2026-09-30 2026, Volume 2 Issue 3

  • Select all
  • Review
    Xudong Zhao, Keyi Chen, Jinquan Ma, Bicheng Wang, Chen Xu, Bo Hu, Huajiang Chen

    Spinal cord injury (SCI) is a severe neurological disorder that often results in long-term motor and sensory dysfunction, leaving patients highly dependent on external care, while current rehabilitation approaches remain limited in their efficacy. In recent years, brain–computer interfaces (BCIs), as a multidisciplinary frontier technology, have demonstrated unique advantages in SCI rehabilitation. By bypassing the damaged spinal pathways and establishing a direct communication channel between the brain and external devices, BCIs enable motor recovery, sensory restoration, and induction of neuroplasticity, while also showing potential in alleviating chronic neuropathic pain and improving psychological well-being. A growing body of experimental and clinical studies has confirmed that BCIs can restore partial motor control and tactile perception, thereby enhancing rehabilitation efficiency and improving patients' quality of life. However, their clinical translation remains constrained by challenges such as signal stability, long-term electrode safety, adaptive decoding algorithms, and ethical concerns. With the continuous development of artificial intelligence, novel biomaterials, and immersive technologies such as virtual reality, BCIs are expected to evolve toward more personalized, home-based, and intelligent rehabilitation solutions, accelerating their clinical application and offering new therapeutic hope for SCI patients.

  • Review
    Haoming Sun, Chunwang Jia, Hongwei Wang, Xianghe Wang, Haocheng Xu, Xiaosheng Ma, Feizhou Lyu, Jianyuan Jiang, Shuo Yang, Hongli Wang

    Spinal infectious diseases can lead to severe disability and adverse outcomes without timely and accurate diagnosis and treatment. Artificial intelligence (AI) and deep learning (DL) have brought new progress to imaging-based diagnosis of spinal infectious diseases. This review aimed to evaluate the diagnostic and differential diagnostic performance of AI models for spinal infectious diseases. A systematic literature search was performed in PubMed, Web of Science, and Embase for studies published between 2018 and 2026. Eligible studies focused on AI or DL models for detecting or differentiating spinal infectious diseases, and data including sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve were extracted and analyzed. A total of eight studies were included. AI models showed generally favorable efficacy in the identification and differential diagnosis of spinal infections. For differentiating tuberculous spondylitis (TS) from brucellar spondylitis (BS), sensitivity ranged from 0.588 to 0.940 and specificity from 0.679 to 0.964. Most models were based on magnetic resonance imaging (MRI) or computed tomography (CT) and achieved high diagnostic accuracy, but the majority were retrospective, single-centered, and lacked large-scale external validation. AI serves as a reliable auxiliary tool for the rapid and accurate diagnosis of spinal infectious diseases, especially when combined with advanced imaging techniques. However, further multicenter, prospective validation is required before widespread clinical application. AI should complement rather than replace clinical expertise in routine diagnostic workflows.

  • Review
    Zhidi Lin, Kaiwen Chen, Chi Sun, Shuo Yang, Fei Zou, Xiaosheng Ma, Hongli Wang, Jianyuan Jiang

    The standardized reporting of postoperative complications is essential for improving surgical quality and comparing outcomes. However, spinal surgery has historically lacked a universal system and has instead relied on ambiguous terms. This narrative review synthesizes the literature on the adaptation, validation, and application of the therapy-based Clavien-Dindo classification (CDC) and its modified versions in spinal surgery. It focuses on studies across various spinal procedures and patient populations. The evidence shows that adapted CDC systems demonstrate good to excellent inter- and intrarater reliability, particularly for severe complications. These systems also facilitate direct comparison of surgical techniques. Moreover, they reveal strong correlations between complication severity, prolonged hospital stay, and patient factors such as frailty. Nevertheless, key limitations include poor correlation with certain long-term patient-reported outcomes. The CDC cannot capture intraoperative events or the cumulative burden of multiple complications. Subjectivity also exists in grading milder events. Recent spine-specific modifications that improve neurologic deficit assessment have enhanced clinical relevance. Overall, the CDC provides a critical framework for standardizing complication reporting in spinal surgery. Optimal application requires awareness of its limitations. In complex scenarios, the CDC may be applied most effectively as a component within more comprehensive, spine-specific taxonomies that incorporate surgical complexity and structured neurological assessment. The CDC remains a vital tool for improving communication, benchmarking, and ultimately enhancing patient care.

  • Original Research
    Yunzhong Cheng, Jixuan Huang, Lijin Zhou, Yong Hai

    Background: This study systematically present the overall landscape and evolutionary trends of international research in the field of scoliosis rehabilitation over the past 30 years.

    Methods: This study used the Science Citation Index Expanded (SCI-E) database of the Web of Science Core Collection as the data source. Target literature was screened according to pre-set criteria, and bibliometric analysis platforms, VOSviewer 1.6.11 and CiteSpace software 6.1.R6 were employed to conduct a quantitative analysis and visual presentation of annual publication volume, citation frequency, core authors, institutions, countries, research directions, funding sources, and journal distribution.

    Results: The results showed that the United States and China were the major contributing countries in this field. A total of 308 valid documents were included, with 4,333 total citations, 14.07 citations per paper on average, and an H-index of 35. Negrini S, Italian Spine Science Institute, the United States, Orthopedics, National Natural Science Foundation of China, and Spine ranked first in terms of publication output, citations, core research directions, funding, and journals, respectively.

    Conclusions: Keyword clustering analysis indicated that adolescent idiopathic scoliosis, Society on Scoliosis Orthopaedic and Rehabilitation Treatment (SOSORT) guidelines, and Asian adolescent idiopathic scoliosis were core topics in the field, representing future research hotspots. Meanwhile, the burst intensity of keywords such as Cobb angle, posterior spinal fusion, and risk factors increased significantly, suggesting that surgical treatment and its safety are becoming frontier research trends.

  • Original Research
    Chuncheung Chan, Xiaojie Liu, Dacheng He, Hongfei Xiang, Jiyu Li, Fuxin Wei

    Background: Lumbar spondylolysis is commonly evaluated initially with lumbar radiography, but radiographs may miss pars interarticularis defects, particularly at L5. This study aimed to characterize radiographic and morphological features associated with bilateral L5 spondylolysis and to develop and externally evaluate an X-ray-based screening nomogram for radiographic risk stratification.

    Methods: This cross-sectional study included patients who underwent lumbar radiography and computed tomography (CT) between 2018 and 2025. Patients with CT-confirmed bilateral L5 spondylolysis were compared with controls without spondylolysis. Radiographic parameters and CT-based morphological features were measured and analyzed separately, and interobserver reliability was assessed. Propensity score matching (PSM) (1:1) was used to balance baseline characteristics. Factors associated with L5 spondylolysis were identified using multivariable logistic regression, and an X-ray-only nomogram was constructed using significant radiographic variables. Model performance in the development cohort was assessed using discrimination, calibration, decision curve analysis (DCA), and clinical impact curves (CICs) with internal bootstrap validation. External validation applied the original development-cohort coefficients without refitting and primarily assessed discrimination in an independent dataset.

    Results: After PSM, 114 patients were included in the development cohort, with 57 patients in each group. Compared with controls, patients with L5 spondylolysis had larger sacral slope (SS), L5 lamina inclination angle (L5 LIA), and pedicle internal inclination angle, as well as smaller sacral inclination (SI) and L5 lumbar index. Multivariable analysis identified smaller SI, smaller L5 lumbar index, larger SS, larger L5 LIA, and larger pedicle internal inclination angle as factors independently associated with L5 spondylolysis. L5 LIA showed the strongest individual diagnostic performance (area under the curve [AUC] = 0.879; 95% CI, 0.813–0.944). The X-ray screening model incorporating the L5 lumbar index, SI, SS, and L5 LIA achieved excellent discrimination in the development cohort (AUC = 0.916; bootstrap-corrected concordance index = 0.911) and maintained strong discrimination in the external validation dataset when the development-cohort coefficients were applied without refitting (AUC/C-index = 0.949; 95% CI, 0.922–0.977).

    Conclusions: An X-ray-based nomogram incorporating key sagittal alignment and morphological parameters showed strong diagnostic discrimination for bilateral L5 spondylolysis in both the development cohort and the external validation dataset. This model should be interpreted as a radiographic screening and risk-stratification aid rather than a tool for prospective prognosis. CT-based analysis also showed that patients with L5 spondylolysis had a larger pedicle internal inclination angle, a morphological difference with potential relevance to preoperative planning.

  • Original Research
    Wenbo Wu, Hanmo Fang, Kun Wang, Xiaobo Feng, Xianlin Zeng, Cao Yang, Lin Xie

    Background: This study investigates the feasibility and effectiveness of a fusion cage with ipsilateral-impaction bone grafting (IIBG) for treating lumbar spinal stenosis (LSS) combined with segmental instability.

    Methods: From January 2018 to December 2021, 282 LSS patients undergoing minimally invasive transforaminal lumbar interbody fusion (MIS-TLIF) at Union Hospital were enrolled (164 IIBG vs. 118 controls). The cohort comprised 97 males and 185 females (mean age 62.75 ± 6.88 years; body mass index [BMI] 24.02 ± 2.58 kg/m2). Surgical levels included 218 single-level (L3–L4: 5, L4–L5: 113, L5–S1: 100), 42 double-level (L3–L5), and 22 multi-level (L3–S1) fusions. Visual analogue scale (VAS) and Oswestry disability index (ODI) were used to assess the severity of pain and spinal nerve function before surgery, as well as at one month, six months, and 12 months after surgery. Besides, the MacNab criteria were used to evaluate clinical efficacy at the last follow-up. Imaging indicators refer to the preoperative and postoperative ratio of intervertebral height and bone grafting area to vertebral body area. Fusion status was evaluated via the Bridwell Grade on 1-year computed tomography (CT).

    Results: The mean operative time was 94.11 ± 38.11 min (range 68–238 min), with an intraoperative blood loss of 90.82 ± 40.35 mL (range 20–150 mL). Follow-up visits were conducted among all 282 patients, with an average follow-up time of 22.44 ± 6.84 months (range 12–34 months). Preoperative VAS of lower back pain and lower limb pain was 6.99 ± 1.42 points and 7.80 ± 1.33 points, respectively, which showed statistical differences (F = 2,290.00, p < 0.001; F = 3,307.17, p < 0.001) compared with postoperative results. The scores decreased to 2.20 ± 0.85 and 2.03 ± 0.84 at one month after surgery, 1.56 ± 0.60 and 1.51 ± 0.50 at six months after surgery, 0.89 ± 0.69 and 0.89 ± 0.69 at one year after surgery, and 0.85 ± 0.70 and 0.87 ± 0.69 at the last follow-up. ODI was 54.47 ± 2.89% (range 50%–59%) before surgery, 25.50 ± 2.88% (range 21%–30%) at one month after surgery, 15.02 ± 3.14% (range 10%–20%) at six months after surgery, 12.50 ± 1.71% (range 10%–15%) at one year after surgery, and 12.51 ± 1.71% (range 50%–59%) at the last follow-up; the difference was significant (F = 734.432, p < 0.001). MacNab criteria indicated excellent outcomes in 209 cases, good outcomes in 58 cases, and fair outcomes in 15 cases, with excellent and good results of 92.74% among all 282 patients. Mean intervertebral height was 5.72 ± 1.39 mm (range 3.14–8.34 mm) before surgery, 13.21 ± 0.43 mm (range 12.46–13.98 mm) after surgery, and 11.62 ± 0.22 mm (range 11.10–11.99 mm) at the last follow-up; the difference was significant (F = 6,082.37, p < 0.001). Postoperative CT revealed that the bone grafting area (including cage) at the ipsilateral level accounted for 33.63 ± 3.42% (range 32.56%–39.02%) of the vertebral endplate area in the IIBG group, which was significantly different (T = 31.35, p < 0.001) from the control group (22.83 ± 1.79%, range 18.57%–26.30%). CT at one year after surgery pointed out that significantly more patients in the IIBG group (91.5%, range 87.3%–95.7%) achieved successful fusion (Grade I or II in Bridwell) than the control group (76.53%, range 69.1%–83.9%), with a K value of 12.50 and a p value less than 0.001. Furthermore, a 3-year postoperative CT follow-up in a subset of patients consistently demonstrated more robust and extensive bridging bone formation across the entire disc space in the IIBG group, in contrast to the control group, where bridging bone was primarily confined within the cage windows. Besides, eight cases (4.8%) and 11 cases (9.3%) in the two groups appeared to have cage subsidence, but the difference was insignificant (K = 2.03, p = 0.154). Three patients (1.0%) developed bone graft particles breaking through the anterior longitudinal ligament, but this did not cause vascular injury. Complications included CSF leakage (0.7%), pedicle screw misplacement (0.7%), and screw revision (0.3%).

    Conclusions: A fusion cage with ipsilateral impaction bone grafting ensures not only a large bone graft area outside the cage but also a shorter bone bridging distance between endplates and denser impaction of bone graft particles. Thus, this technique appears to be a safer impaction bone grafting method that provides a better environment for intervertebral fusion.

  • Original Research
    Mufeng Li, Quezhu Danzeng, Yi Shen, Jiang Yi, Jie Meng, Zhihui Qian, Minfei Wu, Weibo Jiang

    Background: Ossification of the nuchal ligament (ONL) is a radiographic finding frequently observed in patients with degenerative cervical myelopathy (DCM). Although its etiology remains unclear, posterior cervical anatomical factors, including paraspinal muscle degeneration and spinous process length parameters, may alter the local mechanical environment of the nuchal ligament and contribute to ONL formation. This study aimed to evaluate whether cervical paraspinal muscle fatty infiltration (FI) and C6 spinous process length are associated with ONL presence, exploratory early-stage ONL, and ONL length in patients with DCM.

    Methods: This single-center retrospective cohort study included 152 patients who underwent surgery for DCM between 2022 and 2024. Patients were classified into ONL and non-ONL groups using sagittal computed tomography images. ONL length was analyzed as a continuous radiographic variable. Because no validated ONL grading system is currently available, the ONL-positive cohort was additionally divided by median ONL length into shorter- and longer-length subgroups; the shorter-length subgroup was considered an exploratory morphological surrogate for early-stage ONL rather than a validated clinical stage. Paraspinal muscle FI was assessed using the Goutallier grading system on T2-weighted magnetic resonance imaging at C4–C5, C5–C6, and C6–C7 levels. Spinous process length parameters included absolute C5, C6, and C7 spinous process lengths and C5:C6 and C6:C7 ratios. Multivariable regression, variance inflation factor analysis, and receiver operating characteristic (ROC) analysis were performed.

    Results: Among 152 patients, 96 had ONL and 56 did not. Male sex and smoking status were more common in the ONL group. The ONL group showed higher C4–C5 and C5–C6 multifidus FI grades, greater absolute C5, C6, and C7 spinous process lengths, and a higher C6:C7 ratio. Multivariable logistic regression showed that the C6:C7 ratio was independently associated with ONL presence (p = 0.012, OR = 1.051, 95% CI = 1.011–1.092). In the exploratory early-stage ONL model, C4–C5 multifidus FI and the C6:C7 ratio were associated with shorter-length ONL. ROC analyses showed modest discriminatory performance, with area under the curve values ranging from 0.658 to 0.682. Exploratory continuous ONL length analysis showed associations with age, C4–C5 multifidus FI, and the C6:C7 ratio in the adjusted model.

    Conclusions: Relative C6 spinous process length was associated with ONL presence, while C4–C5 multifidus FI and relative C6 spinous process length were associated with exploratory early-stage ONL in patients with DCM. These parameters may serve as auxiliary radiographic markers and provide a hypothesis-generating framework for understanding the anatomical, biomechanical, and mechanobiological background of ONL formation.

  • Erratum
    Chunfeng Shang, Longyu Li, Songfeng Chen, Hongwei Kou, Guowei Shang, Zikuan Leng, Menghan Wang, Hongjian Liu