Abstract
Background Placement of iliosacral screws is technically challenging. This is especially the case with dysmorphic anatomy. Intraoperative computer navigation can be used to facilitate. No studies have evaluated computer navigation accuracy for minimally invasive sacroiliac joint (SIJ) fusion using iliosacral screws. We sought to evaluate the outcomes of navigated SIJ fusion and the accuracy of virtual screw placement (navigation projection) compared with actual screw placement across the SIJ.
Methods We retrospectively reviewed consecutive minimally invasive SIJ fusions at our institution from July 2020 to November 2023. The virtual trajectory of the iliosacral screws on the computer-guided navigation (Stealth Navigation, Medtronic, Fridley, MN, USA) was compared with actual intraoperative cone beam (O-arm) imaging of the iliosacral screws after placement. The virtual iliosacral screw and actual iliosacral screw angulation and trajectories were compared with respect to the SIJ.
Results Thirteen patients (4 men and 9 women) with 38 total iliosacral screws were analyzed. Five patients (38%) had dysmorphic anatomy. No iliosacral screw malpositions were identified, and no complications occurred. The mean axial angular difference (% variation) between the Stealth projections and the actual iliosacral screw placement was 0.3° (2.7%) on axial images (P = 0.69). The mean coronal angular difference (% variation) between the Stealth projections and the actual iliosacral screw placement was 0.1° (2.6%) on coronal images (P = 0.76).
Conclusion The Stealth navigation iliosacral screw projections accurately represented final implant placement within 1° in this cohort. Results suggest that SIJ fusion iliosacral screw placement using intraoperative Stealth navigation is accurate and reliable.
Clinical Relevance This study provides evidence supporting the accuracy of intraoperative computer navigation for iliosacral screw placement in iliosacral fusion procedures. The findings demonstrate precision, with minimal average angular differences of iliosacral fusion implants in the axial and coronal views. Intraoperative navigation systems can offer an advantage in achieving optimal implant positioning, especially in patients with anatomically complex, dysmorphic anatomy.
Level of Evidence 3.
Introduction
Pelvic anatomy is complex, and as a result, the placement of iliosacral screws (ISSs) is technically challenging. Successful placement requires consideration of pelvic and iliosacral anatomy and knowledge of pertinent anatomic landmarks.1 In the general population, the ISS perforation rate is as high as 68%, and neurological injury can be as high as 7.9%.1 Additional challenges may be caused by bowel gas, obesity, osteoporosis, and transitional anatomy/sacral dysmorphism.1
Sacral dysmorphism further complicates the already technically challenging placement of implants and is essential to consider prior to surgery. Sacral dysmorphism is common, with reported prevalence rates between 3.9% and 35.6%.2
ISSs are commonplace in the trauma and SI joint fusion settings, and historically, various methods have been utilized to ensure proper screw placement without damage to surrounding structures. The use of C-arm for the implantation of ISSs has demonstrated low but still prevalent risks for screw perforation of cortical bone, neurovascular damage, and wound site infection.3–5 Increased body mass index and sacral dysmorphism lead to difficulties with visualization.5 Three-dimensional (3D) fluoroscopy has shown comparable results to C-arm fluoroscopy, as far as complications.5 3D navigation systems allow the surgeon to visualize screw trajectory or a screw projection in real time, not only reducing radiation exposure but also improving screw position with similar operative times.1,6–9
Navigation has been associated with reduced rates of postoperative neurological injury and lower rates of allogeneic transfusion in the deformity setting.10 With navigation, screw position is also improved in patients with sacral dysmorphism undergoing iliosacral fusion.11 The accuracy of 3D navigation has been evaluated and well demonstrated in the thoracic spine, lumbar spine, and S1 level. Namely, the planned trajectory, or virtual representation, and actual implant trajectory have been compared and shown to be highly accurate.12
Although studies have compared conventional and 3D-navigation-assisted techniques1,7–9 for screw placement across the sacroiliac joint (SIJ), no studies have evaluated computer navigation accuracy for minimally invasive SIJ fusion. Namely, no studies have compared the accuracy of virtual screw placement (navigation projection) compared with actual screw placement (intraoperative cone beam scan, O-arm). We investigated the accuracy of implantation in 3D-navigated SIJ fusion based on differences in angulation between the navigation (virtual screw) and actual screw in both the axial and coronal planes. Moreover, our study investigates the accuracy in a population with a high prevalence of sacral dysmorphism.
Materials and Methods
Patient Population
A retrospective evaluation of prospectively collected data was conducted from July 2020 to November 2023 after obtaining approval from the University of Minnesota Institutional Review Board (study #0016469). Patients who underwent minimally invasive SIJ fusions using O-arm and Stealth navigation during the study period at our university were considered. Only patients who had data on virtual as well as actual ISS placement available in our picture archiving system were then included. Any patient without a virtual projection for each screw and a postinstrumentation intraoperative computed tomography (CT) scan, or whose images were of inadequate quality, were excluded. For the purposes of our study, an ISS stabilizing a unilateral SIJ was defined as a standard iliosacral implant.
Surgical Technique and Navigation Utilization
After induction of general anesthesia, the patients were placed prone on a radiolucent table (TRIOS, Mizuho OSI, Union City, CA, USA). A percutaneous reference pin was placed in the contralateral posterior superior iliac spine. A cone beam CT scanner (O-arm, Medtronic Inc., Fridley, MN, USA) was utilized to obtain an intraoperative CT scan. The CT scan information was transferred to the computer-guided navigation system (Stealth, Medtronic). The navigation system allowed the surgeon to visualize virtual projections of instruments in the axial and coronal planes, all overlaid onto the initial CT scan. Using the computer navigation system imaging, the surgeon identified a desired starting point for the ISS. A virtual implant projection was laid along the trajectory of the instrument. A Kirschner wire was placed using a navigated drill guide. The wire was then checked by fluoroscopy using the imaging system. A tract was made through the desired screw entry point. The implant was then placed over the K wire. 2D fluoroscopy and an intraoperative CT scan were performed to confirm the desired implant position once all implants had been placed. Implants were either screw fusion devices (TORQ, SI Bone, Santa Clara, CA) or nonscrew fusion devices (iFuse 3D, SI Bone, Santa Clara, CA).
Statistical Analysis
Our group reviewed all virtual implant projections and intraoperative CT scans after screw placement along with the associated operative reports. Our study utilized a variation of the angle deviation from ideal orientation technique popularized by Chen et al to evaluate the angulation of all virtual projections and iliosacral implants.3 Theoretically, as Chen et al discuss, the ideal axial and coronal plane angles between ISSs and SIJ are 90°; the angles deviated from these ideal orientations are the angle deviation from ideal orientation of ISS.3 In our study, however, we measured the angulation between all virtual projections and ISSs with relation to the SIJ. Our desired implant placement was not always exactly perpendicular to the joint, especially in patients with dysmorphic sacra. Measurements were completed in the axial and coronal planes. Figure 1 demonstrates an example of virtual implantation, or, in other words, the navigation feedback that the surgeon receives during the implantation process. Furthermore, Figure 1 demonstrates the resulting screw. IntelliSpace Radiology 4.7 Enterprise (Philips, Amsterdam, Netherlands) was used for angle measurements. The percent deviation for both axial and coronal angles was calculated as follows:
Virtual screw (left) angulation and actual screw (right) angulation in the axial (top) and coronal (bottom) planes.
Independent t tests were used to compare the accuracy of the virtual screw projections and actual screw placements between the 2 planes (axial and coronal). The linear correlation between the percent variation and the screw angle difference was assessed using the Spearman rank correlation for both axial and coronal planes. Results were considered statistically significant at a probability value of <0.05. Analyses were performed using SPSS Statistics version 28.0 (IBM Corp., Armonk, NY, USA).
Results
A total of 13 consecutive patients (4 men and 9 women) who underwent SIJ fusion with the use of the O-arm and Stealth navigation system at our institution were included. Each screw represented an opportunity for malposition and therefore constituted a legitimate separate event. This process follows the analysis by Miller et al12 for pedicle screw accuracy, where each screw was treated as a separate event. Five patients had dysmorphic anatomy (38%); an example patient with dysmorphic anatomy is depicted in Figure 2.
A total of 38 unilateral standard ISSs (36 screw fusion devices and 2 nonscrew fusion devices) were included in the study. The screws were inserted in either S1 or S2. We routinely use navigation for screw placement and have redundant systems in place, so a system malfunction is quite rare (<1%). No complications (n = 0) occurred in any of the included cases. Mean age was 55.8 ± 16.7 years, body mass index was 29.1 ± 6.5 kg/m2, and height was 1.7 ± 0.1 m. ISSs ranged from 10.0 to 11.5 mm in diameter and 35 to 75 mm in length, and average blood loss was 20.5 ± 13.5 mL. S1 (n = 24) and S2 (n = 14) screws averaged 60.2 mm and 54.6 mm in length, respectively. The screw insertion angle in both the axial and coronal planes correlated with the deviation of the screw position at the sacral level (Table 1).
(a) Axial computed tomography scan slice demonstrating an iliosacral implant in a patient with a dysmorphic sacrum. The implant shown is 10 mm in diameter by 65 mm in length screw fusion device (TORQ, SI Bone, Santa Clara, CA). (b) Postoperative outlet view of the pelvis demonstrating the 4 iliosacral implants the patient received intraoperatively.
Screw insertion angle and its correlation with the percent variation of iliosacral screws at each sacral level.
In the present study, we considered all screws holistically, regardless of sacral level. After doing so, we considered screws at individual sacral levels; in other words, S1 and S2 level screws were considered separately. Mean angular difference and percent variation were calculated for each individual screw, and thereafter, for each individual level. When evaluating all screws holistically, in the axial plane, the average angles between the virtual and actual were 103.9° and 104.2°, respectively. The mean axial angular difference (% variation) between the Stealth projections and the actual screw placement was 0.3° (2.7% variation, 95% CI: −1.60 to 0.99, P = 0.69). In the coronal plane, average angles between the virtual and actual were 94.0° and 93.9°, respectively. The mean coronal angular difference (% variation) between the Stealth projections and the actual screw placement was 0.1° (2.6% variation, 95% CI: −0.88 to 1.20, P = 0.75; Table 2).
Screw percent variation rates of iliosacral screws inserted using Stealth (virtual implant) and intraoperative O-arm (actual implant) navigation.
When considering the S1 level, the average axial plane angles between the projection (virtual) and screws (actual) were 103.6° and 104.4°, respectively. The mean axial angular difference (% variation) between the Stealth projections and the screws was 0.8° (3.0% variation, 95% CI: 1.97–4.13, P < 0.001). In the coronal plane, the projection (virtual) and screw (actual) angles averaged 94.7° and 95.2°, respectively. The mean coronal angular difference (% variation) between the Stealth projections and the screws was 0.5° (2.5% variation, 95% CI: 1.59–3.50, P < 0.001; Table 2).
When considering the S2 level, the average axial plane angles between the projection (virtual) and screws (actual) were 104.5° and 103.8°, respectively. The mean axial angular difference (% variation) between the Stealth projections and the screws was 0.7° (2.1% variation, 95% CI: 0.65–3.49, P = 0.008). In the coronal plane, the projection (virtual) and screw (actual) angles averaged 92.8° and 91.7°, respectively. The mean coronal angular difference (% variation) between the Stealth projections and the screws was 1.1° (2.8% variation, 95% CI: 1.58–3.95, P < 0.001; Table 2).
Aside from these measurements, we considered the need for screw repositioning and any novel neurological deficits for patients. In our study, no screws required repositioning, and no new neurological deficits occurred.
Discussion
This study showed the accuracy of intraoperative cone beam CT and computer-guided navigation in the placement of ISSs in a lateral-to-medial trajectory. The variance between the virtual and actual angle was generally less than 1°. This degree of accuracy enabled highly reliable screw placement.
No screw malpositions occurred. Malposition rates after implantation of ISSs with fluoroscopic guidance have been reported to range from 2% to 24%.8,11,13,14 We acknowledge that certain studies include the traumatic setting, which may impose a greater degree of technical difficulty due to associated fracture displacement. Our patient population had a high degree of sacral dysmorphism (5/13 patients [38%]), which is greater than the high end of the reported sacral dysmorphism prevalence in the population (3.9%–35.6%).2 Sacral dysmorphism is a risk factor for ISS malposition, as it can lead to alterations of normal anatomy and reduction of safe corridors for implants.11,15,16 Therefore, we achieved these results in an otherwise anatomically challenging population.
In addition to a high degree of sacral dysmorphism, our study contained rather large screws (usually 10 mm in outer diameter or greater) in all patients. Prior studies typically utilized 6.5 to 8.0 mm outer diameter screws.11,17 Table 3 demonstrates the size of screws utilized at each corresponding sacral level. A total of 35 out of 38 (approximately 92%) of our screws had a diameter of 10 mm.
Size of iliosacral fusion implants.
Finally, a high concordance between virtual and actual screw angulation is also seen in our study (0.3° mean angle difference in the axial plane and 0.1° mean angle difference in the coronal plane). These results are consistent with and improved compared with a similar study evaluating the accuracy of pedicle screws.12 The difference in angulation calculated in their study was approximately 2.8° and 2.17° in the axial and sagittal planes, respectively. In this case series, there were no foraminal violations. Foraminal violation is an implant length issue vs a navigation accuracy issue. In the senior author’s experience of >330 primary SI fusions with an average of 2.7 implants, the rate of foraminal violation was 1/891. This was recognized intraoperatively and repositioned. No unplanned return to the OR occurred due to implant malposition. Unfortunately, the navigation screen data were not saved in all these cases to allow for a greater n. Our study’s improved accuracy is perhaps in part attributed to the less mobile iliosacral joint, whereas the thoracic and lumbar anatomy are subject to increased motion during implantation, with the reference frame being remote from some of the instrumented levels. While we believe our study is unique, it is not without limitations. These parameters have never been well established. Of note, the senior author’s approach has been to position screw trajectories so that any pin advancement of screws being too long would still be contained within bone.
Although our study offers a new mode in evaluating the accuracy of ISSs, it does have limitations. Our study had a relatively small sample size (n = 38), which limits the power and generalizability of our findings. There is no comparable literature in this space looking at image-guided accuracy of ISS placement. This analysis was done similar to pedicle screw analysis but was more complex, as there is not a clearly established methodology for grading screw placement. The variation of actual vs virtual placement with 38 iterations is a reasonable number, especially given the absence of any other literature on the topic. Because this study was not prospective, our cohort is a sample where sufficient data points were available to allow for analysis. Notably, saving navigation images for later analysis is not part of our surgical workflow. Significant prospective planning efforts occurred to collect data for our study. Nonetheless, our study was characterized by minimal angular difference between the navigation and screw placement. Considering our patient population with high degrees of sacral dysmorphism and the large diameter screws that were used, we believe that valid conclusions can be drawn from this study regarding the accuracy of intraoperative navigation for the placement of ISSs.
Additionally, the clinical manifestations of navigation and actual screw trajectory differences in our study are poorly studied to date. It is unclear how much of an angular difference correlates with screw malposition resulting in clinical complications (Figure 3). Our cases did not require revision of screws, and no patients sustained neurological complications; trajectory differences in our study appear to be minimal and likely clinically unimportant.
Varying degrees of accuracy from an ideal screw trajectory in the sacrum.
Lastly, our study may not be generalizable to some centers that carry out similar surgeries. At our institution, cone beam CT and intraoperative navigation have been used routinely for spinal instrumentation since 2007 and for SI fusion since 2010. The operating room personnel, the radiology technologists, and spine surgeons alike have significant experience and comfort with intraoperative navigation, which may not be reflective of other sites, especially those with recent integration of navigation. We think high volume affects workflow rather than accuracy. Experience may affect the ability to discern sources of error, which would most likely be due to loosening of the percutaneous pin reference frame. Also of note, the surgeon who performed the placement of ISSs is highly experienced, having placed around 890 ISSs, with a single implant requiring 2 mm of backout for neuroforaminal encroachment discerned intraoperatively. This patient had no neurological deficits postoperatively.
Conclusion
Our study provides evidence supporting the accuracy of intraoperative computer navigation for screw placement in iliosacral fusion procedures, with no associated neurological complications or need for revision. The findings demonstrate precision, with minimal average angular differences of ISSs in the axial and coronal views. Our study suggests that intraoperative navigation systems can offer an advantage in achieving optimal implant positioning, with relatively large implants. It appears this conclusion may especially be the case in patients with anatomically complex, dysmorphic anatomy.
Footnotes
Funding The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests D.P. has consultancies with SI-Bone, Globus, Spinal Stabilization Technologies, and Curiteva; royalties from SI Bone and Springer (textbook); research support from SI Bone and Medtronic; and stock in SeeAll AI (<0.03%).
Ethical Approval/Review This project was approved by the University of Minnesota Institutional Review Board (study #0016469). All procedures were conducted in accordance with the Declaration of Helsinki. The requirement for individual informed consent was waived due to the retrospective nature of the study and the use of deidentified clinical data.
- This manuscript is generously published free of charge by ISASS, the International Society for the Advancement of Spine Surgery. Copyright © 2026 ISASS. To see more or order reprints or permissions, see http://ijssurgery.com.
References
- 1.↵Takao M , Nishii T , Sakai T , Yoshikawa H , Sugano N . Iliosacral screw insertion using CT-3D-fluoroscopy matching navigation. Injury. 2014;45(6):988–994. 10.1016/j.injury.2014.01.015
- 2.↵Matson DM , Maccormick LM , Sembrano JN , Polly DW . Sacral dysmorphism and lumbosacral transitional vertebrae (LSTV) review. Int J Spine Surg. 2020;14(Suppl 1):14–19. 10.14444/6075
- 3.↵Chen J-P , Tsai P-J , Su C-Y , et al . Percutaneous iliosacral screw and trans-iliac trans-sacral screw with single C-arm fluoroscope intensifier is a safe treatment for pelvic ring injuries. Sci Rep. 2022;12(1):368. 10.1038/s41598-021-04351-z
- 4.↵Rommens PM , Nolte EM , Hopf J , Wagner D , Hofmann A , Hessmann M . Safety and efficacy of 2D-fluoroscopy-based iliosacral screw osteosynthesis: results of a retrospective monocentric study. Eur J Trauma Emerg Surg. 2021;47(6):1687–1698. 10.1007/s00068-020-01362-9
- 5.↵Kaiser SP , Gardner MJ , Liu J , Routt MLC , Morshed S . Anatomic determinants of sacral dysmorphism and implications for safe iliosacral screw placement. J Bone Joint Surg Am. 2014;96(14):e120. 10.2106/JBJS.M.00895
- 6.↵Zwingmann J , Konrad G , Kotter E , Südkamp NP , Oberst M . Computer-navigated iliosacral screw insertion reduces malposition rate and radiation exposure. Clin Orthop Relat Res. 2009;467(7):1833–1838. 10.1007/s11999-008-0632-6
- 7.↵Takeba J , Umakoshi K , Kikuchi S , et al . Accuracy of screw fixation using the O-arm® and stealthstation® navigation system for unstable pelvic ring fractures. Eur J Orthop Surg Traumatol. 2018;28(3):431–438. 10.1007/s00590-017-2075-9
- 8.↵Zwingmann J , Konrad G , Mehlhorn AT , Südkamp NP , Oberst M . Percutaneous iliosacral screw insertion: malpositioning and revision rate of screws with regards to application technique (navigated vs. conventional). J Trauma. 2010;69(6):1501–1506. 10.1097/TA.0b013e3181d862db
- 9.↵Guerin G , Laghmouche N , Moreau PE , Upex P , Jouffroy P , Riouallon G . Iliosacral screwing under navigation control: technical note. Orthop Traumatol Surg Res. 2020;106(5):877–880. 10.1016/j.otsr.2020.03.024
- 10.↵Chan V , Etigunta S , Gausper A , et al . Navigation is associated with lower risk of neurological injury and transfusion in pediatric idiopathic scoliosis surgery. Spine Deform. 2025;13(6):1911–1919. 10.1007/s43390-025-01140-w
- 11.↵Boudissa M , Carmagnac D , Kerschbaumer G , Ruatti S , Tonetti J . Screw misplacement in percutaneous posterior pelvic iliosacral screwing with and without navigation: a prospective clinical study of 174 screws in 127 patients. Orthop Traumatol Surg Res. 2022;108(2):103213. 10.1016/j.otsr.2022.103213
- 12.↵Miller CA , Ledonio CG , Hunt MA , Siddiq F , Polly DW . Reliability of the planned pedicle screw trajectory versus the actual pedicle screw trajectory using intra-operative 3D CT and image guidance. Int J Spine Surg. 2016;10:38. 10.14444/3038
- 13.↵Hinsche AF , Giannoudis PV , Smith RM . Fluoroscopy-based multiplanar image guidance for insertion of sacroiliac screws. Clin Orthop Relat Res. 2002;395(395):135–144. 10.1097/00003086-200202000-00014
- 14.↵Templeman D , Schmidt A , Freese J , Weisman I . Proximity of iliosacral screws to neurovascular structures after internal fixation. Clin Orthop Relat Res. 1996;329(329):194–198. 10.1097/00003086-199608000-00023
- 15.↵Berger-Groch J , Lueers M , Rueger JM , et al . Accuracy of navigated and conventional iliosacral screw placement in B- and C-type pelvic ring fractures. Eur J Trauma Emerg Surg. 2020;46(1):107–113. 10.1007/s00068-018-0990-z
- 16.↵Cai H , Yin Y , Zhang R , Liu L , Wang T , Hou Z . Simplifying the orientation of S1 iliosacral screws for placement in the dysmorphic sacrum. J Bone Joint Surg Am. 2025;107(6):621–627. 10.2106/JBJS.23.01416
- 17.↵Bastian JD , Jost J , Cullmann JL , Aghayev E , Keel MJB , Benneker LM . Percutaneous screw fixation of the iliosacral joint: optimal screw pathways are frequently not completely intraosseous. Injury. 2015;46(10):2003–2009. 10.1016/j.injury.2015.06.044
