Treffer: Development and validation of machine learning models to predict vertebral artery injury by C2 pedicle screws.

Title:
Development and validation of machine learning models to predict vertebral artery injury by C2 pedicle screws.
Authors:
Ye B; Department of Spine Surgery, Honghui Hospital, Xi'an Jiaotong University, Shaanxi, China.; Yan'an University, Yan'an, China., Sun Y; Department of Spine Surgery, Honghui Hospital, Xi'an Jiaotong University, Shaanxi, China., Chen G; Department of Orthopedics, Chongqing Kaizhou District People's Hospital of Chongqing, Shaanxi, China., Wang B; Department of Spine Surgery, Honghui Hospital, Xi'an Jiaotong University, Shaanxi, China., Meng H; Department of Spine Surgery, Honghui Hospital, Xi'an Jiaotong University, Shaanxi, China., Shan L; Department of Spine Surgery, Honghui Hospital, Xi'an Jiaotong University, Shaanxi, China. drshanlq@163.com.
Source:
European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society [Eur Spine J] 2025 Sep; Vol. 34 (9), pp. 3950-3961. Date of Electronic Publication: 2025 Aug 12.
Publication Type:
Journal Article; Validation Study
Language:
English
Journal Info:
Publisher: Springer-Verlag Country of Publication: Germany NLM ID: 9301980 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1432-0932 (Electronic) Linking ISSN: 09406719 NLM ISO Abbreviation: Eur Spine J Subsets: MEDLINE
Imprint Name(s):
Original Publication: Heidelberg, Federal Republic of Germany : Springer-Verlag, c1992-
References:
Kuris EO, McDonald CL, Palumbo MA, Daniels AH (2021) Evaluation and management of cauda equina syndrome. Am J Med 134:1483–1489. https://doi.org/10.1016/j.amjmed.2021.07.021. (PMID: 10.1016/j.amjmed.2021.07.02134473966)
Abumi K, Shono Y, Taneichi H et al (1999) Correction of cervical kyphosis using pedicle screw fixation systems. Spine 24:2389–2396. https://doi.org/10.1097/00007632-199911150-00017. (PMID: 10.1097/00007632-199911150-0001710586466)
Abumi K, Shono Y, Kotani Y, Kaneda K (2000) Indirect posterior reduction and fusion of the traumatic herniated disc by using a cervical pedicle screw system. J Neurosurg 92:30–37. https://doi.org/10.3171/spi.2000.92.1.0030. (PMID: 10.3171/spi.2000.92.1.003010616055)
Abumi K, Kaneda K, Shono Y, Fujiya M (1999) One-stage posterior decompression and reconstruction of the cervical spine by using pedicle screw fixation systems. J Neurosurg 90:19–26. https://doi.org/10.3171/spi.1999.90.1.0019. (PMID: 10.3171/spi.1999.90.1.001910413121)
Abumi K, Kaneda K (1997) Pedicle screw fixation for nontraumatic lesions of the cervical spine. Spine 22:1853–1863. https://doi.org/10.1097/00007632-199708150-00010. (PMID: 10.1097/00007632-199708150-000109280021)
Abumi K, Takada T, Shono Y et al (1999) Posterior occipitocervical reconstruction using cervical pedicle screws and plate-rod systems. Spine 24:1425–1434. https://doi.org/10.1097/00007632-199907150-00007. (PMID: 10.1097/00007632-199907150-0000710423787)
Abumi K, Itoh H, Taneichi H, Kaneda K (1994) Transpedicular screw fixation for traumatic lesions of the middle and lower cervical spine: description of the techniques and preliminary report. J Spinal Disord 7:19–28. https://doi.org/10.1097/00002517-199407010-00003. (PMID: 10.1097/00002517-199407010-000038186585)
Neo M, Sakamoto T, Fujibayashi S, Nakamura T (2005) The clinical risk of vertebral artery injury from cervical pedicle screws inserted in degenerative vertebrae. Spine 30:2800–2805. https://doi.org/10.1097/01.brs.0000192297.07709.5d. (PMID: 10.1097/01.brs.0000192297.07709.5d16371908)
Garcia Alzamora M, Rosahl SK, Lehmberg J, Klisch J (2005) Life-threatening bleeding from a vertebral artery pseudoaneurysm after anterior cervical spine approach: endovascular repair by a triple stent-in-stent method. Case report. Neuroradiology 47:282–286. https://doi.org/10.1007/s00234-005-1343-2. (PMID: 10.1007/s00234-005-1343-215789201)
Bertalanffy H, Eggert HR (1989) Complications of anterior cervical discectomy without fusion in 450 consecutive patients. Acta Neurochir (Wien) 99:41–50. https://doi.org/10.1007/BF01407775. (PMID: 10.1007/BF014077752667284)
de los Reyes RA, Moser FG, Sachs DP, Boehm FH (1990) Direct repair of an extracranial vertebral artery pseudoaneurysm: case report and review of the literature. Neurosurgery 26:528–533. https://doi.org/10.1097/00006123-199003000-00025. (PMID: 10.1097/00006123-199003000-000252181338)
Tumialan LM, Wippold FJ, Morgan RA (2004) Tortuous vertebral artery injury complicating anterior cervical spinal fusion in a symptomatic rheumatoid cervical spine. Spine 29:E343–348. https://doi.org/10.1097/01.brs.0000134593.98662.d0. (PMID: 10.1097/01.brs.0000134593.98662.d015303043)
Abumi K, Shono Y, Ito M et al (2000) Complications of pedicle screw fixation in reconstructive surgery of the cervical spine. Spine 25:962–969. https://doi.org/10.1097/00007632-200004150-00011. (PMID: 10.1097/00007632-200004150-0001110767809)
Wright NM, Lauryssen C (1998) Vertebral artery injury in C1-2 transarticular screw fixation: results of a survey of the AANS/CNS section on disorders of the spine and peripheral nerves. American association of neurological surgeons/congress of neurological surgeons. J Neurosurg 88:634–640. https://doi.org/10.3171/jns.1998.88.4.0634. (PMID: 10.3171/jns.1998.88.4.06349525707)
Deen HG, Birch BD, Wharen RE, Reimer R (2003) Lateral mass screw-rod fixation of the cervical spine: a prospective clinical series with 1-year follow-up. Spine J Off J North Am Spine Soc 3:489–495.
Pateder DB, Carbone JJ (2006) Lateral mass screw fixation for cervical spine trauma: associated complications and efficacy in maintaining alignment. Spine J Off J North Am Spine Soc 6:40–43. https://doi.org/10.1016/j.spinee.2005.04.010. (PMID: 10.1016/j.spinee.2005.04.010)
Lee S-H, Park D-H, Kim S-D et al (2014) Analysis of 3-dimensional course of the intra-axial vertebral artery for C2 pedicle screw trajectory: a computed tomographic study. Spine 39:E1010–1014. https://doi.org/10.1097/BRS.0000000000000418. (PMID: 10.1097/BRS.000000000000041824859580)
Bloch O, Holly LT, Park J et al (2001) Effect of frameless stereotaxy on the accuracy of C1-2 transarticular screw placement. J Neurosurg 95:74–79. https://doi.org/10.3171/spi.2001.95.1.0074. (PMID: 10.3171/spi.2001.95.1.007411453435)
Wakao N, Takeuchi M, Nishimura M et al (2014) Vertebral artery variations and osseous anomaly at the C1-2 level diagnosed by 3D CT angiography in normal subjects. Neuroradiology 56:843–849. https://doi.org/10.1007/s00234-014-1399-y. (PMID: 10.1007/s00234-014-1399-y25001076)
Maki S, Koda M, Iijima Y et al (2016) Medially-shifted rather than high-riding vertebral arteries preclude safe pedicle screw insertion. J Clin Neurosci Off J Neurosurg Soc Australas 29:169–172. https://doi.org/10.1016/j.jocn.2015.11.026. (PMID: 10.1016/j.jocn.2015.11.026)
Agrawal M, Devarajan LJ, Singh PK et al (2018) Proposal of a new safety margin for placement of C2 pedicle screws on computed tomography angiography. World Neurosurg 120:e282–e289. https://doi.org/10.1016/j.wneu.2018.08.052. (PMID: 10.1016/j.wneu.2018.08.05230144596)
Guo J, He Q, Peng C et al (2023) Machine learning algorithms to predict risk of postoperative pneumonia in elderly with hip fracture. J Orthop Surg 18:571. https://doi.org/10.1186/s13018-023-04049-0. (PMID: 10.1186/s13018-023-04049-0)
Guo J, He Q, Li Y (2024) Development and validation of machine learning models to predict perioperative transfusion risk for hip fractures in the elderly. Ann Med 56:2357225. https://doi.org/10.1080/07853890.2024.2357225. (PMID: 10.1080/07853890.2024.23572253890284711191839)
Chen T, Liu C, Zhang Z et al (2023) Using machine learning to predict surgical site infection after lumbar spine surgery. Infect Drug Resist 16:5197–5207. https://doi.org/10.2147/IDR.S417431. (PMID: 10.2147/IDR.S4174313758116710423613)
Neo M, Matsushita M, Iwashita Y et al (2003) Atlantoaxial transarticular screw fixation for a high-riding vertebral artery. Spine 28:666–670. https://doi.org/10.1097/01.BRS.0000051919.14927.57. (PMID: 10.1097/01.BRS.0000051919.14927.5712671353)
Chin KR, Mills MV, Seale J, Cumming V (2014) Ideal starting point and trajectory for C2 pedicle screw placement: a 3D computed tomography analysis using perioperative measurements. Spine J 14:615–618. https://doi.org/10.1016/j.spinee.2013.06.077. (PMID: 10.1016/j.spinee.2013.06.07724200408)
Elliott RE, Tanweer O, Boah A et al (2012) Comparison of safety and stability of C-2 Pars and pedicle screws for atlantoaxial fusion: meta-analysis and review of the literature. J Neurosurg Spine 17:577–593. https://doi.org/10.3171/2012.9.SPINE111021. (PMID: 10.3171/2012.9.SPINE11102123039110)
Chiapparelli E, Bowen E, Okano I et al (2022) Spinal cord medial safe zone for C2 pedicle instrumentation: an MRI measurement analysis. Spine 47:E101–E106. https://doi.org/10.1097/BRS.0000000000004137. (PMID: 10.1097/BRS.000000000000413734091562)
Madawi AA, Casey AT, Solanki GA et al (1997) Radiological and anatomical evaluation of the atlantoaxial transarticular screw fixation technique. J Neurosurg 86:961–968. https://doi.org/10.3171/jns.1997.86.6.0961. (PMID: 10.3171/jns.1997.86.6.09619171174)
Sanelli PC, Tong S, Gonzalez RG, Eskey CJ (2002) Normal variation of vertebral artery on CT angiography and its implications for diagnosis of acquired pathology. J Comput Assist Tomogr 26:462–470. https://doi.org/10.1097/00004728-200205000-00027. (PMID: 10.1097/00004728-200205000-0002712016382)
Shimizu T, Koda M, Abe T et al (2021) Correlation between osteoarthritis of the atlantoaxial facet joint and a high-riding vertebral artery. BMC Musculoskelet Disord 22:406. https://doi.org/10.1186/s12891-021-04275-9. (PMID: 10.1186/s12891-021-04275-9339411428091766)
Bransford RJ, Russo AJ, Freeborn M et al (2011) Posterior C2 instrumentation: accuracy and complications associated with four techniques. Spine 36:E936–943. https://doi.org/10.1097/BRS.0b013e3181fdaf06. (PMID: 10.1097/BRS.0b013e3181fdaf0621289565)
Yoshida M, Neo M, Fujibayashi S, Nakamura T (2006) Comparison of the anatomical risk for vertebral artery injury associated with the C2-pedicle screw and atlantoaxial transarticular screw. Spine 31:E513–517. https://doi.org/10.1097/01.brs.0000224516.29747.52. (PMID: 10.1097/01.brs.0000224516.29747.5216816753)
Lau SW, Sun LK, Lai R et al (2010) Study of the anatomical variations of vertebral artery in C2 vertebra with magnetic resonance imaging and its application in the C1-C2 transarticular screw fixation. Spine 35:1136–1143. https://doi.org/10.1097/BRS.0b013e3181bb4f21. (PMID: 10.1097/BRS.0b013e3181bb4f2120118834)
Contributed Indexing:
Keywords: Cervical pedicle screw; Machine learning; Neural network; Risk prediction; Vertebral artery injury
Entry Date(s):
Date Created: 20250812 Date Completed: 20251216 Latest Revision: 20251216
Update Code:
20251216
DOI:
10.1007/s00586-025-09168-2
PMID:
40794115
Database:
MEDLINE

Weitere Informationen

Purpose: Cervical 2 pedicle screw (C2PS) fixation is widely used in posterior cervical surgery but carries risks of vertebral artery injury (VAI), a rare yet severe complication. This study aimed to identify risk factors for VAI during C2PS placement and develop a machine learning (ML)-based predictive model to enhance preoperative risk assessment.
Methods: Clinical and radiological data from 280 patients undergoing head and neck CT angiography were retrospectively analyzed. Three-dimensional reconstructed images simulated C2PS placement, classifying patients into injury (n = 98) and non-injury (n = 182) groups. Fifteen variables, including characteristic of patients and anatomic variables were evaluated. Eight ML algorithms were trained (70% training cohort) and validated (30% validation cohort). Model performance was assessed using AUC, sensitivity, specificity, and SHAP (SHapley Additive exPlanations) for interpretability.
Results: Six key risk factors were identified: pedicle diameter, high-riding vertebral artery (HRVA), intra-axial vertebral artery (IAVA), vertebral artery diameter (VAD), distance between the transverse foramen and the posterior end of the vertebral body (TFPEVB) and distance between the vertebral artery and the vertebral body (VAVB). The neural network model (NNet) demonstrated optimal predictive performance, achieving AUCs of 0.929 (training) and 0.936 (validation). SHAP analysis confirmed these variables as primary contributors to VAI risk.
Conclusion: This study established an ML-driven predictive model for VAI during C2PS placement, highlighting six critical anatomical and radiological risk factors. Integrating this model into clinical workflows may optimize preoperative planning, reduce complications, and improve surgical outcomes. External validation in multicenter cohorts is warranted to enhance generalizability.
(© 2025. The Author(s).)

Declarations. Competing interests: The authors declare no competing interests.