Application of Multitask Deep Learning Model in Grading the Severity of Spinal Facet Joint Degeneration

Spinal facet joint osteoarthritis is a disease with high incidence among people over 40 years old. It is a disease characterized by a series of degenerative pathological changes and clinical features of synovium, articular cartilage, subchondral bone, joint space and accessory tissues of spinal facet joints under the action of multiple factors. Some physiological or pathological factors can lead to osteoarthritis of spinal facet joints. Patients with spinal facet osteoarthritis often have different degrees of clinical manifestations such as back pain and dyskinesia, which significantly affect the physical and mental health of patients. The severity of spinal facet osteoarthritis not only has a certain impact on low back pain and changes in low back muscle density, but also affects patient management and treatment plan. At present, different doctors have certain subjectivity in the grading reading of lumbar facet osteoarthritis, and the consistency and repeatability of the results are poor. Moreover, doctors need to read image images and judge the grading is very time-consuming and repetitive work. In recent years, the application of deep learning technology in medical image analysis has been widely concerned by clinicians. Deep learning has great potential benefits in medical imaging diagnosis. It can provide semi-automatic reports under the supervision of radiologists, so as to improve the accuracy, consistency, objectivity and rapidity of disease degree assessment, and further support clinical decision-making on this basis. This project plans to develop an intelligent diagnosis and classification system for degenerative diseases of small joints of the spine with multi task and in-depth learning, and verify its clinical feasibility, aiming to help clinicians improve the accuracy, consistency, objectivity and rapidity of the corresponding disease degree evaluation, and further support the follow-up clinical decision-making.

Study Overview

• Status: Active, not recruiting
• Conditions: Facet Joints; Degeneration ; Deep Learning ;Artificial Intelligence
• Intervention / Treatment: Other

Detailed Description

This project is a retrospective clinical study. From 2020 to 2022, DICOM-format images and basic information of X-ray, CT, and magnetic resonance (MR) images were collected from outpatients and inpatients with suspected low back pain at the Fifth Affiliated Hospital of Sun Yat-sen University. After obtaining the DICOM image mode, data were exported from the information module upon the successful submission of OA batches; basic patient information was collected from inpatient medical records.

This study plans to include 1,132 patients from a single center, who will be randomly divided into a training set, a validation set, and a test set according to the proportion for automatic diagnosis by the computer deep learning model, aiming to test the stability and reliability of the model. Among these 1,132 patients, two doctors separately conducted graded image reading for joint stenosis, hypertrophy, osteophytes, articular surface erosion, and subchondral cysts. Controversial results were determined by another more experienced doctor, and results of the reference standard group were confirmed by the senior doctor group. The data analysis methods for other centers were consistent with those described above.

By comparing the diagnostic results of clinicians and the model, the performance and clinical feasibility of the deep learning model for the automatic diagnosis of lumbar facet joint degeneration were evaluated. The doctors' judgment results were compared with the model's prediction results, and statistical analysis was performed on the performance of the model's automatic diagnosis. Performance evaluation indicators included accuracy, sensitivity, specificity, precision, recall, F1 score, and AUC value. Among them, F1 score and AUC value are the main indicators for the comprehensive evaluation of model performance; the higher their values, the stronger the model performance.

• Study Type: Observational
• Enrollment (Estimated): 1132

Contacts and Locations

Study Locations

• China
  • Guangdong
    • Zhuhai, Guangdong, China
      • The fifth affiliated hospital of SYSU

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

For patients receiving imaging examination due to low back pain, the degree of degeneration of the facet joints of the patients is no to severe. Remove patients meeting the exclusion criteria to avoid poor image quality affecting judgment.

Description

Inclusion Criteria:

- From 2020 to 2023, data of patients receiving lumbar imaging examination in the Fifth Affiliated Hospital of Sun Yat sen University and other hospital.

Exclusion criteria: patients with conditions like spondylolysis, lumbar spondylolisthesis, lumbar spine fractures, previous lumbar surgery, or severe scoliosis.

Study Plan

How is the study designed?

Design Details

• Observational Models: Other
• Time Perspectives: Retrospective

Number of groups / cohorts

3

Cohorts and Interventions

Group / Cohort
Training group; 70% of the participants were randomly divided into training groups to train the learning performance of the machine
Validation group; 15% of the participants were randomly divided into validation groups to enhance the learning performance of the machine and avoid over fitting
Test group; 15% of the participants were randomly divided into test groups to test the learning performance of the machine and draw research conclusions

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
To compare the accuracy of multitask deep learning model and clinicians in judging spinal facet joint degeneration	It is mainly used to indicate the number of correctly predicted samples in the total number of samples.True Positive(TP),False Negative(FN), False Positive(FP),True Negative(TN).Accuracy = (TP + TN) / (TP + FN + FP +TN)	2022.12.01-2023.07.31
To compare the precision of multitask deep learning model and clinicians in judging spinal facet joint degeneration	True Positive(TP),False Negative(FN), False Positive(FP),True Negative(TN).Precision = TP / (TP+FP)	2022.12.01-2023.07.31
To compare the sensitivity of multitask deep learning model and clinicians in assessing spinal facet joint degeneration	True Positive(TP),False Negative(FN), False Positive(FP),True Negative(TN).Sensitivity=TP / (TP+FN)	2022.12.01-2023.07.31
To compare the specificity of multitask deep learning model and clinicians in assessing spinal facet joint degeneration	True Positive(TP),False Negative(FN), False Positive(FP),True Negative(TN).Specificity=TN / (TN+FP)	2022.12.01-2023.07.31
Calculate the F1 score for evaluating the severity of facet joints degeneration in the multitask deep learning model	F1 score is an important evaluation indicator for automatic classification,F1 =2*Precision*Sensitivity/（Precision+Sensitivity）=2TP/(2TP+FP+FN)	2022.12.01-2023.07.31
ROC (Receiver Operation Characteristic) is called receiver operation characteristic curve, which is an index to evaluate the performance of deep learning model	ROC (Receiver Operation Characteristic) is called receiver operation characteristic curve. The closer the curve is to the upper left corner, the better the classifier is. The area under the ROC curve is called AUC. The larger the AUC is, the better the classification effect of the classifier will be.	2022.12.01-2023.07.31

Collaborators and Investigators

• Sponsor: Hai Lv

Study record dates

Study Major Dates

• Study Start (Actual): December 31, 2022
• Primary Completion (Estimated): December 31, 2026
• Study Completion (Estimated): December 31, 2026

Study Registration Dates

• First Submitted: November 23, 2022
• First Submitted That Met QC Criteria: December 1, 2022
• First Posted (Actual): December 2, 2022

Study Record Updates

• Last Update Posted (Actual): August 13, 2025
• Last Update Submitted That Met QC Criteria: August 8, 2025
• Last Verified: August 1, 2025

More Information

Terms related to this study

• Other Study ID Numbers: ZDWY.JZWK.004

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: If necessary, it can be provided

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No
• product manufactured in and exported from the U.S.: No