Xiang Liao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xiang Liao has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
Zhouyang Hu
National Key Clinical Pain Medicine of China, Shenzhen Nanshan People’s Hospital, Shenzhen University Medical School, Shenzhen, 518052, People’s Republic of China
7 shared publications- Guoxin Fan
Department of Pain Medicine, Shenzhen Nanshan District People's Hospital;
5 shared publications Jianjin Zhang
National Key Clinical Pain Medicine of China, Shenzhen Nanshan People’s Hospital, Shenzhen University Medical School, Shenzhen, 518052, People’s Republic of China
5 shared publicationsGuoxin Fan
4 shared publicationsZhipeng Xu
Department of Pain Medicine, Shenzhen Nanshan People’s Hospital, Shenzhen University Medical School, Shenzhen, China
4 shared publicationsHong Wang
Department of Pain Medicine, Shenzhen Nanshan People’s Hospital, Shenzhen University Medical School, Shenzhen, China
3 shared publications
Latest Publications
- Open Access
- 19 September 2026
- Published by Elsevier in Asian Journal of Surgery
Automated multi-class segmentation of anatomical structures in percutaneous endoscopic cervical foraminotomy using deep learning
- 19 September 2026
- Published by Elsevier in Asian Journal of Surgery
Open AccessAutomated multi-class segmentation of anatomical structures in percutaneous endoscopic cervical foraminotomy using deep learning
- Open Access
- 1 May 2026
- Published by Taylor & Francis in Journal of Pain Research
- Vol. ume 19, 1-13
Dissecting the Causal Pathway from Herpes Zoster to Postherpetic Neuralgia: A Multi-Stage Mendelian Randomization Study Implicating VZV-Specific IgE and Cerebrospinal Fluid Metabolites
Background: Postherpetic neuralgia (PHN) develops in 10– 20% of herpes zoster (HZ) patients, yet the causal mechanisms driving this transition from acute infection to chronic neuropathic pain remain elusive. VZV-specific immunoglobulin E (IgE) and cerebrospinal fluid (CSF) metabolites represent candidate intermediate phenotypes that may mediate neuroimmune dysfunction, but their sequential causal relationships remain unexplored. We hypothesized that IgE and CSF metabolites constitute a hierarchical causal pathway from HZ exposure to PHN risk. Methods: We conducted a multi-stage Mendelian randomization study. First, two-sample Mendelian Randomization (MR) with MRlap correction established causal effects of HZ on IgE and screened 435 CSF metabolites for PHN associations. Given substantial HZ-PHN sample overlap precluding total effect estimation, multivariable MR (MVMR) evaluated whether metabolites mediate IgE effects on PHN. Results: HZ exhibited a significant causal effect on VZV-specific IgE (IVW: β = 0.237, 95% CI: 0.104– 0.371, P = 4.95× 10− 4). Among 435 metabolites, two showed effects significant after FDR correction: 1-stearoyl-2-oleoyl-GPC (β = 2.448, FDR-adjusted P = 0.043) and N-methylproline (β = − 0.425, FDR-adjusted P = 0.046). Critically, 5-methyluridine—nominally significant in univariable analysis (P = 0.095)—demonstrated a robust independent protective effect in MVMR (MV-IVW: β = − 3.227, 95% CI: − 5.765 to − 0.688, FDR-adjusted P = 0.025; OR = 0.040). The IgE→PHN effect attenuated to non-significance (P = 0.202) upon adjusting for 5-methyluridine, consistent with metabolic mediation. Conclusion: These findings support a sequential causal pathway from HZ to PHN involving IgE and CSF metabolites, identifying 5-methyluridine as a potential modifiable protective factor and highlighting neuroimmune-metabolic crosstalk in chronic pain pathogenesis.
- 1 May 2026
- Published by Taylor & Francis in Journal of Pain Research
- Vol. ume 19, 1-13
Open AccessDissecting the Causal Pathway from Herpes Zoster to Postherpetic Neuralgia: A Multi-Stage Mendelian Randomization Study Implicating VZV-Specific IgE and Cerebrospinal Fluid Metabolites
Background: Postherpetic neuralgia (PHN) develops in 10– 20% of herpes zoster (HZ) patients, yet the causal mechanisms driving this transition from acute infection to chronic neuropathic pain remain elusive. VZV-specific immunoglobulin E (IgE) and cerebrospinal fluid (CSF) metabolites represent candidate intermediate phenotypes that may mediate neuroimmune dysfunction, but their sequential causal relationships remain unexplored. We hypothesized that IgE and CSF metabolites constitute a hierarchical causal pathway from HZ exposure to PHN risk. Methods: We conducted a multi-stage Mendelian randomization study. First, two-sample Mendelian Randomization (MR) with MRlap correction established causal effects of HZ on IgE and screened 435 CSF metabolites for PHN associations. Given substantial HZ-PHN sample overlap precluding total effect estimation, multivariable MR (MVMR) evaluated whether metabolites mediate IgE effects on PHN. Results: HZ exhibited a significant causal effect on VZV-specific IgE (IVW: β = 0.237, 95% CI: 0.104– 0.371, P = 4.95× 10− 4). Among 435 metabolites, two showed effects significant after FDR correction: 1-stearoyl-2-oleoyl-GPC (β = 2.448, FDR-adjusted P = 0.043) and N-methylproline (β = − 0.425, FDR-adjusted P = 0.046). Critically, 5-methyluridine—nominally significant in univariable analysis (P = 0.095)—demonstrated a robust independent protective effect in MVMR (MV-IVW: β = − 3.227, 95% CI: − 5.765 to − 0.688, FDR-adjusted P = 0.025; OR = 0.040). The IgE→PHN effect attenuated to non-significance (P = 0.202) upon adjusting for 5-methyluridine, consistent with metabolic mediation. Conclusion: These findings support a sequential causal pathway from HZ to PHN involving IgE and CSF metabolites, identifying 5-methyluridine as a potential modifiable protective factor and highlighting neuroimmune-metabolic crosstalk in chronic pain pathogenesis.
- Open Access
- 18 January 2026
- Published by Elsevier in Asian Journal of Surgery
- Vol. 49 (7) , 4482-4483
Automated segmentation of keyhole for percutaneous endoscopic cervical decompression using deep learning
- 18 January 2026
- Published by Elsevier in Asian Journal of Surgery
- Vol. 49 (7) , 4482-4483
Open AccessAutomated segmentation of keyhole for percutaneous endoscopic cervical decompression using deep learning
- Open Access
- 2 October 2025
- Published by Springer Nature in Scientific Data
- Vol. 12 (1) , 1-11
Spine endoscopic atlas: an open-source dataset for surgical instrument segmentation
Endoscopic spine surgery (ESS) is a minimally invasive procedure used for spinal nerve decompression, herniated disc removal, and spinal fusion. Despite its many advantages, its steep learning curve poses a challenge to widespread adoption. The development of artificial intelligence (AI) systems is crucial for enhancing the precision and safety of ESS. The automatic segmentation of surgical instruments is a key step towards realizing intelligent surgical assistance systems. As such, this paper has created the Spine Endoscopic Atlas (SEA) dataset, a comprehensive collection of annotated images encompassing all instruments commonly used in spinal endoscopic surgery. In total, SEA contains 48,510 images and 10,662 instrument segmentations derived from real-world ESS. This dataset is specifically designed to train deep learning models for precise instrument segmentation. Through validation of five models, we demonstrate the dataset’s value in improving segmentation accuracy under complex conditions, providing a foundation for future AI advancements in ESS.
- 2 October 2025
- Published by Springer Nature in Scientific Data
- Vol. 12 (1) , 1-11
Open AccessSpine endoscopic atlas: an open-source dataset for surgical instrument segmentation
Endoscopic spine surgery (ESS) is a minimally invasive procedure used for spinal nerve decompression, herniated disc removal, and spinal fusion. Despite its many advantages, its steep learning curve poses a challenge to widespread adoption. The development of artificial intelligence (AI) systems is crucial for enhancing the precision and safety of ESS. The automatic segmentation of surgical instruments is a key step towards realizing intelligent surgical assistance systems. As such, this paper has created the Spine Endoscopic Atlas (SEA) dataset, a comprehensive collection of annotated images encompassing all instruments commonly used in spinal endoscopic surgery. In total, SEA contains 48,510 images and 10,662 instrument segmentations derived from real-world ESS. This dataset is specifically designed to train deep learning models for precise instrument segmentation. Through validation of five models, we demonstrate the dataset’s value in improving segmentation accuracy under complex conditions, providing a foundation for future AI advancements in ESS.
- Open Access
- 16 April 2025
- Published by Wiley in Health Science Reports
- Vol. 8 (4) , e70664
Complementary Repair Therapy as a Trending Topic in Discogenic Pain: A Bibliometric Study Over the Past 40 Years
Background and Aims: Bibliometric analysis has been frequently employed for visualizing trends within a particular scientific domain. The pain associated with discogenic origins has a significant impact on one's quality of life, but there is currently a lack of bibliometric analysis in the literature. Hence, this study aimed to examine related research in the field and identify the latest topics that are currently trending by conducting a bibliometric analysis.Methods: The Bibliometrix which developed in the statistical R‐packages was used for the data analysis. All related eligible publications were identified, and studies published from 1982 to 2023 were extracted from the Web of Science database.Results: Disc repair, bone marrow cells, platelet‐rich plasma, and the activation of inflammatory responses were identified as the trending topics after analyzing 977 journal articles. The most productive and influential journal was SPINE (Phila Pa 1976), which accounted for the largest publications and highest H index. The most productive and locally cited authors were Takahashi K., Ohtori S., and Aoki Y. from Chiba University. The top three productive and globally cited institutions were Chiba University, followed by University of California San Francisco and Korea University. The USA, China, and Japan were demonstrated as the most productive and globally cited countries.Conclusion: This study performed the first bibliometric analysis on discogenic pain and provided valuable insights into the latest trending topics in the field. Analysis reveals that recent research has primarily focused on complementary and regenerative approaches for repairing painful discs, as well as the role of inflammatory responses in disc pathology.
- 16 April 2025
- Published by Wiley in Health Science Reports
- Vol. 8 (4) , e70664
Open AccessComplementary Repair Therapy as a Trending Topic in Discogenic Pain: A Bibliometric Study Over the Past 40 Years
Background and Aims: Bibliometric analysis has been frequently employed for visualizing trends within a particular scientific domain. The pain associated with discogenic origins has a significant impact on one's quality of life, but there is currently a lack of bibliometric analysis in the literature. Hence, this study aimed to examine related research in the field and identify the latest topics that are currently trending by conducting a bibliometric analysis.Methods: The Bibliometrix which developed in the statistical R‐packages was used for the data analysis. All related eligible publications were identified, and studies published from 1982 to 2023 were extracted from the Web of Science database.Results: Disc repair, bone marrow cells, platelet‐rich plasma, and the activation of inflammatory responses were identified as the trending topics after analyzing 977 journal articles. The most productive and influential journal was SPINE (Phila Pa 1976), which accounted for the largest publications and highest H index. The most productive and locally cited authors were Takahashi K., Ohtori S., and Aoki Y. from Chiba University. The top three productive and globally cited institutions were Chiba University, followed by University of California San Francisco and Korea University. The USA, China, and Japan were demonstrated as the most productive and globally cited countries.Conclusion: This study performed the first bibliometric analysis on discogenic pain and provided valuable insights into the latest trending topics in the field. Analysis reveals that recent research has primarily focused on complementary and regenerative approaches for repairing painful discs, as well as the role of inflammatory responses in disc pathology.
- Open Access
- 29 November 2024
- Published by SAGE Publications in Journal of Central Nervous System Disease
- Vol. 16
Percutaneous spinal cord stimulation cylindrical lead placement for managing refractory neuropathic pain: A case series with an endoscopic-assisted approach
Background: The paddle lead (PL) and cylindrical lead (CL) remain the main implant categories in spinal cord stimulation (SCS) for treating neuropathic pain. Surgeons often complain about the greater trauma associated with PL implantation, while percutaneous endoscopic technique offers a promising approach for minimizing the trauma associated to PL implantation. However, there remains a dearth of real-world case study on endoscopy-assisted CL implantation. Purpose: This study aimed to demonstrate the endoscopic-assisted approach and outcomes of CL implantation in SCS for managing neuropathic pain. Research Design: A retrospective case series. Study Sample: Patients aged 18 years and above with chronic neuropathic pain persisting for at least three months, refractory to standardized conservative treatment, were enrolled between January 2021 and March 2023. Data Collection and Analysis: The surgical key steps including puncture, working cannula placement, endoscopic laminotomy and endoscopic CL introduction were demonstrated. Characteristics as demographics, follow-up time, visual analog scale (VAS) score, pain disability index (PDI) score and patient-reported outcomes measurement information system (PROMIS) scale were assessed. Results: Successful CL implantation under endoscopy was achieved in all patients, including 3 with failed back surgery syndrome, 2 with complex regional pain syndrome and 2 with chronic pelvic pain. No spinal cord injuries, dural tears, lead migration, lead fractures, or postoperative infections were observed. VAS score of regional pain, PDI score as well as PROMIS of patient’s quality of life were all significantly improved after surgery. Conclusion: Percutaneous endoscope-assisted CL implantation offered a new alternative technique for SCS in managing neuropathic pain.
- 29 November 2024
- Published by SAGE Publications in Journal of Central Nervous System Disease
- Vol. 16
Open AccessPercutaneous spinal cord stimulation cylindrical lead placement for managing refractory neuropathic pain: A case series with an endoscopic-assisted approach
Background: The paddle lead (PL) and cylindrical lead (CL) remain the main implant categories in spinal cord stimulation (SCS) for treating neuropathic pain. Surgeons often complain about the greater trauma associated with PL implantation, while percutaneous endoscopic technique offers a promising approach for minimizing the trauma associated to PL implantation. However, there remains a dearth of real-world case study on endoscopy-assisted CL implantation. Purpose: This study aimed to demonstrate the endoscopic-assisted approach and outcomes of CL implantation in SCS for managing neuropathic pain. Research Design: A retrospective case series. Study Sample: Patients aged 18 years and above with chronic neuropathic pain persisting for at least three months, refractory to standardized conservative treatment, were enrolled between January 2021 and March 2023. Data Collection and Analysis: The surgical key steps including puncture, working cannula placement, endoscopic laminotomy and endoscopic CL introduction were demonstrated. Characteristics as demographics, follow-up time, visual analog scale (VAS) score, pain disability index (PDI) score and patient-reported outcomes measurement information system (PROMIS) scale were assessed. Results: Successful CL implantation under endoscopy was achieved in all patients, including 3 with failed back surgery syndrome, 2 with complex regional pain syndrome and 2 with chronic pelvic pain. No spinal cord injuries, dural tears, lead migration, lead fractures, or postoperative infections were observed. VAS score of regional pain, PDI score as well as PROMIS of patient’s quality of life were all significantly improved after surgery. Conclusion: Percutaneous endoscope-assisted CL implantation offered a new alternative technique for SCS in managing neuropathic pain.
- Open Access
- 2 October 2024
- Published by Elsevier in Asian Journal of Surgery
- Vol. 48 (3) , 1977-1979
Automatic segmentation of surgical instruments in endoscopic spine surgery: A deep learning-based analysis
- 2 October 2024
- Published by Elsevier in Asian Journal of Surgery
- Vol. 48 (3) , 1977-1979
Open AccessAutomatic segmentation of surgical instruments in endoscopic spine surgery: A deep learning-based analysis
- Open Access
- 1 July 2024
- Published by Taylor & Francis in Journal of Multidisciplinary Healthcare
- Vol. ume 17, 3057-3069
Global Trends in Orthopedic Biofilm Research: A Bibliometric Analysis of 1994-2022
Bibliometric analysis is commonly used to visualize the knowledge foundation, trends, and patterns in a specific scientific field by performing a quantitative evaluation of the relevant literature. The purpose of this study was to perform a bibliometric analysis of recent studies in the field of orthopedic biofilm research and identify its current trends and hotspots. Research studies were retrieved from the Web of Science Core Collection and Scopus databases and analyzed in bibliometrix with R package (4.2.2). A total of 2426 literature were included in the study. Journal of orthopaedic research and Clinical orthopaedics and related research ranked first in terms of productivity and impact, with 57 published articles and 32 h-index, respectively. Trampuz A, Ohio State Univ and the United States ranked as the most productive authors, institutions, and countries. Biofilm formation, role of sonication, biomaterial mechanism and antibiotic loading have been investigated as the trend and hotspots in the field of orthopedic biofilm research. This study provides a thorough overview of the state of the art of current orthopedic biofilm research and offers valuable insights into recent trends and hotspots in this field.
- 1 July 2024
- Published by Taylor & Francis in Journal of Multidisciplinary Healthcare
- Vol. ume 17, 3057-3069
Open AccessGlobal Trends in Orthopedic Biofilm Research: A Bibliometric Analysis of 1994-2022
Bibliometric analysis is commonly used to visualize the knowledge foundation, trends, and patterns in a specific scientific field by performing a quantitative evaluation of the relevant literature. The purpose of this study was to perform a bibliometric analysis of recent studies in the field of orthopedic biofilm research and identify its current trends and hotspots. Research studies were retrieved from the Web of Science Core Collection and Scopus databases and analyzed in bibliometrix with R package (4.2.2). A total of 2426 literature were included in the study. Journal of orthopaedic research and Clinical orthopaedics and related research ranked first in terms of productivity and impact, with 57 published articles and 32 h-index, respectively. Trampuz A, Ohio State Univ and the United States ranked as the most productive authors, institutions, and countries. Biofilm formation, role of sonication, biomaterial mechanism and antibiotic loading have been investigated as the trend and hotspots in the field of orthopedic biofilm research. This study provides a thorough overview of the state of the art of current orthopedic biofilm research and offers valuable insights into recent trends and hotspots in this field.
- Open Access
- 10 May 2024
- Published by Wiley in Journal of Applied Clinical Medical Physics
- Vol. 25 (7) , e14378
Automatic segmentation of dura for quantitative analysis of lumbar stenosis: A deep learning study with 518 CT myelograms
Background: The diagnosis of lumbar spinal stenosis (LSS) can be challenging because radicular pain is not often present in the culprit‐level localization. Accurate segmentation and quantitative analysis of the lumbar dura on radiographic images are key to the accurate differential diagnosis of LSS. The aim of this study is to develop an automatic dura‐contouring tool for radiographic quantification on computed tomography myelogram (CTM) for patients with LSS.Methods: A total of 518 CTM cases with or without lumbar stenosis were included in this study. A deep learning (DL) segmentation algorithm 3‐dimensional (3D) U‐Net was deployed. A total of 210 labeled cases were used to develop the dura‐contouring tool, with the ratio of the training, independent testing, and external validation datasets being 150:30:30. The Dice score (DCS) was the primary measure to evaluate the segmentation performance of the 3D U‐Net, which was subsequently developed as the dura‐contouring tool to segment another unlabeled 308 CTM cases with LSS. Automatic masks of 446 slices on the stenotic levels were then meticulously reviewed and revised by human experts, and the cross‐sectional area (CSA) of the dura was compared.Results: The mean DCS of the 3D U‐Net were 0.905 ± 0.080, 0.933 ± 0.018, and 0.928 ± 0.034 in the five‐fold cross‐validation, the independent testing, and the external validation datasets, respectively. The segmentation performance of the dura‐contouring tool was also comparable to that of the second observer (the human expert). With the dura‐contouring tool, only 59.0% (263/446) of the automatic masks of the stenotic slices needed to be revised. In the revised cases, there were no significant differences in the dura CSA between automatic masks and corresponding revised masks (p = 0.652). Additionally, a strong correlation of dura CSA was found between the automatic masks and corresponding revised masks (r = 0.805).Conclusions: A dura‐contouring tool was developed that could automatically segment the dural sac on CTM, and it demonstrated high accuracy and generalization ability. Additionally, the dura‐contouring tool has the potential to be applied in patients with LSS because it facilitates the quantification of the dural CSA on stenotic slices.
- 10 May 2024
- Published by Wiley in Journal of Applied Clinical Medical Physics
- Vol. 25 (7) , e14378
Open AccessAutomatic segmentation of dura for quantitative analysis of lumbar stenosis: A deep learning study with 518 CT myelograms
Background: The diagnosis of lumbar spinal stenosis (LSS) can be challenging because radicular pain is not often present in the culprit‐level localization. Accurate segmentation and quantitative analysis of the lumbar dura on radiographic images are key to the accurate differential diagnosis of LSS. The aim of this study is to develop an automatic dura‐contouring tool for radiographic quantification on computed tomography myelogram (CTM) for patients with LSS.Methods: A total of 518 CTM cases with or without lumbar stenosis were included in this study. A deep learning (DL) segmentation algorithm 3‐dimensional (3D) U‐Net was deployed. A total of 210 labeled cases were used to develop the dura‐contouring tool, with the ratio of the training, independent testing, and external validation datasets being 150:30:30. The Dice score (DCS) was the primary measure to evaluate the segmentation performance of the 3D U‐Net, which was subsequently developed as the dura‐contouring tool to segment another unlabeled 308 CTM cases with LSS. Automatic masks of 446 slices on the stenotic levels were then meticulously reviewed and revised by human experts, and the cross‐sectional area (CSA) of the dura was compared.Results: The mean DCS of the 3D U‐Net were 0.905 ± 0.080, 0.933 ± 0.018, and 0.928 ± 0.034 in the five‐fold cross‐validation, the independent testing, and the external validation datasets, respectively. The segmentation performance of the dura‐contouring tool was also comparable to that of the second observer (the human expert). With the dura‐contouring tool, only 59.0% (263/446) of the automatic masks of the stenotic slices needed to be revised. In the revised cases, there were no significant differences in the dura CSA between automatic masks and corresponding revised masks (p = 0.652). Additionally, a strong correlation of dura CSA was found between the automatic masks and corresponding revised masks (r = 0.805).Conclusions: A dura‐contouring tool was developed that could automatically segment the dural sac on CTM, and it demonstrated high accuracy and generalization ability. Additionally, the dura‐contouring tool has the potential to be applied in patients with LSS because it facilitates the quantification of the dural CSA on stenotic slices.
- Open Access
- 26 December 2023
- Published by MDPI AG in Diagnostics
- Vol. 14 (1) , 53
Machine Learning Predicts Decompression Levels for Lumbar Spinal Stenosis Using Canal Radiomic Features from Computed Tomography Myelography
Background: The accurate preoperative identification of decompression levels is crucial for the success of surgery in patients with multi-level lumbar spinal stenosis (LSS). The objective of this study was to develop machine learning (ML) classifiers that can predict decompression levels using computed tomography myelography (CTM) data from LSS patients. Methods: A total of 1095 lumbar levels from 219 patients were included in this study. The bony spinal canal in CTM images was manually delineated, and radiomic features were extracted. The extracted data were randomly divided into training and testing datasets (8:2). Six feature selection methods combined with 12 ML algorithms were employed, resulting in a total of 72 ML classifiers. The main evaluation indicator for all classifiers was the area under the curve of the receiver operating characteristic (ROC-AUC), with the precision–recall AUC (PR-AUC) serving as the secondary indicator. The prediction outcome of ML classifiers was decompression level or not. Results: The embedding linear support vector (embeddingLSVC) was the optimal feature selection method. The feature importance analysis revealed the top 5 important features of the 15 radiomic predictors, which included 2 texture features, 2 first-order intensity features, and 1 shape feature. Except for shape features, these features might be eye-discernible but hardly quantified. The top two ML classifiers were embeddingLSVC combined with support vector machine (EmbeddingLSVC_SVM) and embeddingLSVC combined with gradient boosting (EmbeddingLSVC_GradientBoost). These classifiers achieved ROC-AUCs over 0.90 and PR-AUCs over 0.80 in independent testing among the 72 classifiers. Further comparisons indicated that EmbeddingLSVC_SVM appeared to be the optimal classifier, demonstrating superior discrimination ability, slight advantages in the Brier scores on the calibration curve, and Net benefits on the Decision Curve Analysis. Conclusions: ML successfully extracted valuable and interpretable radiomic features from the spinal canal using CTM images, and accurately predicted decompression levels for LSS patients. The EmbeddingLSVC_SVM classifier has the potential to assist surgical decision making in clinical practice, as it showed high discrimination, advantageous calibration, and competitive utility in selecting decompression levels in LSS patients using canal radiomic features from CTM.
- 26 December 2023
- Published by MDPI AG in Diagnostics
- Vol. 14 (1) , 53
Open AccessMachine Learning Predicts Decompression Levels for Lumbar Spinal Stenosis Using Canal Radiomic Features from Computed Tomography Myelography
Background: The accurate preoperative identification of decompression levels is crucial for the success of surgery in patients with multi-level lumbar spinal stenosis (LSS). The objective of this study was to develop machine learning (ML) classifiers that can predict decompression levels using computed tomography myelography (CTM) data from LSS patients. Methods: A total of 1095 lumbar levels from 219 patients were included in this study. The bony spinal canal in CTM images was manually delineated, and radiomic features were extracted. The extracted data were randomly divided into training and testing datasets (8:2). Six feature selection methods combined with 12 ML algorithms were employed, resulting in a total of 72 ML classifiers. The main evaluation indicator for all classifiers was the area under the curve of the receiver operating characteristic (ROC-AUC), with the precision–recall AUC (PR-AUC) serving as the secondary indicator. The prediction outcome of ML classifiers was decompression level or not. Results: The embedding linear support vector (embeddingLSVC) was the optimal feature selection method. The feature importance analysis revealed the top 5 important features of the 15 radiomic predictors, which included 2 texture features, 2 first-order intensity features, and 1 shape feature. Except for shape features, these features might be eye-discernible but hardly quantified. The top two ML classifiers were embeddingLSVC combined with support vector machine (EmbeddingLSVC_SVM) and embeddingLSVC combined with gradient boosting (EmbeddingLSVC_GradientBoost). These classifiers achieved ROC-AUCs over 0.90 and PR-AUCs over 0.80 in independent testing among the 72 classifiers. Further comparisons indicated that EmbeddingLSVC_SVM appeared to be the optimal classifier, demonstrating superior discrimination ability, slight advantages in the Brier scores on the calibration curve, and Net benefits on the Decision Curve Analysis. Conclusions: ML successfully extracted valuable and interpretable radiomic features from the spinal canal using CTM images, and accurately predicted decompression levels for LSS patients. The EmbeddingLSVC_SVM classifier has the potential to assist surgical decision making in clinical practice, as it showed high discrimination, advantageous calibration, and competitive utility in selecting decompression levels in LSS patients using canal radiomic features from CTM.