Prospective Real-Time Screw Placement Using O-Arm Navigation

Abstract

Background/Objectives: A variety of techniques for pedicle screw placement exist. Efficiency claims have varied, but limited data are available to support or refute these claims. Our goal was to study our screw placement efficiency, reporting real-time screw placement using O-arm 3D navigation. Methods: We prospectively enrolled patients from July 2019 to February 2022 who were undergoing primary procedures involving thoracolumbar pedicle and pelvic screw placement with O-arm navigation. Screw time began at the first placement of the navigated probe/awl and ended once the navigated screwdriver was removed from the screw head. Confirmatory 3D scans were performed to assess all screw placements. Results: The real-time average to place pedicle screws was 2 min 9 s (SD ± 1 min 5 s); for pelvic screws, this was 3 min 36 s. Screw placement was slightly faster in pediatric patients (2 min 3 s) vs. adults (2 min 24 s), p < 0.001. Screw placement was faster in the thoracic spine (2 min 2 s) vs. the lumbosacral spine (2 min 22 s), p < 0.001. Screw placement was faster in adolescent idiopathic scoliosis (2 min 0 s) vs. all other diagnoses (2 min 24 s), p < 0.001. Screw placement performed by a single attending surgeon (2 min 24 s) was no different from dual-surgeon placement(2 min 13 s), p = 0.35. Conclusions: Our screw placement time is shorter than previously published estimates, and has a very high accuracy rate. While there are variations in how time is reported compared to the previous literature, our study serves as a benchmark for real-time screw placement for future studies. The use of navigation technology for pedicle and pelvic screw placement can be efficient.

1. Introduction

The use of pedicle screws has steadily grown since its first description by Boucher in 1959 to become the standard of care [1]. Pedicle screws provide definitive fixation for various spinal pathologies due to increased fusion rates, construct rigidity, and deformity-correction capability compared to other fixation devices, such as sub-laminar wires and pedicle hooks [2,3].

Multiple techniques for pedicle screw placement exist. These techniques can be broken into three categories: freehand, conventional fluoroscopy, and image-guided navigation. Freehand screw placement relies on understanding anatomic landmarks in the thoracolumbar spine. The benefits are decreased radiation exposure and presumed efficiency. Reported accuracy rates for this technique have ranged from 69% to 94% [4,5,6,7,8]. Conventional fluoroscopy is a common method of placing pedicle screws. Fluoroscopy facilitates anatomic accuracy of screw placement but increases radiation exposure to the patient and surgical team [9,10,11]. The rate of accuracy for this technique has been reported to be between 88.4 and 93.8% [12]. Image-guided navigation utilizes computer-generated virtual images for tools and screws. Versions include 2D and 3D fluoroscopic-based navigation [13], along with surgeon-controlled and robotic computed tomography (CT)-based navigation. The rate of accuracy for navigated screw placement is between 94.2 and 96.5% [12]. When compared against other screw placement techniques, image-guided navigation has higher reported accuracy rates [12,13,14].

While CT-based navigation techniques have been reported to have higher accuracy rates, adoption of this technology has been limited. Reasons for this include radiation exposure and the perceived increase in screw placement time due to workflow issues. Radiation exposure to the patient and surgeon have been well reported in the literature. Depending on the number of levels fused, the total patient radiation exposure to intraoperative CT scans for navigation can be less than or roughly similar to pre-operative CT scans [15]. By comparison, surgical team radiation exposure can be nearly eliminated using intraoperative lead shields [16]. Workflow efficiency issues leading to increased screw placement time or operative time have been debated, with limited clinical studies to support or refute claims.

Limited prospective data for screw placement time using intraoperative CT-based (O-arm) 3D navigation exists. From the literature, seven studies were found to investigate freehand, using fluoroscopy and C-arm techniques; these studies investigated individual screw placement time. To date, only six studies (two prospective [17,18]; three retrospective [19,20,21]; one cadaveric [22]) have investigated the OR time or pedicle placement time using the O-arm. Furthermore, these studies have only recorded the time for all screws to be placed during surgery, while ours focuses on the time for each individual screw being placed. Rajasekaran et al. performed a randomized controlled trial comparing non-navigated screw placement to navigated; the times taken was 4 min 37 s (±1 min 3 s) versus 2 min 22 s (±43 s). The mean data acquisition time was 24 min 36 s (±6 min 18 s) [23].

However, the number of studies [24,25,26,27] investigating different 3D imaging devices and navigation systems for navigated pedicle screw placement is limited. Furthermore, there have been multiple comments at national meetings that navigation takes longer. This has not been our clinical experience. However, most studies have been retrospective reviews where the methodology has included looking at OR time per number of screws placed. Therefore, we sought to study our accuracy of pedicle screw placement as well as the time needed for screw placement with real-time video recording of screw placement using O-arm navigation. Our institution has been using this technology for twelve years prior to this study. This represents a mature workflow in a high-case volume institution. We hypothesize that our institution’s screw placement time is shorter than what is reported in the available literature (2 min 34 s to 20 min 12 s). This study intends to provide a benchmark for real-time screw placement in the thoracic, lumbar, and sacral spine and pelvis.

2. Materials and Methods

We obtained Institutional Review Board approval before conducting this study. We performed prospective enrollment at our academic institution from July 2019 to February 2022 (Figure 1). We invited patients undergoing navigated screw placement to participate in an intraoperative observation study to determine screw placement time and intraoperative malposition/reposition rates. Eligible patients had to be at least eight years old and receiving primary pedicle screw placement (not revision screw placement) utilizing intraoperative CT-based 3D navigation (O-arm/StealthStation, Medtronic, Minneapolis, MN, USA), for which the application is on-label. We excluded pregnant patients, patients from vulnerable populations (prisoners, unable to consent, etc.), and those with a fracture, tumor, or infection as their primary diagnosis. For adolescent subjects, parental consent, as well as child assent, were obtained.

2.1. Screw Placement Workflow

Surgeons performed pedicle screw placement utilizing the standard technique [28]. This was carried out via an open posterior midline approach and subperiosteal spine exposure (Figure 2). Once complete, the spinous process tracking arc was attached to a spinous process. Then, intraoperative 3D images were obtained with the O-arm and transferred to the image-guided workstation.

Pedicle screws were placed by identifying the starting point and utilizing a burr to create a cortical defect. Then, a navigated awl or navigated drill was used to create a tract, followed by navigated taps, and, lastly, screw placement using a navigated driver. Surgeons determined the appropriate screw sizes using the navigation software. Screw placement is typically carried out from the caudal to the cephalad direction.

For long segment fusions to the sacrum, additional points of pelvic fixation were placed. When placing pelvic screws, the surgeons utilized S2-alar-iliac (S2AI) screws with or without an iliac screw for a kickstand rod. Beginning on one side at the caudal aspect of S2, a navigated awl was used to identify the starting point, and a tract through the sacrum to the sacroiliac (SI) joint was created. Next, the tract was drilled and then serially tapped with increasing diameter using navigated tools. Finally, a navigated screw was placed just above the sciatic notch and then parallel to the iliopectineal line. Depending on the case, a second pelvic screw was placed just below the iliopectineal line, cephalad to the first implanted screw creating a stacked S2AI configuration. This was typically placed bilaterally.

All screws in this study were from the CD HORIZON^® SOLERA^® 5.5/6.0 system (Medtronic SOFAMOR DANEK, Memphis, TN, USA) and were placed with intraoperative CT-based 3D navigation.

2.2. Data Collection

Video recording began at the initial O-arm spin when navigation was established. Individual screw placement times were determined by the following. Timing began when a proposed screw tract was established, either using a burr to make a cortical defect or placing a navigated probe/awl to visualize screw trajectory. Time ended once the navigated driver was removed from the screw head. In the event of a screw removal or repositioning, additional time was added to the initial placement time, beginning when the driver was threaded back onto the screw and ending after the driver was removed from the screw head. Our operative time did not take into consideration instrument changes, intraoperative troubleshooting, or recalibration of the navigation system as this is a rarity in our OR experiences.

After screw placement, intraoperative fluoroscopic radiographs and confirmatory check spins were performed to ensure accurate screw positioning as is our routine. After verifying screw positioning, intraoperative data collection was complete, and we stopped video recording. The surgeons performed the remainder of the surgery per standard of care.

Recorded videos were stored on an encrypted hard drive and stored for subsequent analysis by an independent study team member other than the surgeon. A study team member reviewed each surgical video fully to collect screw placement times. These were carried out for each vertebral level and stratified by laterality.

We collected demographic and surgical information, including age, gender, body mass index (BMI), number of spinal levels instrumented, number of screws placed, and size and diameter of the pedicle and pelvic screws. For this intraoperative observational study, no clinical follow-up data were collected.

2.3. Data Analysis

The primary outcomes included screw placement time, the incidence of intraoperative screw revision, return to OR for screw revisions, and neurologic complications due to screw malposition. Additional analyses were performed to compare thoracic, lumbosacral, and pelvic levels; pediatric (≤18 y.o.) versus adult; adolescent idiopathic scoliosis (AIS) patients versus non-AIS patients; and single- vs. dual-surgeon cases. Descriptive statistics were collected for all patients. The Kruskal–Wallis test was used to determine if there were significant differences across and between thoracic, lumbosacral, and pelvic screw placement times. The Mann–Whitney U test was used for the remainder of the analyses. SPSS Statistics (version 28, IBM (Armonk, New York, NY, USA) was used for all calculations. A p-value of <0.05 was considered significant.

3. Results

We enrolled sixty-three patients; one patient ultimately did not receive surgery, and one patient received surgery, but data were unable to be collected. Forty-two cases were performed as a dual-surgeon team. Primary pathology included degenerative scoliosis, adolescent idiopathic scoliosis, flatback deformity, Scheuermann’s kyphosis, lumbar spondylosis, lumbar spondylolisthesis, and thoracic spondylosis with myelopathy. All patient and surgical demographics are described in

Table 1. Demographics and list of primary diagnoses by frequency of the studied group (N = 61).

Variable	Value
Mean Age	39 ± 26
No. of Females	40
BMI	26.44 ± 6.97
Primary Diagnosis
Adolescent idiopathic scoliosis	22
Degenerative scoliosis	8
Scheuermann’s kyphosis	6
Degenerative spondylolisthesis	5
Flatback syndrome	4
Lumbar stenosis	4
Idiopathic scoliosis	3
Neuromuscular scoliosis	2
Thoracolumbar scoliosis	2
Congenital kyphoscoliosis	1
Congenital scoliosis	1
Isthmic spondylolisthesis	1
Lumbar radiculopathy	1
Thoracic spondylosis	1

.

The overall average time to place an individual pedicle screw (T2-S1) was 2 min 10 s (±1 min 4 s), and the overall average time to place an individual pelvic screw (iliac or S2AI) was 3 min 33 s (±1 min 33 s) (p < 0.001). Breaking down pedicle screw placement time by region, the thoracic spine (T2-T13) was 2 min 2 s and the lumbosacral spine (L1-S1) was 2 min 22 s. There were significant differences across and between thoracic, lumbosacral, and pelvic screw placement times (p < 0.001) (Figure 3). An in-depth breakdown of average screw placement time, screw sizes, malpositions, and intraoperative screw revisions by laterality and vertebral level are shown in

Table 2. Breakdown of average screw placement time, screw sizes, malpositions, and screw revision type by laterality and vertebral level (N = 1092).

LEFT					RIGHT
Change	Size Range	N	Average Screw Placement Time (min:s)	Vertebrae	Average Screw Placement Time (min:s)	N	Size Range	Change
1 redirected	4.5 mm × 25 mm–5.5 mm × 30 mm	8	3:09	T2	2:20	8	4.0 mm × 25 mm–5.5 mm × 35 mm
	4.5 mm × 30 mm–5.5 mm × 30 mm	13	1:50	T3	3:07	11	4.0 mm × 30 mm–5.5 mm × 30 mm	1 redirected
1 removed	4.5 mm × 30 mm–5.5 mm × 35 mm	19	1:55	T4	1:59	13	4.5 mm × 30 mm–5.5 mm × 35 mm
	4.0 mm × 30 mm–6.5 mm × 40 mm	29	1:42	T5	2:12	27	4.5 mm × 30 mm–6.5 mm × 40 mm
1 redirected, 1 removed	4.5 mm × 30 mm–6.5 mm × 45 mm	32	2:01	T6	1:51	32	4.5 mm × 30 mm–6.6 mm × 45 mm	1 lateral cortical engagement, 1 partial medial breach
	4.5 mm × 30 mm–6.5 mm × 40 mm	32	2:04	T7	1:52	30	4.5 mm × 35 mm–6.5 mm × 45 mm
1 redirected	4.5 mm × 30 mm–7.5 mm × 45 mm	34	2:09	T8	2:10	33	4.5 mm × 35 mm–7.5 mm × 45 mm	1 redirected
1 (Malposition and Removed)	4.5 mm × 35 mm–7.5 mm × 45 mm	36	1:56	T9	2 min 2 s	34	4.5 mm × 35 mm–7.5 mm × 45 mm
1 (Removed)	4.5 mm × 35 mm–7.5 mm × 45 mm	40	2:00	T10	1 min 49 s	39	5.5 mm × 35 mm–7.5 mm × 45 mm
	4.0 mm × 25 mm–7.5 mm × 50 mm	42	2:07	T11	2 min	41	4.5 mm × 35 mm–7.5 mm × 50 mm
	4.5 mm × 35 mm–7.5 mm × 50 mm	42	1:52	T12	2 min 11 s	40	4.5 mm × 40 mm–7.5 mm × 50 mm	1 (Partial Medial Breach)
	5.5 mm × 45 mm	1	3:52	T13	2:38	1	5.5 mm × 35 mm
	4.5 mm × 50 mm–7.5 mm × 60 mm	41	2:03	L1	2:17	41	5.0 mm × 35 mm–7.5 mm × 50 mm	1 partial medial breach
1 redirected	4.5 mm × 35 mm–7.5 mm × 65 mm	43	2:18	L2	2:22	43	4.5 mm × 35 mm–7.5 mm × 65 mm	1 removed, 1 shortened, 1 redirected
	5.0 mm × 35 mm–7.5 mm × 65 mm	38	2:07	L3	2:13	38	4.5 mm × 55 mm–7.5 mm × 65 mm
	5.5 mm × 35 mm–7.5 mm × 60 mm	34	2:08	L4	2:29	34	5.5 mm × 30 mm–7.5 mm × 60 mm
	6.5 mm × 35 mm–7.5 mm × 60 mm	26	2:49	L5	2:57	26	6.5 mm × 35 mm–7.5 mm × 60 mm	1 redirected
	6.5 mm × 30 mm–8.5 mm × 40 mm	18	2:38	S1	2:36	19	6.5 mm × 30 mm–7.5 mm × 60 mm	1 shortened
	---	0	---	Iliac	2:29	3	5.5 mm × 50 mm–8.5 mm × 70 mm
2 removed	9.5 mm × 80 mm–9.5 mm × 100 mm	8	3:50	Cephalad S2AI	3:43	9	8.5 mm × 80 mm–9.5 mm × 100 mm	1 removed
	7.5 mm × 70 mm–9.5 mm × 100 mm	17	3:46	Caudal S2AI	3:17	17	9.5 mm × 80 mm–9.5 mm × 100 mm

N = number of screws placed; mm = millimeters; min = minutes; s = seconds; T = thoracic; L = lumbar; S = Sacral; S2AI = Sacral 2 Alar Iliac.

.

We subsequently performed several sub-analyses. These are further broken down into pedicle versus pelvic screw placement time and are summarized in

Table 3. Average screw placement time in pediatric patients vs. adult patients by screw type.

Screw Type	Pediatric Screw Times min:s	Adult Screw Times min:s	p-Value
Pedicle screw (T2-S1)	2:03 (N = 550)	2:17 (N = 448)	0.003
Thoracic	2:01 (N = 424)	2:05 (N = 213)	0.990
Lumbosacral	2:11 (N = 126)	2:27 (N = 275)	0.019
Pelvic screw (Iliac/S2AI)	2:52 (N = 3)	3:35 (N = 51)	0.610
Overall	2:02 (N = 553)	2:24 (N = 539)	<0.001

,

Table 4. Average screw placement time in AIS patients vs. non-AIS patients by screw type.

Screw Type	AIS Screw Times min:s	Non-AIS Screw Times min:s	p-Value
Pedicle screw (T2-S1)	2:00 (N = 460)	2:17 (N = 578)	<0.001
Thoracic	1:59 (N = 358)	2:07 (N = 279)	0.450
Lumbosacral	2:05 (N = 102)	2:27 (N = 299)	0.012
Pelvic screw (Iliac/S2AI)	-	3:33 (N = 54)	-
Overall	2:00 (N = 460)	2:24 (N = 632)	<0.001

and

Table 5. Average screw placement time in single- vs. dual-surgeon cases by screw type.

Screw Type	Single-Surgeon Screw Times min:s	Dual-Surgeon Screw Times min:s	p-Value
Pedicle screw (T2-S1)	2:15 (N = 303)	2:08 (N = 735)	0.150
Thoracic	2:03 (N = 173)	2:02 (N = 464)	0.640
Lumbosacral	2:31 (N = 130)	2:17 (N = 271)	0.019
Pelvic screw (Iliac/S2AI)	3:09 (N = 11)	3:39 (N = 43)	0.260
Overall	2:17 (N = 314)	2:13 (N = 778)	0.350

. The Mann–Whitney U test did not show screw placement time in the AIS thoracic vertebrae group to be statistically different from the non-AIS thoracic vertebrae group (U = 51,661.0, p = 0.45), while screw placement time in the AIS lumbar vertebrae group was statistically significantly higher than the non-AIS lumbar vertebrae group (U = 17,793.0, p = 0.012). Additionally, an independent samples Kruskal–Wallis test did show a statistically significant difference in screw placement time between the three non-AIS vertebral segment groups (p = 0.001).

The incidence of intraoperative screw revision was 21/1092 (1.9%). Of these, 9 (0.82%) were redirected, 9 (0.82%) were removed and not replaced, and 3 were replaced using a shorter screw (0.27%) (

Table 6. Frequency of intraoperative screw changes by type (N = 1092).

Screw Type	Frequency	Percentage (%)
Redirected	9	0.82
Removed	9	0.82
Shortened	3	0.27

). Four screws (0.36%) were malpositioned medially or laterally by less than 2 mm and accepted in accordance with the Gertzbein and Robbins classification scheme [29].

4. Discussion

To our knowledge, this is the largest prospective study on thoracic and lumbosacral pedicle screw placement time, and the only prospective study on pelvic screw placement time using intraoperative video recording and intraoperative CT-based (O-arm) 3D navigation.

4.1. Pedicle Screw Placement

Our pedicle screw placement time is shorter than previously reported in the literature. Studies utilizing a freehand with fluoroscopy technique range from 2 min 44 s to 6 min 38 s [17,20,21,22,23,24,25,26]. A navigated C-arm study reported a 2 min 22 s placement time [23]. Robot-assisted placement times range from 2 min 34 s to 5 min 29 s [30,31,32]. Retrospective studies reporting O-arm times vary between 3 min 33 s and 7 min 48 s [17,18,19,20,22].

Our lumbosacral screw placement time was 2 min 22 s. Our findings contrast with those of Ding et al. [18]. Their prospective study showed an average time of 7 min 48 s for lumbar and sacral screws using O-arm navigation. The studies had a similar number of patients (61 vs. 69), but our study had more screws (401 vs. 380). Possible variables affecting this include institutional, surgical team, navigation technologist, and radiation technologist experience. Moreover, their method of calculating screw placement time differed from ours and was similar to Feng et al. [30].

Our thoracic screw placement time was 2 min 2 s for 637 screws. Liu et al. compared thoracic pedicle screw placement time using conventional fluoroscopy (156 screws) and O-arm navigation (208 screws) and found them to be 6 min 38 s and 5 min 19 s, respectively [19]. These figures did not account for the time spent adjusting the navigation guide probe and pedicle screws, which is different from how we recorded our time. Their screw placement time was longer than ours, but their cohort focused on thoracic spine fracture patients, which is also different from our cohort. A prospective randomized controlled trial by Rajasekaran et al. compared thoracic pedicle screw placement with non-navigated fluoroscopy to C-arm navigation; the reported times were 4 min 37 s vs. 2 min 22 s, respectively [23]. This was consistent with our results.

Shin et al. completed a retrospective study to assess O-arm navigation’s accuracy and clinical benefits for thoracolumbar spine procedures [17]. They reported an average screw insertion time of 4 min 30 s and an average preparatory time of 19 min. However, the report did not specify how these times were collected. In a prospective follow-up study by the same group, their reported average screw insertion time was 4 min 18 s, and the average preparatory time was 14.2 min [20]. Tabaraee et al. showed an average pedicle screw placement of 3 min 33 s utilizing the O-arm compared to 4 min 46 s with the C-arm [22]. This was a cadaveric study with the placement of 80 pedicle screws in the thoracic and lumbar spine.

We compared screw placement times for thoracic versus lumbosacral pedicle screws. There was a 20 s difference. Typically, thoracic screw placement is considered more challenging because the pedicles are smaller in the thoracic region than in the lumbar region. However, this finding was not substantiated by our data.

4.2. Pelvic Screw Placement

Our pelvic screw placement time (3 min 33 s) was longer than our pedicle screw time. This is because we sequentially tapped our pelvic screw requiring at least one additional instrument pass. However, this is still shorter than times found in the literature. Excluding the three iliac screws in our cohort, our S2AI placement time was 3 min 36 s, which is shorter than Zhou et al. [33]. Our S2AI screw placement time did not vary substantially from when our group first reported it at 3 min 42 s [21]. The difference is how we methodically recorded and measured pelvic screw placement for all patients using a real-time approach for our current study.

4.3. Additional Analyses

To our knowledge, there is no existing literature comparing pedicle or pelvic screw placement time in pediatric vs. adult patients using O-arm-based 3D navigation. In this cohort, the mean combined (T2-S1) pedicle screw placement time was significantly shorter in pediatric patients vs. adults. This finding was unexpected given the smaller sized pedicles in pediatric patients. Our placement times are shorter than Morse et al., who reported a 3 min 36 s robot-assisted thoracolumbar pedicle screw placement time in pediatric patients [34]. We found AIS patients had significantly shorter combined pedicle screw placement time, lumbar pedicle screw placement time, and combined screw placement time than non-AIS patients. Our combined pedicle screw placement time in AIS patients was again shorter than that reported by Morse et al. (2 min vs. 3 min 50 s) [34].

Several studies have shown a dual-surgeon strategy results in reduced with a dual-surgeon approach. However, a comparison of individual pedicle screw insertion times with a single-surgeon vs. a dual-surgeon team using O-arm-based 3D navigation using has not been reported. In this cohort, we found only lumbosacral pedicle screw placement time was significantly different when a dual-surgeon team operated vs. a single attending surgeon (2 min 17 s vs. 2 min 31 s). We routinely utilize a dual-surgeon approach for cases we expect to require a long operative time or to be especially challenging, which may have influenced our results.

4.4. Screw Malposition

The reported pedicle screw accuracy of 97.9% in the present study is consistent with both retrospective and prospective data available in the literature. Hagan et al. recently published a large retrospective case series with a reported accuracy rate of 96.9% in 1400 pedicle screws using intraoperative CT and navigation (AIRO mobile; Brainlab, Munich, Germany) [35]. Thirty-seven screws were revised intraoperatively. Knafo et al. performed a retrospective review of pedicle screws placed across four surgeons at one institution using O-arm navigation [36]. An accuracy rate of 95.5% was reported (633/663), while Van De Kelft et al. reported 97.5% accuracy of all pedicle screws placed (1834/1881) [37].

Our 94.1% (48/51) S2AI screw accuracy rate was lower than previous retrospective studies [38,39,40] using intraoperative O-arm-based 3D navigation. However, the previous studies did not specify if any stacked S2AI constructs were attempted; however, we feel this would certainly influence their results. As pelvic fixation techniques have evolved, we have adopted the stacked configuration. While technically demanding, it provides more points of pelvic fixation and allows additional (>2) rods to the pelvis with the goal of preventing construct failure.

4.5. Limitations

There are several limitations to this study. First, we did not prospectively account for the time required for instrument registration (performed as part of the surgical setup not adding time to the case), navigation establishment, and intraoperative 3D confirmatory spins. We routinely obtain a scan with the O-arm after all instrumentation is inserted to verify placement. Thus, patients were not exposed to additional radiation outside the standard of care at our institution due to this study. We did not prospectively record O-arm acquisition time, which was typically less than five minutes, and did not have episodes of having to re-spin the same segment for navigation. Second, although we were able to navigate up to six levels away with reasonable accuracy, we did not perform independent postoperative grading of screw malposition. We recognize that our institution has a mature workflow and workflow definitely affects timing. The quality and experience of the radiology and navigation technologists during the study was also quite high, which enhanced the efficacy of our screw placements, thus leading to greater efficiency. Surgeons deemed a screw to be accurately placed if it was contained entirely within the bone. They assessed all screw placement intraoperatively. Any screw malposition was confirmed by reviewing the operative note and CT scan and was reported. In addition, surgeon experience in this case ranged from 4 years in practice to greater than 30 years in practice; moreover, patient anatomy reflected our practice, which includes pediatric and adult patients. There was significant inclusion of coronal and sagittal plane alignment in both primary and revision cases. Third, we did not quantify the total patient radiation exposure per procedure.

Our real-time average for pedicle and pelvic screw placement was shorter than previously reported in the literature. However, this did vary by level, with the fastest at T10 (1 min 55 s) and the slowest at cephalad S2AI (3 min 15 s). Longer placement time for pelvic screws can be explained by a different workflow involving multiple tapping before placement. We identified disruptions to workflow, such as aberrant anatomy, body habitus, neuromonitoring issues, and navigation difficulties, as potential factors affecting screw placement efficiency.

5. Conclusions

Our real-time average screw placement is shorter than previously reported in the literature. While variations in how time is reported from previous literature exist, our study serves as a benchmark for real-time screw placement for future studies.