Haiying Liu

Objective: Postoperative loss of cervical lordosis remains a common and clinically relevant complication following laminoplasty, negatively affecting neck pain, neurological recovery, and long‐term sagittal balance. However, reliable and easily applicable preoperative predictors for identifying patients at high risk of cervical lordosis deterioration remain limited. This study aims to investigate whether preoperative C2 slope (C2S) independently predicts cervical lordosis deterioration following laminoplasty. Methods: This retrospective cohort study included 179 patients who underwent cervical laminoplasty for cervical spondylotic myelopathy at our institution between April 2014 and December 2020, with a minimum follow‐up of 24 months. Radiological parameters including C2‐7 Cobb angle, cervical sagittal vertical axis (cSVA), C7 slope, and C2S are measured preoperatively and at final follow‐up. Patients are divided into lordosis deterioration group (> 5°) and control group (≤ 5°). B Between‐group comparisons are performed using independent‐samples t tests and χ ² tests. Binary logistic regression analysis is conducted to identify independent predictors of postoperative cervical lordosis loss. Receiver operating characteristic (ROC) curve analysis is used to evaluate predictive performance and determine the optimal cutoff value. Results: The lordosis loss group ( n = 55) shows significantly higher preoperative C2S (14.57° ± 3.47° vs. 9.52° ± 7.30°, p < 0.001), lower preoperative Cobb angle (13.01° ± 4.91° vs. 16.1° ± 6.50°, p < 0.001), and greater cSVA (2.58 ± 1.45 cm vs. 2.13 ± 1.42 cm, p = 0.027) compared to controls. The lordosis loss group demonstrates worse postoperative neck pain VAS scores (3.31 ± 1.63 vs. 2.40 ± 1.56, p < 0.001) and slightly lower JOA scores (15.45 ± 1.14 vs. 15.78 ± 1.12, p = 0.037). Multivariate analysis reveals preoperative C2S as the only independent predictor (OR = 1.176, p < 0.001), with 11.49 as cutoff value for C2S. Conclusion: Elevated preoperative C2S independently predicts postoperative cervical lordosis deterioration. C2S measurement provides a simple, effective tool for identifying high‐risk patients and optimizing surgical planning.

Axial symptoms after cervical laminoplasty significantly impact patient quality of life, with incidence rates of 20–40%. The relationship between cervical facet joint degeneration and postoperative axial symptoms remains unclear. This retrospective study analyzed 189 patients who underwent cervical laminoplasty from 2012 to 2022. Patients are divided into axial symptom (n = 69) and non-axial symptom (n = 120) groups based on persistent neck/shoulder pain (VAS ≥ 4) lasting ≥ 3 months postoperatively. Cervical facet joint degeneration is assessed using a 4-grade CT-based scoring system from C2/C3 to C6/C7. Total facet degeneration score and C4-C7 segmental score are calculated and compared between groups. No significant differences exist in upper cervical segments (C2/C3, C3/C4) between groups. However, the axial symptom group shows significantly higher degeneration scores in lower cervical segments: C4/C5 (1.92 ± 1.20 vs. 1.44 ± 0.71, P < 0.001), C5/C6 (2.16 ± 0.86 vs. 1.52 ± 0.84, P < 0.001), and C6/C7 (2.12 ± 0.91 vs. 1.57 ± 0.84, P < 0.001). Multivariate analysis reveal that both total degeneration score (OR = 1.164, P < 0.001) and C4-C7 score (OR = 1.269, P < 0.001) are independent predictors of axial symptoms. ROC analysis shows C4-C7 score has superior predictive accuracy (AUC = 0.741) compared to total score (AUC = 0.694). Lower cervical facet joint degeneration, particularly C4-C7 segments, is a strong independent predictor of postoperative axial symptoms, providing valuable guidance for preoperative risk assessment and surgical planning.

Background: Sagittal balance deterioration is major concern after cervical laminoplasty, which affects neurological recovery and quality of life. Preoperative bone quality may play a key role. CT-derived Hounsfield Unit (HU) values reflect bone mineral density, while MRI-derived vertebral bone quality (VBQ) values indicate bone marrow fat content. Their combined predictive value remains unclear. Methods: This study retrospectively analyzes 104 patients (mean age 58.4 ± 11.6 years) who undergo posterior single-door laminoplasty with preoperative cervical CT and MRI. Mean HU and VBQ values (C2–T1) are measured, and patients are classified by postoperative change in cervical sagittal vertical axis (ΔcSVA). Logistic regression and ROC analyses are used to evaluate predictive performance. Results: HU and VBQ differ significantly among improvement, stable, and deterioration groups ( P < 0.001). Both HU (−0.007, P = 0.009) and VBQ (2.521, P = 0.018) independently predict postoperative imbalance. VBQ correlates negatively with HU ( r = −0.306, P = 0.0016). ROC analysis shows AUCs of 0.690 for HU, 0.747 for VBQ, and 0.791 for the combined VBQ + HU model, with improved sensitivity (0.838) and specificity (0.597). Conclusion: Decreased HU and elevated VBQ independently predict postoperative sagittal balance deterioration, reflecting complementary aspects of vertebral bone quality. The combined MRI- and CT-based assessment improves prediction accuracy and assists in preoperative risk stratification and individualized surgical planning.

This study explores the use of radiomic features extracted from preoperative T2-weighted MRI and CT images, combined with machine learning models, to predict the risk of vertebral refracture after percutaneous kyphoplasty (PKP) in postmenopausal women. We retrospectively collect data from 156 postmenopausal women with osteoporotic vertebral compression fractures (OVCFs) who underwent PKP (35 refracture cases, 121 non-refracture controls). All patients had preoperative lumbar T2-weighted MRI and CT scans. We extract MRI and CT radiomic features and constructed radiomic signatures through feature selection. Key clinical factors (age, body mass index [BMI], vertebral CT Hounsfield unit [HU] values, smoking history, diabetes history, alcohol use, etc.) are used to build clinical prediction models. Various machine learning classifiers (Support Vector Machine [SVM], K-Nearest Neighbors [KNN], Random Forest [RF], ExtraTrees, XGBoost, LightGBM, Multi-layer Perceptron [MLP]) are trained on the radiomic signatures and clinical factors. Model performance was evaluated on an independent test set using area under the ROC curve (AUC) as the primary metric. Accuracy, sensitivity, specificity, and other measures on the test set were compared between radiomic models, clinical models, and a combined model. The refracture group (n = 35, 22.4%) is significantly older (72.09 ± 4.25 vs 70.11 ± 3.31 years, P = 0.002) with lower vertebral bone density (97.00 ± 6.31 vs 102.49 ± 4.68 HU, P < 0.001). Among individual algorithms, the KNN clinical model achieves optimal performance (AUC = 0.74), while the SVM radiomics model demonstrates the best accuracy (AUC = 0.798, accuracy = 0.839, sensitivity = 0.857, specificity = 0.833). The combined model achieves superior performance (AUC = 0.886), significantly outperforming both standalone models. Multi-modal radiomics combined with key clinical factors provides superior prediction of refracture risk after PKP. This approach offers clinicians an objective tool for individualized risk stratification, representing a meaningful step toward precision medicine in managing osteoporotic fractures.

Background: Thoracic spinal stenosis (TSS) is a rare yet debilitating condition, often requiring surgical decompression. Prognostic assessments traditionally rely on single clinical or imaging features, limiting prediction accuracy. This study explores whether radiomics-based models enhance outcome prediction in TSS. Methods: We retrospectively enrolled 106 surgically treated TSS patients (2012–2022), collecting clinical data and T2 axial MRI scans. Radiomics features were extracted from the most stenotic level, followed by rigorous feature selection (ICC > 0.9, U-test, Spearman, mRMR, and LASSO). Six machine learning classifiers were trained using radiomics and/or clinical data. Model performance was evaluated using AUC on an independent test set. Results: Radiomics models outperformed clinical models (SVM AUC: 0.824 vs. 0.731). The combined radiomics–clinical model achieved the highest test-set AUC of 0.867, offering improved sensitivity and specificity. Conclusion: In this preliminary exploratory study, integrating MRI radiomics with clinical data appeared to improve prediction of neurological recovery in TSS. These findings suggest that radiomics may enable objective, high-dimensional assessment of spinal cord pathology and potentially support individualized surgical decision-making, although further validation in larger, multicenter prospective cohorts is required.

To develop machine-learning model that combines pre-operative vertebral-body CT radiomics with clinical data to predict cage subsidence after single-level ACDF with Zero-P. We retrospectively review 253 patients (2016–2023). Subsidence is defined as ≥ 3 mm loss of fused-segment height at final follow-up. Patients are split 8:2 into a training set (n = 202; 39 subsidence) and an independent test set (n = 51; 14 subsidence). Vertebral bodies adjacent to the target level are segmented on pre-operative CT, and high-throughput radiomic features are extracted with PyRadiomics. Features are z-score–normalized, then reduced by variance, correlation and LASSO. Age, vertebral Hounsfield units (HU) and T1-slope entered a clinical model. Eight classifiers are tuned by cross-validation; performance is assessed by AUC and related metrics, with thresholds optimized on the training cohort. Subsidence patients are older, lower HU and higher T1-slope (all P < 0.05). LASSO retained 11 radiomic features. In the independent test set, the clinical model had limited discrimination (AUC 0.595). The radiomics model improved performance (AUC 0.775; sensitivity 100%; specificity 60%). The combined model is best (AUC 0.813; sensitivity 80%; specificity 80%) and surpassed both single-source models (P < 0.05). A pre-operative model integrating CT-based radiomic signatures with key clinical variables predicts cage subsidence after ACDF with good accuracy. This tool may facilitate individualized risk stratification and guide strategies—such as endplate protection, implant choice and bone-quality optimization—to mitigate subsidence risk. Multicentre prospective validation is warranted.

Background: This bibliometric analysis aimed to map the knowledge network of artificial intelligence in scoliosisMethods: Studies on artificial intelligence published from January 2003 to December 2024 are retrieved from Web of Science Core Collection (WoSCC). The contributions of countries, institutions, authors, and journals are identified using VOSviewer, Online Analysis Platform of Literature Metrology (http://biblimetric.com) and Microsoft Excel. Tendencies, hotspots and knowledge networks are analyzed and visualized using VOS-viewer and CiteSpace.Results: 718 publications are included in the final analysis. The leading country in this field is China. Royal Hospital for Sick Children featured the highest number of publications among all institutions and National University of Singapore featured the highest citations of publications. Co-citation cluster labels revealed characteristics of three main clusters: (1) Image process and classification of scoliosis, (2) AI application in surgical treatment of scoliosis, (3) predict postoperative complications and scoliosis development. Keyword burst detection indicated that machine learning and deep learning are the newly emerging research hot spots.Conclusion: This study compiled 718 publications covering AI in scoliosis and showed that the direction of these studies is likely in transition from cerebral palsy to machine learning and deep learning. It provides guidance for further research and clinical applications on AI application in scoliosis.

Background Tranexamic acid (TXA) is widely used to reduce blood loss in spine surgery. Different administration routes, including intravenous (IV), topical, and oral, have been studied, but a comprehensive comparison of their efficacy and safety remains lacking. This network meta-analysis aims to compare the effectiveness and safety of various TXA administration methods in spine surgery, focusing on blood loss, transfusion rates and other perioperative parameters. Methods We systematically reviewed randomized controlled trials (RCTs) published until January 2025, comparing different TXA regimens in spine surgery. Data on blood loss, transfusion rates, postoperative complications, post-drainage, length of hospital stay and operative time are extracted. A network meta-analysis is conducted using a Bayesian random effects model to rank the efficacy and safety of TXA routes. Results A total of 12 RCTs involving 1980 patients are included. Topical TXA + IV TXA demonstrated the highest efficacy in reducing intraoperative blood loss, post-operative blood loss, overall blood loss, Hct, Hgb, transfusion rates, post-drainage, operative time and hospital stay, with the highest probability of being the best treatment. Oral TXA showed fewer complications, making it a safer option for low-risk patients. Subgroup analysis confirmed the superiority of topical TXA + IV TXA in spine fusion surgeries, particularly for overall blood loss control and reducing transfusion requirements. Conclusion For open thoracolumbar fusion surgery, Topical TXA + IV TXA is the most effective regimen for reducing blood loss and enhancing recovery in high-risk surgeries. Oral TXA offers a safer alternative for low-risk patients. Extrapolating these findings to other spine procedures still requires further randomized controlled trials. Tailoring TXA administration to individual patient needs optimizes outcomes and reduces complications.

Background: This study systematically reviews the literature and performs a meta-analysis to evaluate and compare the intra-operative outcomes, clinical efficacy, safety, and cost of alternative-level and all-level laminoplasty.Methods: A systematic review is conducted according to the PRISMA guidelines. Searches are performed in PubMed, Cochrane Library, OVID, and Embase databases from inception to August 2024, using search terms “laminoplasty” AND “all OR skip OR alternative.” Data extraction and risk-of-bias assessment are conducted independently by two researchers using the Newcastle-Ottawa Scale. Statistical analysis is performed with RevMan 5.4.Results: Four retrospective Chinese studies (337 patients: 176 alternative-level, 161 all-level) meet the criteria. Meta-analysis shows no significant difference in intra-operative outcomes: operative time (P = 0.23) and blood loss (P = 0.11). Clinical efficacy, measured by Japanese Orthopaedic Association(JOA) Score (P = 0.08), JOA recovery(P = 0.08), and Visual Analog Scale (P = 0.26), also shows no significant difference. Similarly, safety outcomes, including complications(P = 0.64), C5 palsy(P = 1.00), and axial symptoms(P = 0.57), are comparable between the two fixation methods. Cervical sagittal parameters are also equivalent: Cervical Curvature Index (P = 0.18) and overall range of motion (P = 0.29). However, alternative-level laminoplasty demonstrates lower cost (P < 0.00001) and is inferior in cervical canal outcomes, including anterior–posterior diameter (P = 0.01), Pavlov ratio(P = 0.007) and open angle (P < 0.00001).Conclusion: Alternative-level laminoplasty matches all-level fixation in operative efficiency, neurological recovery, and complication rates while substantially reducing implant costs. Its slightly lesser canal expansion does not translate into inferior clinical outcomes. Evidence strength is limited by the small number of single-center retrospective studies from one country. Multi-center randomized trials in other countries are needed to confirm generalizability.