Qiang Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Qiang Zhang has a total of 98 co-authors. Here are the top co-authors by number of shared publications:
Yongxiang Cheng
Shandong Public Health Clinical Center, Shandong University, Jinan, People’s Republic of China
4 shared publications- Deqiang Chen
Shandong University of Traditional Chinese Medicine Affiliated Hospital, Jinan City, Shandong Province, China.
3 shared publications Jin Wang
Shanghai University of Traditional Chinese Medicine
3 shared publicationsTing Li
Shandong Public Health Clinical Center, Shandong University, Jinan, People’s Republic of China
3 shared publicationsXin Xiu
Shandong University of Traditional Chinese Medicine, Shandong, China
3 shared publications- Yanke Hao
Affiliated Hospital of Shandong University of Traditional Chinese Medicine, Jinan, China
3 shared publications
Latest Publications
- Open Access
- 30 April 2026
- Published by The Korean Spinal Neurosurgery Society in Neurospine
- Vol. 23 (2) , 487-499
Diagnostic Value of Metagenomic Next-Generation Sequencing for Suspected Native Spinal Brucella Infection: A Multicenter Study
Objective: The aim is to study the diagnostic positive rates of metagenomic next-generation sequencing (mNGS), microbial culture, and serologic testing in suspected native spinal brucellosis, and to evaluate the clinical value of their combined application.Methods: In this multicenter, retrospective observational study, 128 patients with suspected native spinal brucellosis from 6 medical centers (February 2020 to February 2025) were enrolled. Specimens from infection sites were subjected to microbial culture, mNGS, and serological testing (agglutination test).Results: Of the 128 patients with suspected native spinal Brucella infections, 118 patients were diagnosed with Brucella spondylitis. Among the 118 confirmed Brucella spondylitis cases, mNGS demonstrated a positivity rate of 92.37% (109 of 118), significantly higher than that of culture (26.27%, 31 of 118) and agglutination test (83.05%, 98 of 118). In the 87 culture-negative samples, mNGS detected Brucella in 91.95% (80 of 87), compared to 82.76% (72 of 87) by agglutination test. mNGS confirmed Brucella infection in all 16 cases that were agglutination test negative. mNGS combined with agglutination tests can effectively complement each other, improving the sensitivity of diagnosis and thereby minimizing missed diagnoses to the greatest extent. Among the 10 nonbrucellar spinal pathologies, agglutination test showed a high false-positive rate of 90% (9 of 10), whereas mNGS had a 10% (1 of 10) false-positive rate. Therefore, the agglutination test has a relatively high rate of false positives.Conclusion: mNGS detection represents an effective adjunct to microbial culture and the agglutination test. The concurrent use of all 3 methods enhances diagnostic accuracy and reduces the likelihood of missed and incorrect diagnoses, significantly improving patient prognosis and guiding personalized clinical treatment.
- 30 April 2026
- Published by The Korean Spinal Neurosurgery Society in Neurospine
- Vol. 23 (2) , 487-499
Open AccessDiagnostic Value of Metagenomic Next-Generation Sequencing for Suspected Native Spinal Brucella Infection: A Multicenter Study
Objective: The aim is to study the diagnostic positive rates of metagenomic next-generation sequencing (mNGS), microbial culture, and serologic testing in suspected native spinal brucellosis, and to evaluate the clinical value of their combined application.Methods: In this multicenter, retrospective observational study, 128 patients with suspected native spinal brucellosis from 6 medical centers (February 2020 to February 2025) were enrolled. Specimens from infection sites were subjected to microbial culture, mNGS, and serological testing (agglutination test).Results: Of the 128 patients with suspected native spinal Brucella infections, 118 patients were diagnosed with Brucella spondylitis. Among the 118 confirmed Brucella spondylitis cases, mNGS demonstrated a positivity rate of 92.37% (109 of 118), significantly higher than that of culture (26.27%, 31 of 118) and agglutination test (83.05%, 98 of 118). In the 87 culture-negative samples, mNGS detected Brucella in 91.95% (80 of 87), compared to 82.76% (72 of 87) by agglutination test. mNGS confirmed Brucella infection in all 16 cases that were agglutination test negative. mNGS combined with agglutination tests can effectively complement each other, improving the sensitivity of diagnosis and thereby minimizing missed diagnoses to the greatest extent. Among the 10 nonbrucellar spinal pathologies, agglutination test showed a high false-positive rate of 90% (9 of 10), whereas mNGS had a 10% (1 of 10) false-positive rate. Therefore, the agglutination test has a relatively high rate of false positives.Conclusion: mNGS detection represents an effective adjunct to microbial culture and the agglutination test. The concurrent use of all 3 methods enhances diagnostic accuracy and reduces the likelihood of missed and incorrect diagnoses, significantly improving patient prognosis and guiding personalized clinical treatment.
- Open Access
- 31 January 2026
- Published by Springer Nature in BMC Medical Imaging
- Vol. 26 (1)
Enhancing WSI image classification with graph convolutional neural networks and model uncertainty modeling
Background: The primary research question addresses whether integrating Graph Convolutional Neural Networks with model uncertainty modeling can improve the accuracy and robustness of Whole Slide Imaging (WSI) classifications in pathology. Methods: This study employed a novel framework combining GCNs with uncertainty quantification techniques to classify WSI images of spinal infections. We constructed a graph from segmented regions of WSI, where nodes represented segmented pathological features and edges represented spatial relationships. The model was trained on a dataset of 422 cases from the Shandong Provincial Center for Disease Control and Prevention, annotated for tuberculosis, brucellosis, and purulent spondylitis. Performance metrics included accuracy, precision, recall, and F1 score. Results: The integrated GCN model demonstrated a classification accuracy of 87%, with a recall of 85% and an F1 score of 0.86. These metrics signify an improvement over traditional CNN models, which showed a 10% lower performance in comparative analyses. The model also effectively quantified uncertainty, enhancing confidence in diagnostic decisions. Conclusions: Integrating GCNs with model uncertainty modeling enhances the accuracy and reliability of WSI image classification in pathology. This approach significantly improves the capture of spatial relationships and pathological feature recognition, offering a robust framework for supporting diagnostic and therapeutic decisions in medical practice. Clinical relevance: The enhanced ability to classify and understand WSI images using this method has significant implications for pathology, potentially leading to more accurate and reliable diagnoses. This approach could be particularly useful in remote diagnostics and in environments where expert pathological consultation is limited.
- 31 January 2026
- Published by Springer Nature in BMC Medical Imaging
- Vol. 26 (1)
Open AccessEnhancing WSI image classification with graph convolutional neural networks and model uncertainty modeling
Background: The primary research question addresses whether integrating Graph Convolutional Neural Networks with model uncertainty modeling can improve the accuracy and robustness of Whole Slide Imaging (WSI) classifications in pathology. Methods: This study employed a novel framework combining GCNs with uncertainty quantification techniques to classify WSI images of spinal infections. We constructed a graph from segmented regions of WSI, where nodes represented segmented pathological features and edges represented spatial relationships. The model was trained on a dataset of 422 cases from the Shandong Provincial Center for Disease Control and Prevention, annotated for tuberculosis, brucellosis, and purulent spondylitis. Performance metrics included accuracy, precision, recall, and F1 score. Results: The integrated GCN model demonstrated a classification accuracy of 87%, with a recall of 85% and an F1 score of 0.86. These metrics signify an improvement over traditional CNN models, which showed a 10% lower performance in comparative analyses. The model also effectively quantified uncertainty, enhancing confidence in diagnostic decisions. Conclusions: Integrating GCNs with model uncertainty modeling enhances the accuracy and reliability of WSI image classification in pathology. This approach significantly improves the capture of spatial relationships and pathological feature recognition, offering a robust framework for supporting diagnostic and therapeutic decisions in medical practice. Clinical relevance: The enhanced ability to classify and understand WSI images using this method has significant implications for pathology, potentially leading to more accurate and reliable diagnoses. This approach could be particularly useful in remote diagnostics and in environments where expert pathological consultation is limited.
- Open Access
- 19 December 2025
- Published by Frontiers Media SA in Frontiers in Pharmacology
- Vol. 16, 1684771
Rifapentine- and moxifloxacin-containing short-course regimens for mild spinal tuberculosis: study protocol for a multicenter, randomized, non-inferiority phase II clinical trial
Background: Spinal tuberculosis is the most common form of osteoarticular infection, with recommended anti-tuberculosis treatment durations typically being long (9–18 months), even for mild cases, increasing the risks of drug resistance, toxicity, and poor patient adherence. This study aims to explore, in patients with mild spinal tuberculosis, whether a new combination of anti-tuberculosis drugs (containing rifapentine and moxifloxacin) could shorten the treatment duration. Methods and Analysis: This trial is an open-label, randomized, controlled, non-inferiority trial comparing the efficacy and safety of a short-course regimen [6-month anti-TB regimen] containing rifapentine and moxifloxacin with empirical long-course regimen [12-month anti-TB regimen] in the treatment of patients with mild spinal tuberculosis. Patients diagnosed with mild spinal tuberculosis who meet the inclusion and exclusion criteria will be recruited and randomized in a 1:1 ratio to either of the two arms. Empirical long-course regimen includes rifampin, isoniazid, pyrazinamide and ethambutol (2RHZE/10RH), while short-course regimen includes rifapentine, moxifloxacin, isoniazid and pyrazinamide (2HRtZM/4HRtM). The primary outcomes are TB-recurrence rate at 24 months after completion of treatment and proportion of participants with grade 3 or higher adverse events during treatment with study medications. Discussion: This trial will provide evidence whether a short-course regimen of anti-TB drugs (2HRtZM/4HRtM) is non-inferior to the empirical long-course regimen (2RHZE/10RH) in patients with mild spinal tuberculosis. If non-inferiority is established, it will contribute to a more optimized treatment for spinal tuberculosis. Clinical Trial Registration: This study is registered with https://clinicaltrials.gov/ct2/show/NCT06917495 as NCT06917495.
- 19 December 2025
- Published by Frontiers Media SA in Frontiers in Pharmacology
- Vol. 16, 1684771
Open AccessRifapentine- and moxifloxacin-containing short-course regimens for mild spinal tuberculosis: study protocol for a multicenter, randomized, non-inferiority phase II clinical trial
Background: Spinal tuberculosis is the most common form of osteoarticular infection, with recommended anti-tuberculosis treatment durations typically being long (9–18 months), even for mild cases, increasing the risks of drug resistance, toxicity, and poor patient adherence. This study aims to explore, in patients with mild spinal tuberculosis, whether a new combination of anti-tuberculosis drugs (containing rifapentine and moxifloxacin) could shorten the treatment duration. Methods and Analysis: This trial is an open-label, randomized, controlled, non-inferiority trial comparing the efficacy and safety of a short-course regimen [6-month anti-TB regimen] containing rifapentine and moxifloxacin with empirical long-course regimen [12-month anti-TB regimen] in the treatment of patients with mild spinal tuberculosis. Patients diagnosed with mild spinal tuberculosis who meet the inclusion and exclusion criteria will be recruited and randomized in a 1:1 ratio to either of the two arms. Empirical long-course regimen includes rifampin, isoniazid, pyrazinamide and ethambutol (2RHZE/10RH), while short-course regimen includes rifapentine, moxifloxacin, isoniazid and pyrazinamide (2HRtZM/4HRtM). The primary outcomes are TB-recurrence rate at 24 months after completion of treatment and proportion of participants with grade 3 or higher adverse events during treatment with study medications. Discussion: This trial will provide evidence whether a short-course regimen of anti-TB drugs (2HRtZM/4HRtM) is non-inferior to the empirical long-course regimen (2RHZE/10RH) in patients with mild spinal tuberculosis. If non-inferiority is established, it will contribute to a more optimized treatment for spinal tuberculosis. Clinical Trial Registration: This study is registered with https://clinicaltrials.gov/ct2/show/NCT06917495 as NCT06917495.
- 2 October 2025
- Published by Springer Nature in European Spine Journal
- Vol. 35 (3) , 1270-1282
Multimodal imaging fusion and machine learning model development: differential diagnosis of spinal inflammatory lesions using combined CT hounsfield units and MRI features
Objective The objective is to develop a differential diagnosis model for tuberculous spondylitis (TS) and pyogenic spondylitis (PS) by integrating MRI morphological features and computed tomography (CT) density parameters (Hounsfield Units, HU). This study aims to leverage multimodal data complementarity to achieve fusion of qualitative and quantitative information, thereby providing clinicians with a rapid and objective decision support tool for spinal inflammatory lesion characterization. Methods Imaging data were extracted from MRI and CT scans of patients with TS and PS, then compared and summarized. Receiver operating characteristic (ROC) curves were used to determine optimal HU value thresholds. The least absolute shrinkage and selection operator (Lasso) regression was applied to identify the most predictive features for model construction. A logistic regression-based predictive model was developed and visualized as a nomogram. Model validation was performed using bootstrap resampling, ROC analysis, and decision curve analysis (DCA). Results A total of 171 patients with TS (n = 91) or PS (n = 80) were included. Statistically significant differences in MRI features were observed between the two groups (P < 0.05). Additionally, significant HU value differences were found in diseased vertebral endplates, small cavitary abscesses, large cavitary abscesses, and intravertebral abscesses between TS and PS patients (P < 0.05). The predictive model incorporated seven independent predictors. Calibration curves, ROC analysis, and DCA all demonstrated excellent model performance. Conclusion Combined MRI and CT HU value analysis effectively differentiates TS from PS. The predictive model integrating imaging features and quantitative parameters demonstrates high accuracy and clinical utility, offering a novel approach to optimize diagnostic and treatment strategies for spinal infectious diseases.
- 2 October 2025
- Published by Springer Nature in European Spine Journal
- Vol. 35 (3) , 1270-1282
Multimodal imaging fusion and machine learning model development: differential diagnosis of spinal inflammatory lesions using combined CT hounsfield units and MRI features
Objective The objective is to develop a differential diagnosis model for tuberculous spondylitis (TS) and pyogenic spondylitis (PS) by integrating MRI morphological features and computed tomography (CT) density parameters (Hounsfield Units, HU). This study aims to leverage multimodal data complementarity to achieve fusion of qualitative and quantitative information, thereby providing clinicians with a rapid and objective decision support tool for spinal inflammatory lesion characterization. Methods Imaging data were extracted from MRI and CT scans of patients with TS and PS, then compared and summarized. Receiver operating characteristic (ROC) curves were used to determine optimal HU value thresholds. The least absolute shrinkage and selection operator (Lasso) regression was applied to identify the most predictive features for model construction. A logistic regression-based predictive model was developed and visualized as a nomogram. Model validation was performed using bootstrap resampling, ROC analysis, and decision curve analysis (DCA). Results A total of 171 patients with TS (n = 91) or PS (n = 80) were included. Statistically significant differences in MRI features were observed between the two groups (P < 0.05). Additionally, significant HU value differences were found in diseased vertebral endplates, small cavitary abscesses, large cavitary abscesses, and intravertebral abscesses between TS and PS patients (P < 0.05). The predictive model incorporated seven independent predictors. Calibration curves, ROC analysis, and DCA all demonstrated excellent model performance. Conclusion Combined MRI and CT HU value analysis effectively differentiates TS from PS. The predictive model integrating imaging features and quantitative parameters demonstrates high accuracy and clinical utility, offering a novel approach to optimize diagnostic and treatment strategies for spinal infectious diseases.
- Open Access
- 14 August 2025
- Vol. 20 (1) , 768
Mid- to short-term efficacy of Uniportal Video-Assisted Thoracoscopic Surgery in the treatment of thoracic spinal tuberculosis.
Surgical treatment of thoracic spine tuberculosis using traditional procedures is often challenged by high trauma and slow recovery. This study aimed to investigate the short- and medium-term efficacy and safety of Uniportal Video-Assisted Thoracoscopic Surgery, a minimally invasive single-port thoracoscopy combined with posterior internal fixation for the treatment of thoracic spinal tuberculosis. Nineteen patients with thoracic spinal tuberculosis (15 men and four women, mean age of 52 years) were retrospectively included. All patients underwent single-port thoracoscopic lesion removal and intervertebral implant fusion combined with posterior internal fixation. Surgical time, intraoperative bleeding volume, hospitalization time, and complications were recorded, and the rate of bone fusion and correction of the kyphosis angle were assessed using imaging, with a follow-up period of 6-12 months. All patients achieved postoperative bone fusion without serious complications. The average operative time was 149.58 min (75-212 min), the average intraoperative bleeding volume was 210.53 mL (50-600 mL), and the average hospital stay was 15.58 d (2-29 d). The mean thoracic kyphosis angle improved from 23.33° preoperatively to 19.00° at the final follow-up. Single-port thoracoscopic technique combined with posterior internal fixation can effectively remove thoracic tuberculosis foci and correct kyphosis deformity. Moreover, it has the advantages of less trauma, quicker recovery, and significant short- and medium-term efficacy, making it a feasible, minimally invasive surgical procedure for treating thoracic spinal tuberculosis. However, further optimizing the instrumentation design and expanding the sample size in the future is necessary to validate its long-term effects.
- 14 August 2025
- Vol. 20 (1) , 768
Open AccessMid- to short-term efficacy of Uniportal Video-Assisted Thoracoscopic Surgery in the treatment of thoracic spinal tuberculosis.
Surgical treatment of thoracic spine tuberculosis using traditional procedures is often challenged by high trauma and slow recovery. This study aimed to investigate the short- and medium-term efficacy and safety of Uniportal Video-Assisted Thoracoscopic Surgery, a minimally invasive single-port thoracoscopy combined with posterior internal fixation for the treatment of thoracic spinal tuberculosis. Nineteen patients with thoracic spinal tuberculosis (15 men and four women, mean age of 52 years) were retrospectively included. All patients underwent single-port thoracoscopic lesion removal and intervertebral implant fusion combined with posterior internal fixation. Surgical time, intraoperative bleeding volume, hospitalization time, and complications were recorded, and the rate of bone fusion and correction of the kyphosis angle were assessed using imaging, with a follow-up period of 6-12 months. All patients achieved postoperative bone fusion without serious complications. The average operative time was 149.58 min (75-212 min), the average intraoperative bleeding volume was 210.53 mL (50-600 mL), and the average hospital stay was 15.58 d (2-29 d). The mean thoracic kyphosis angle improved from 23.33° preoperatively to 19.00° at the final follow-up. Single-port thoracoscopic technique combined with posterior internal fixation can effectively remove thoracic tuberculosis foci and correct kyphosis deformity. Moreover, it has the advantages of less trauma, quicker recovery, and significant short- and medium-term efficacy, making it a feasible, minimally invasive surgical procedure for treating thoracic spinal tuberculosis. However, further optimizing the instrumentation design and expanding the sample size in the future is necessary to validate its long-term effects.
- 17 May 2025
- Published by Springer Nature in European Spine Journal
- Vol. 34 (10) , 4638-4649
An imaging-based clinical prediction model to differentiate brucellar spondylitis from pyogenic spondylitis: a multicenter retrospective observational study
Purpose Differentiating between pyogenic (PS) and brucellar (BS) spondylitis is clinically challenging due to their similar clinical symptoms, with delayed diagnosis or misdiagnosis common, causing trouble for surgeons in selecting appropriate treatment strategies. Currently, radiology-based diagnostic models for PS and BS are lacking. This study aimed to combine magnetic resonance (MR) and radiographic imaging to elucidate the differences between PS and BS and develop a novel diagnostic model for differential diagnosis. Methods We collected and analyzed the differences between MR and radiological images of patients with PS and BS from two medical institutions. A nomogram was constructed using least absolute shrinkage and selection operator (LASSO) regression, alongside univariate and multivariate analyses to select the best features of the predictive model. Model discrimination, calibration, and clinical utility were assessed using receiver operating characteristic, calibration, and decision curve analyses. Results Among the enrolled 342 patients with PS (n = 167) or BS (n = 175), we found significant differences in MR and radiological characteristics between the two groups. LASSO regression analysis revealed that thoracic involvement, involved vertebrae number, parrot beak osteophyte presence, endplate destruction, and intervertebral disc signal strength on T1-weighted sequences were independent predictive factors for differentiating between PS and BS. The imaging-based clinical prediction model showed high accuracy in the training and validation sets, with the area under the curve achieving 0.861 and 0.908, respectively, and a significant net benefit in the threshold probability, indicating high clinical potential of the model. Conclusion This imaging-based model offers a useful tool for efficiently differentiating PS and BS, facilitating prompt diagnosis and treatment and mitigating incorrect or delayed diagnosis.
- 17 May 2025
- Published by Springer Nature in European Spine Journal
- Vol. 34 (10) , 4638-4649
An imaging-based clinical prediction model to differentiate brucellar spondylitis from pyogenic spondylitis: a multicenter retrospective observational study
Purpose Differentiating between pyogenic (PS) and brucellar (BS) spondylitis is clinically challenging due to their similar clinical symptoms, with delayed diagnosis or misdiagnosis common, causing trouble for surgeons in selecting appropriate treatment strategies. Currently, radiology-based diagnostic models for PS and BS are lacking. This study aimed to combine magnetic resonance (MR) and radiographic imaging to elucidate the differences between PS and BS and develop a novel diagnostic model for differential diagnosis. Methods We collected and analyzed the differences between MR and radiological images of patients with PS and BS from two medical institutions. A nomogram was constructed using least absolute shrinkage and selection operator (LASSO) regression, alongside univariate and multivariate analyses to select the best features of the predictive model. Model discrimination, calibration, and clinical utility were assessed using receiver operating characteristic, calibration, and decision curve analyses. Results Among the enrolled 342 patients with PS (n = 167) or BS (n = 175), we found significant differences in MR and radiological characteristics between the two groups. LASSO regression analysis revealed that thoracic involvement, involved vertebrae number, parrot beak osteophyte presence, endplate destruction, and intervertebral disc signal strength on T1-weighted sequences were independent predictive factors for differentiating between PS and BS. The imaging-based clinical prediction model showed high accuracy in the training and validation sets, with the area under the curve achieving 0.861 and 0.908, respectively, and a significant net benefit in the threshold probability, indicating high clinical potential of the model. Conclusion This imaging-based model offers a useful tool for efficiently differentiating PS and BS, facilitating prompt diagnosis and treatment and mitigating incorrect or delayed diagnosis.
- Open Access
- 6 March 2025
- Published by Bio-Byword Scientific Publishing, Pty. Ltd. in Journal of Clinical and Nursing Research
- Vol. 9 (2) , 1-6
Exploration of the Value of Extract of Wuwei Xiaodu Drink on Rabbit Model of Spinal Infection
Objective: To study the therapeutic effect of the Extract of Wuwei Xiaodu Drink on spinal infection and provide the scientific basis for clinical application. Methods: By establishing a rabbit model of spinal infection, this paper observed and analyzed the changes in body mass before and after the intervention and the comparison of inflammation-related factors and blood leukocyte counts among the three groups. Results: There was a significant difference in the changes in body mass of rabbits before and after intervention in the experimental group, control group and blank group (P < 0.05); there was no statistically significant difference in calcitoninogen, C-reactive protein and routine blood leukocyte counts between the experimental group and the control group (P > 0.05), and there was a statistically significant difference in calcitoninogen, C-reactive protein and routine blood leukocyte counts between the experimental group and the blank group (P < 0.05). Conclusion: The Extract of Wuwei Xiaodu Drink can play a protective role by regulating the level of inflammatory factors in blood routine leukocyte count and reducing the inflammatory reaction in the spinal cord injury area.
- 6 March 2025
- Published by Bio-Byword Scientific Publishing, Pty. Ltd. in Journal of Clinical and Nursing Research
- Vol. 9 (2) , 1-6
Open AccessExploration of the Value of Extract of Wuwei Xiaodu Drink on Rabbit Model of Spinal Infection
Objective: To study the therapeutic effect of the Extract of Wuwei Xiaodu Drink on spinal infection and provide the scientific basis for clinical application. Methods: By establishing a rabbit model of spinal infection, this paper observed and analyzed the changes in body mass before and after the intervention and the comparison of inflammation-related factors and blood leukocyte counts among the three groups. Results: There was a significant difference in the changes in body mass of rabbits before and after intervention in the experimental group, control group and blank group (P < 0.05); there was no statistically significant difference in calcitoninogen, C-reactive protein and routine blood leukocyte counts between the experimental group and the control group (P > 0.05), and there was a statistically significant difference in calcitoninogen, C-reactive protein and routine blood leukocyte counts between the experimental group and the blank group (P < 0.05). Conclusion: The Extract of Wuwei Xiaodu Drink can play a protective role by regulating the level of inflammatory factors in blood routine leukocyte count and reducing the inflammatory reaction in the spinal cord injury area.
- Open Access
- 20 November 2024
- Published by Springer Nature in Research Square
Optimizing Spinal Infection Classification in Whole-Slide Images via Graph Convolutional Network and Model Uncertainty Integration
Background Spinal infections such as pyogenic spondylitis, spinal tuberculosis, and brucellar spondylitis are severe conditions that can lead to significant spinal damage and chronic pain. Whole-slide imaging (WSI) provides valuable visual information in pathological diagnoses. However, owing to the complexity and high dimensionality of WSI data, traditional manual diagnostic methods are often time-consuming and prone to errors. Therefore, developing an automated image analysis method is crucial to enhance the diagnostic accuracy and efficiency of WSI for spinal infections. Methods This study employed a novel framework that combines Graph Convolutional Networks (GCNs) with uncertainty quantification techniques to classify WSI images of spinal infections. A graph was constructed from segmented regions of the WSI, where nodes represented segmented pathological features and edges represented spatial relationships. The model was trained using a dataset of 422 cases from a provincial center for disease control and prevention and annotated for tuberculosis, brucellosis, and purulent spondylitis. The performance metrics were accuracy, precision, recall, and F1 scores. Results The integrated GCN model demonstrated a classification accuracy of 87%, recall of 85%, and F1 score of 0.86. Comparative analyses revealed that the GCN model exhibited a 10% higher performance than that of traditional CNN models. Moreover, the GCN model effectively quantified uncertainty and enhanced confidence in diagnostic decisions. Conclusions Integrating GCNs with model uncertainty enhances the accuracy and reliability of WSI image classification in pathology. This method significantly improves the capture of spatial relationships and identification of pathological features of spinal infections, offering a robust framework for supporting diagnostic and therapeutic decisions in medical practice.
- 20 November 2024
- Published by Springer Nature in Research Square
Open AccessOptimizing Spinal Infection Classification in Whole-Slide Images via Graph Convolutional Network and Model Uncertainty Integration
Background Spinal infections such as pyogenic spondylitis, spinal tuberculosis, and brucellar spondylitis are severe conditions that can lead to significant spinal damage and chronic pain. Whole-slide imaging (WSI) provides valuable visual information in pathological diagnoses. However, owing to the complexity and high dimensionality of WSI data, traditional manual diagnostic methods are often time-consuming and prone to errors. Therefore, developing an automated image analysis method is crucial to enhance the diagnostic accuracy and efficiency of WSI for spinal infections. Methods This study employed a novel framework that combines Graph Convolutional Networks (GCNs) with uncertainty quantification techniques to classify WSI images of spinal infections. A graph was constructed from segmented regions of the WSI, where nodes represented segmented pathological features and edges represented spatial relationships. The model was trained using a dataset of 422 cases from a provincial center for disease control and prevention and annotated for tuberculosis, brucellosis, and purulent spondylitis. The performance metrics were accuracy, precision, recall, and F1 scores. Results The integrated GCN model demonstrated a classification accuracy of 87%, recall of 85%, and F1 score of 0.86. Comparative analyses revealed that the GCN model exhibited a 10% higher performance than that of traditional CNN models. Moreover, the GCN model effectively quantified uncertainty and enhanced confidence in diagnostic decisions. Conclusions Integrating GCNs with model uncertainty enhances the accuracy and reliability of WSI image classification in pathology. This method significantly improves the capture of spatial relationships and identification of pathological features of spinal infections, offering a robust framework for supporting diagnostic and therapeutic decisions in medical practice.
- Open Access
- 22 August 2024
- Published by Springer Nature in Annals of Clinical Microbiology and Antimicrobials
- Vol. 23 (1) , 1-10
Evaluation of targeted sequencing for pathogen identification in bone and joint infections: a cohort study from China
Bone and joint tuberculosis (BJTB) is a distinct variant of tuberculosis in which clinical diagnosis often leads to relative misdiagnosis and missed diagnoses. This study aimed to evaluate the diagnostic accuracy of the targeted nanopore sequencing (TNPseq) assay for BJTB patients in China. The study enrolled a cohort of 163 patients with suspected BJTB. Diagnostic testing was performed using the TNPseq assay on samples including punctured tissue, pus, and blood. The diagnostic accuracy of the TNPseq assay was then compared with that of the T-SPOT and Xpert MTB/RIF assays. TNPseq exhibited superior performance in terms of accuracy, demonstrating a sensitivity of 76.3% (95% CI: 71.0-81.6%) and a specificity of 98.8% (95% CI: 93.5–100%) in clinical diagnosis. When evaluated against a composite reference standard, TNPseq demonstrated a sensitivity of 74.4% (95% CI: 69.3–79.5%) and a specificity of 98.8% (95% CI: 93.7–100%). These results exceed the performance of both the T-SPOT and Xpert MTB/RIF tests. Notably, TNPseq demonstrated high specificity and accuracy in puncture specimens, with a sensitivity of 75.0% (95% CI: 70.2–79.8%) and a specificity of 98.3% (95% CI: 92.7–100%), as well as in pus samples, with a sensitivity of 83.3% (95% CI: 78.6–88.1%) and a specificity of 100% (95% CI: 100–100%). Additionally, TNPseq facilitated the detection of mixed infection scenarios, identifying 20 cases of bacterial-fungal co-infection, 17 cases of bacterial-viral co-infection, and two cases of simultaneous bacterial-fungal-viral co-infection. TNPseq demonstrated great potential in the diagnosis of BJTB due to its high sensitivity and specificity. The ability of TNPseq to diagnose pathogens and detect drug resistance genes can also guide subsequent treatment. Expanding the application scenarios and scope of TNPseq will enable it to benefit more clinical treatments.
- 22 August 2024
- Published by Springer Nature in Annals of Clinical Microbiology and Antimicrobials
- Vol. 23 (1) , 1-10
Open AccessEvaluation of targeted sequencing for pathogen identification in bone and joint infections: a cohort study from China
Bone and joint tuberculosis (BJTB) is a distinct variant of tuberculosis in which clinical diagnosis often leads to relative misdiagnosis and missed diagnoses. This study aimed to evaluate the diagnostic accuracy of the targeted nanopore sequencing (TNPseq) assay for BJTB patients in China. The study enrolled a cohort of 163 patients with suspected BJTB. Diagnostic testing was performed using the TNPseq assay on samples including punctured tissue, pus, and blood. The diagnostic accuracy of the TNPseq assay was then compared with that of the T-SPOT and Xpert MTB/RIF assays. TNPseq exhibited superior performance in terms of accuracy, demonstrating a sensitivity of 76.3% (95% CI: 71.0-81.6%) and a specificity of 98.8% (95% CI: 93.5–100%) in clinical diagnosis. When evaluated against a composite reference standard, TNPseq demonstrated a sensitivity of 74.4% (95% CI: 69.3–79.5%) and a specificity of 98.8% (95% CI: 93.7–100%). These results exceed the performance of both the T-SPOT and Xpert MTB/RIF tests. Notably, TNPseq demonstrated high specificity and accuracy in puncture specimens, with a sensitivity of 75.0% (95% CI: 70.2–79.8%) and a specificity of 98.3% (95% CI: 92.7–100%), as well as in pus samples, with a sensitivity of 83.3% (95% CI: 78.6–88.1%) and a specificity of 100% (95% CI: 100–100%). Additionally, TNPseq facilitated the detection of mixed infection scenarios, identifying 20 cases of bacterial-fungal co-infection, 17 cases of bacterial-viral co-infection, and two cases of simultaneous bacterial-fungal-viral co-infection. TNPseq demonstrated great potential in the diagnosis of BJTB due to its high sensitivity and specificity. The ability of TNPseq to diagnose pathogens and detect drug resistance genes can also guide subsequent treatment. Expanding the application scenarios and scope of TNPseq will enable it to benefit more clinical treatments.
- Open Access
- 18 October 2023
- Published by Springer Nature in BMC Infectious Diseases
- Vol. 23 (1) , 1-8
Evaluation of different diagnostic methods for spinal tuberculosis infection
Background and purpose: Tuberculosis (TB) is the most fatal infectious disease worldwide. Approximately 24.6% of tuberculosis cases are extrapulmonary and predominantly affect the spine. It is difficult to diagnose spinal TB (STB). We aimed to evaluate the diagnostic performance of the Mycobacteria Growth Indicator Tube (MGIT)-960 culture, T-SPOT.TB, Xpert Mycobacterium tuberculosis complex (MTB)/resistance to rifampin (RIF), and Metagenomic Next-Generation Sequencing (mNGS) to detect STB. Methods: We assessed 126 patients presumed to have STB using these four methods. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated using clinical diagnosis as a reference. Results: Of the patients, 41 were diagnosed with STB and 85 with non-STB. In the STB group, the sensitivity, specificity, PPV, and NPV of the MGIT-960 culture were 29.3% (12/41), 100% (85/85), 100% (12/12), and 74.6% (85/114), respectively. The sensitivity, specificity, PPV, and NPV of T-SPOT.TB were 92.7% (38/41), 82.4% (70/85), 58.5% (31/53), and 95.9% (70/73), respectively. The sensitivity, specificity, PPV, and NPV of the Xpert MTB/RIF assay were 53.7% (22/41), 100% (85/85), 100% (22/22), and 81.7% (85/104), respectively. The sensitivity, specificity, PPV, and NPV of mNGS were 39.0% (16/41), 98.8% (84/85), 94.1% (16/17), and 77.1% (84/109), respectively. The sensitivity, specificity, PPV, and NPV of mNGS + Xpert MTB/RIF were 73.2% (30/41), 100% (85/85), 96.8% (30/31), and 72.0% (85/118), respectively. The sensitivity, specificity, PPV, and NPV of the mNGS + T-spot assay were 97.6% (40/41), 100% (85/85), 67.9% (38/56), and 75.9% (85/113), respectively. Moreover, the sensitivity, specificity, PPV, and NPV of T-spot + Xpert MTB/RIF were 95.1% (39/41), 100% (85/85), 72.2% (39/54), and 81.0% (85/105), respectively. Conclusions: T-SPOT.TB is the most effective method for diagnosing STB; however, Xpert MTB/RIF is more reliable and can detect RIF resistance. Clinicians can use mNGS to identify pathogens in patients with spinal infections; these pathogens appeared to be more meaningful in guiding the clinical management of patients in the non-STB group. The combination of Xpert MTB/RIF and mNGS can improve the early diagnosis rate and drug resistance detection, reduce the diagnostic cycle, and provide early targeted anti-TB treatment for patients with STB.
- 18 October 2023
- Published by Springer Nature in BMC Infectious Diseases
- Vol. 23 (1) , 1-8
Open AccessEvaluation of different diagnostic methods for spinal tuberculosis infection
Background and purpose: Tuberculosis (TB) is the most fatal infectious disease worldwide. Approximately 24.6% of tuberculosis cases are extrapulmonary and predominantly affect the spine. It is difficult to diagnose spinal TB (STB). We aimed to evaluate the diagnostic performance of the Mycobacteria Growth Indicator Tube (MGIT)-960 culture, T-SPOT.TB, Xpert Mycobacterium tuberculosis complex (MTB)/resistance to rifampin (RIF), and Metagenomic Next-Generation Sequencing (mNGS) to detect STB. Methods: We assessed 126 patients presumed to have STB using these four methods. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated using clinical diagnosis as a reference. Results: Of the patients, 41 were diagnosed with STB and 85 with non-STB. In the STB group, the sensitivity, specificity, PPV, and NPV of the MGIT-960 culture were 29.3% (12/41), 100% (85/85), 100% (12/12), and 74.6% (85/114), respectively. The sensitivity, specificity, PPV, and NPV of T-SPOT.TB were 92.7% (38/41), 82.4% (70/85), 58.5% (31/53), and 95.9% (70/73), respectively. The sensitivity, specificity, PPV, and NPV of the Xpert MTB/RIF assay were 53.7% (22/41), 100% (85/85), 100% (22/22), and 81.7% (85/104), respectively. The sensitivity, specificity, PPV, and NPV of mNGS were 39.0% (16/41), 98.8% (84/85), 94.1% (16/17), and 77.1% (84/109), respectively. The sensitivity, specificity, PPV, and NPV of mNGS + Xpert MTB/RIF were 73.2% (30/41), 100% (85/85), 96.8% (30/31), and 72.0% (85/118), respectively. The sensitivity, specificity, PPV, and NPV of the mNGS + T-spot assay were 97.6% (40/41), 100% (85/85), 67.9% (38/56), and 75.9% (85/113), respectively. Moreover, the sensitivity, specificity, PPV, and NPV of T-spot + Xpert MTB/RIF were 95.1% (39/41), 100% (85/85), 72.2% (39/54), and 81.0% (85/105), respectively. Conclusions: T-SPOT.TB is the most effective method for diagnosing STB; however, Xpert MTB/RIF is more reliable and can detect RIF resistance. Clinicians can use mNGS to identify pathogens in patients with spinal infections; these pathogens appeared to be more meaningful in guiding the clinical management of patients in the non-STB group. The combination of Xpert MTB/RIF and mNGS can improve the early diagnosis rate and drug resistance detection, reduce the diagnostic cycle, and provide early targeted anti-TB treatment for patients with STB.