8): 683–687.

13.

Vissarionov

Kokushin

Khusainov

, et al. Comparing the treatment of congenital spine deformity using freehand techniques in vivo and 3D-printed templates in vitro (Prospective-Retrospective single-center analytical single-cohort study). Adv Ther 2020; 37(1): 402–419.

14.

Jin

Liu

, et al. Does intraoperative navigation improve the accuracy of pedicle screw placement in the apical region of dystrophic scoliosis secondary to neurofibromatosis type I: comparison between O-arm navigation and free-hand technique. Eur Spine J 2016; 25(6): 1729–1737.

15.

Tian

. Image-guided pedicle screw insertion accuracy: a meta-analysis. Int Orthop 2009; 33(4): 895–903.

16.

Chen

, et al. Efficacy and accuracy of a novel rapid prototyping drill template for cervical pedicle screw placement. Comput Aided Surg 2011; 16(5): 240–248.

17.

Kotani

Akazawa

Sakuma

, et al. Accuracy of pedicle screw placement in scoliosis surgery: a comparison between conventional computed tomography-based and O-Arm-Based navigation techniques. Asian Spine J 2014; 8(3): 331–338.

18.

Luo

Wang

Yang

, et al.

Does three-dimensional printing plus pedicle guider technology in severe congenital scoliosis facilitate accurate and efficient pedicle screw placement?

Clin Orthop Relat Res 2019; 477(8): 1904–1912.

19.

Kim

Hsu

De Oliveira

, et al. Operative duration as an independent risk factor for postoperative complications in single-level lumbar fusion: an analysis of 4588 surgical cases. Spine 2014; 39(6): 510–520.

20.

Parsch

Geiger

Brocai

, et al. Surgical management of paralytic scoliosis in myelomeningocele. J Pediatr Orthop B 2001; 10(1): 10–17.

21.

Basobas

Mardjetko

Hammerberg

, et al. Selective anterior fusion and instrumentation for the treatment of neuromuscular scoliosis. Spine 2003; 28(20): S245–S248.

22.

Yazici

Asher

Hardacker

. The safety and efficacy of Isola-Galveston instrumentation and arthrodesis in the treatment of neuromuscular spinal deformities. J Bone Joint Surg Am 2000; 82(4): 524–543.

23.

Karlin

Weinstock

Hedequist

, et al. The surgical treatment of spinal deformity in children with myelomeningocele: the role of personalized three-dimensional printed models. J Pediatr Orthop B 2017; 26(4): 375–382.