Aneysis D. Gonzalez-Suarez, Paymon G. Rezaii, Daniel Herrick, Seth Stravers Tigchelaar, John K. Ratliff, Mirabela Rusu, David Scheinker, Ikchan Jeon, Atman M. Desai
Neurospine 2024;21(2):620-632. Published online May 20, 2024
Objective Readmission rates after posterior cervical fusion (PCF) significantly impact patients and healthcare, with complication rates at 15%–25% and up to 12% 90-day readmission rates. In this study, we aim to test whether machine learning (ML) models that capture interfactorial interactions outperform traditional logistic regression (LR) in identifying readmission-associated factors.
Methods The Optum Clinformatics Data Mart database was used to identify patients who underwent PCF between 2004–2017. To determine factors associated with 30-day readmissions, 5 ML models were generated and evaluated, including a multivariate LR (MLR) model. Then, the best-performing model, Gradient Boosting Machine (GBM), was compared to the LACE (Length patient stay in the hospital, Acuity of admission of patient in the hospital, Comorbidity, and Emergency visit) index regarding potential cost savings from algorithm implementation.
Results This study included 4,130 patients, 874 of which were readmitted within 30 days. When analyzed and scaled, we found that patient discharge status, comorbidities, and number of procedure codes were factors that influenced MLR, while patient discharge status, billed admission charge, and length of stay influenced the GBM model. The GBM model significantly outperformed MLR in predicting unplanned readmissions (mean area under the receiver operating characteristic curve, 0.846 vs. 0.829; p < 0.001), while also projecting an average cost savings of 50% more than the LACE index.
Conclusion Five models (GBM, XGBoost [extreme gradient boosting], RF [random forest], LASSO [least absolute shrinkage and selection operator], and MLR) were evaluated, among which, the GBM model exhibited superior predictive performance, robustness, and accuracy. Factors associated with readmissions impact LR and GBM models differently, suggesting that these models can be used complementarily. When analyzing PCF procedures, the GBM model resulted in greater predictive performance and was associated with higher theoretical cost savings for readmissions associated with PCF complications.
Citations
Citations to this article as recorded by
Predicting Healthcare Utilization Outcomes With Artificial Intelligence: A Large Scoping Review Carlos Gallego-Moll, Lucía A. Carrasco-Ribelles, Marc Casajuana, Laia Maynou, Pablo Arocena, Concepción Violán, Edurne Zabaleta-del-Olmo Value in Health.2026; 29(1): 159. CrossRef
Artificial intelligence in spine surgery: a scoping review Anis Choucha, Morgane Evin, Matteo de Simone, Guillaume Dannhoff, Henry Dufour, Valentin Avinens, Kaissar Farah, Florian Saby, Stephane Fuentes Neurochirurgie.2026; 72(1): 101764. CrossRef
Artificial intelligence in spine surgery Cheng Zhang, Shanshan Liu, Jialin Shi, Xingyu Zhou, Peter Passias, Nanfang Xu, Weishi Li Spine Research.2025; 1(1): 13. CrossRef
Objective We propose that cervical intrafacetal fusion (cIFF) using bone chip insertion into the facetal joint space additional to minimal PLF is a supplementary fusion method to conventional posterolateral fusion (PLF).
Methods Patients who underwent posterior cervical fixation accompanied by cIFF with minimal PLF or conventional PLF for cervical myelopathy from 2012 to 2023 were investigated retrospectively. Radiological parameters including Cobb angle and C2–7 sagittal vertical axis (SVA) were compared between the 2 groups. In cIFF with minimal PLF group, cIFF location and PLF location were carefully divided, and the fusion rates of each location were analyzed by computed tomography scan.
Results Among enrolled 46 patients, 31 patients were in cIFF group, 15 in PLF group. The postoperative change of Cobb angle in 1-year follow-up in cIFF with minimal PLF group and conventional PLF group were 0.1° ± 4.0° and -9.7° ± 8.4° respectively which was statistically lower in cIFF with minimal PLF group (p = 0.022). Regarding the fusion rate in cIFF with minimal PLF group in postoperative 6 months, the rates was achieved in 267 facets (98.1%) in cIFF location, and 244 facets (89.7%) in PLF location (p < 0.001).
Conclusion Postoperative sagittal alignment was more preserved in cIFF with minimal PLF group compared with conventional PLF group. Additionally, in cIFF with minimal PLF group, the bone fusion rate of cIFF location was higher than PLF location. Considering the concerns of bone chip migration onto the spinal cord and relatively low fusion rate in PLF method, applying cIFF method using minimized PLF might be a beneficial alternative for posterior cervical decompression and fixation.
Objective C2 slope (C2S), a cervical parameter mathematically approximated as T1 slope minus cervical lordosis (T1S–CL), predicts functional improvement in cervical deformity patients. Nonetheless, C2S is a positional parameter based only on the horizontal axis. The current study aims to introduce novel odontoid parameters and establish their relationships with patient-reported health-related quality of life (HRQoL).
Methods Lateral plain radiographs of 32 adults who underwent multilevel posterior cervical fusion were analyzed. The odontoid parameters included odontoid incidence (OI), C2S, odontoid tilt (OT), and gravity line-C2 distance (GL-C2), while the cervical parameters were the Cobb angle at C0–1, C1–2, C0–2, C2–7, C2–7 sagittal vertical axis (cSVA), T1 slope, and T1S–CL. The range of motion (ROM) of the occipito-atlantoaxial complex was measured in flexion and extension plain radiographs. Scores on the Neck Disability Index (NDI) and visual analogue scale (VAS) for axial neck (VASn) and arm pain were measured.
Results Compared to asymptomatic subjects, patients had larger C2S, cSVA, and T1S–CL, and smaller OT. Preoperatively, OI was significantly correlated with the ROM of C1–2 (r = 0.37, p < 0.05) and C0–2 (r = 0.46, p < 0.01). OT and C2S had significant correlations with the C0–1, C1–2, and C0–2 angles, GL-C2, and T1S–CL. Postoperative NDI scores were significantly correlated with OI (r = -0.40, p < 0.05) and OT (ρ = -0.37, p < 0.05). VASn was significantly correlated with GL-C2 (r = -0.35, p < 0.05).
Conclusion The odontoid parameters were significantly correlated with established cervical parameters and HRQoL measures. OI is a constant parameter representing the individual's compensatory reservoir at the upper cervical spine.
The effect of cervical spine flexion-extension motion on odontoid parameters Longao Huang, Dun Liu, Hongyuan Xu, Junfei Feng, Tao Kang, Shengwang Wei, Hua Jiang Journal of Orthopaedic Surgery and Research.2025;[Epub] CrossRef
Correlation analysis between preoperative odontoid incidence and clinical outcomes 2 years after anterior cervical discectomy and fusion Yan Gong, Hao Liu, Yanchi Gan, Jiahui He, Zelin Zhou, Hang Zhuo, Yu Liu, De Liang, Hui Ren, Xiaobing Jiang, Zhaojun Cheng Journal of Orthopaedic Surgery and Research.2025;[Epub] CrossRef
Lower C2 slope and milder uncovertebral joint degeneration are risk factors for pseudarthrosis after single-level anterior cervical corpectomy and fusion (ACCF): retrospective study of 102 patients with minimum 2-year follow-up Haoxiang Wang, Tian Xia, Ruomu Qu, Hanbo Geng, Yu Sun, Fengshan Zhang, Shengfa Pan, Xin Chen, Yanbin Zhao, Feifei Zhou Journal of Orthopaedic Surgery and Research.2025;[Epub] CrossRef
Impact of anterior controllable antedisplacement and fusion (ACAF) on cervical lordosis and sagittal alignment in OPLL: A comparative radiographic analysis Shunmin Wang, Dong Liu, Jincui Li, Jingchuan Sun, Tiefeng Li, Jiangang Shi Journal of Orthopaedics.2025; 70: 158. CrossRef
Odontoid parameters in subjects with and without degenerative cervical spondylosis Tao Kang, Weiyou Chen, Longao Huang, Hongyuan Xu, Hua Jiang BMC Musculoskeletal Disorders.2025;[Epub] CrossRef
Introducing the Modified Cobb Angle for Measuring Cervical Spine Curvature During Swallowing Catriona M. Steele, Makaya O' Grady, Sima Farpour, Melanie Peladeau‐Pigeon, Sophia Werden Abrams, Maureen Folsom, Emily K. Plowman, Ashwini M. Namasivayam‐MacDonald Laryngoscope Investigative Otolaryngology.2025;[Epub] CrossRef
The Significance of Odontoid Incidence in Patients With Cervical Spondylotic Myelopathy Hongyu Qin, Weiyou Chen, Longao Huang, Xin Xiao, Qinghua Yang, Hua Jiang Global Spine Journal.2024; 14(8): 2374. CrossRef
Variation of Odontoid Incidence According to Age in Asymptomatic Children Yuwang Du, Weiyou Chen, Hongyuan Xu, Xiao Liang, Chongyang Wang, Hongyu Qin, Hua Jiang World Neurosurgery.2024; 190: e496. CrossRef
Surgical and Clinical Outcomes Associated With the Use of Barbed Sutures and Self-Adhering Mesh System and Polymeric Glue for Wound Closure in Multilevel or Revision Spinal Surgery: A Matched Cohort Comparative Study With Conventional Wound Closure Proced Junho Mun, Seung-Jae Hyun, Jae-Koo Lee, Sungjae An, Ki-Jeong Kim Neurospine.2023; 20(3): 981. CrossRef
Selection of Optimal Lower Instrumented Vertebra for Adolescent Idiopathic Scoliosis Surgery Seung-Ho Seo, Seung-Jae Hyun, Jae-Koo Lee, Yong Jae Cho, Dae Jean Jo, Jin Hoon Park, Ki-Jeong Kim Neurospine.2023; 20(3): 799. CrossRef
Optimizing Surgical Strategy for Cervical Spinal Deformity: Global Alignment and Surgical Targets Jae-Koo Lee, Seung-Jae Hyun, Ki-Jeong Kim Neurospine.2023; 20(4): 1246. CrossRef
Commentary on “Clinical Impact and Correlations of Odontoid Parameters Following Multilevel Posterior Cervical Fusion Surgery” Sang Hun Lee Neurospine.2022; 19(4): 921. CrossRef
Significance of Atlantoaxial and Subaxial Spinal Instability in Cervical Spinal Spondylosis: Commentary on “Clinical Impact and Correlations of Odontoid Parameters Following Multilevel Posterior Cervical Fusion Surgery” Atul Goel Neurospine.2022; 19(4): 924. CrossRef
Objective This retrospective cohort study seeks to identify risk factors associated with complications following posterior cervical laminectomy and fusion (PCLF) surgery.
Methods Adults undergoing PCLF from 2012 through 2018 at a single center were identified. Demographic and radiographic data, surgical characteristics, and complication rates were compared. Multivariate logistic regression models identified independent predictors of complications following surgery.
Results A total of 196 patients met the inclusion criteria and were included in the study. The medical, surgical, and overall complication rates were 10.2%, 23.0%, and 29.1% respectively. Risk factors associated with medical complications in multivariate analysis included impaired ambulation status (odds ratio [OR], 2.27; p=0.02) and estimated blood loss over 500 mL (OR, 3.67; p=0.02). Multivariate analysis revealed preoperative narcotic use (OR, 2.43; p=0.02) and operative time (OR, 1.005; p=0.03) as risk factors for surgical complication, whereas antidepressant use was a protective factor (OR, 0.21; p=0.01). Overall complication was associated with preoperative narcotic use (OR, 1.97; p=0.04) and higher intraoperative blood loss (OR, 1.0007; p=0.03).
Conclusion Preoperative narcotic use and estimated blood loss predicted the incidence of complications following PCLF for CSM. Ambulation status was a significant predictor of the development of a medical complication specifically. These results may help surgeons in counseling patients who may be at increased risk of complication following surgery.
Citations
Citations to this article as recorded by
Risk factors of significant intraoperative blood loss in unilateral expansive open-door cervical laminoplasty for cervical compressive myelopathy: A retrospective observational study Hailong Yu, Hong Yuan, Zhihao Zhang, Xinyuan Zhao, Bin Zheng, Xiaoyu Wang, Yin Hu, Hongwei Wang Medicine.2026; 105(20): e48869. CrossRef
Minimally Invasive Surgical Technique through a Natural Anatomical Corridor for C1-C2 Screw Fixation Andres Mendez-Gutierrez, Mariana Agudelo-Arrieta, Miguel Enrique Berbeo, Roberto Diaz-Orduz Journal of Minimally Invasive Spine Surgery and Technique.2023; 8(1): 55. CrossRef
Risk Factors for Significant Intraoperative Blood Loss during Anterior Cervical Decompression and Fusion for Degenerative Cervical Diseases Hong Yuan, Yuanhang Zhao, Yin Hu, Zhonghua Liu, Yu Chen, Hongwei Wang, Hailong Yu, Liangbi Xiang Orthopaedic Surgery.2023; 15(11): 2822. CrossRef
Laminectomy alone versus laminectomy with lateral mass screw fixation in the treatment of multisegment cervical spinal canal stenosis: a comparative analysis Hany Elkholy, Mohamed Ahmed El Tabl, Osama Saber El Sherif Egyptian Journal of Neurosurgery.2023;[Epub] CrossRef
Complications in posterior cervical spine surgery Julian G. Lugo-Pico, John G. Heller Seminars in Spine Surgery.2022; 34(1): 100921. CrossRef
Smoking Is an Independent Risk Factor for 90-Day Readmission and Reoperation Following Posterior Cervical Decompression and Fusion Ryan K Badiee, Andrew K Chan, Joshua Rivera, Annette Molinaro, Dean Chou, Praveen V Mummaneni, Lee A Tan Neurosurgery.2021; 88(6): 1088. CrossRef
Benefits of Early Ambulation in Elderly Patients Undergoing Lumbar Decompression and Fusion Surgery: A Prospective Cohort Study Jie Huang, Zhan Shi, Fang‐fang Duan, Ming‐xing Fan, Shuo Yan, Yi Wei, Bing Han, Xue‐mei Lu, Wei Tian Orthopaedic Surgery.2021; 13(4): 1319. CrossRef
First
Prev
