Machine learning for image analysis in the cervical spine: Systematic review of the available models and methods

Deep learning model for automated diagnosis of degenerative cervical spondylosis and altered spinal cord signal on MRI

BACKGROUND CONTEXT

A deep learning (DL) model for degenerative cervical spondylosis on MRI could enhance reporting consistency and efficiency, addressing a significant global health issue.

PURPOSE

Create a DL model to detect and classify cervical cord signal abnormalities, spinal canal and neural foraminal stenosis.

STUDY DESIGN/SETTING

Retrospective study conducted from January 2013 to July 2021, excluding cases with instrumentation.

PATIENT SAMPLE

Overall, 504 MRI cervical spines were analyzed (504 patients, mean=58 years±13.7[SD]; 202 women) with 454 for training (90%) and 50 (10%) for internal testing. In addition, 100 MRI cervical spines were available for external testing (100 patients, mean=60 years±13.0[SD];26 women).

OUTCOME MEASURES

Automated detection and classification of spinal canal stenosis, neural foraminal stenosis, and cord signal abnormality using the DL model. Recall(%), inter-rater agreement (Gwet's kappa), sensitivity, and specificity were calculated.

METHODS

Utilizing axial T2-weighted gradient echo and sagittal T2-weighted images, a transformer-based DL model was trained on data labeled by an experienced musculoskeletal radiologist (12 years of experience). Internal testing involved data labeled in consensus by 2 musculoskeletal radiologists (reference standard, both with 12-years-experience), 2 subspecialist radiologists, and 2 in-training radiologists. External testing was performed.

RESULTS

The DL model exhibited substantial agreement surpassing all readers in all classes for spinal canal (κ=0.78, p<.001 vs κ range=0.57-0.70 for readers) and neural foraminal stenosis (κ=0.80, p<.001 vs κ range=0.63-0.69 for readers) classification. The DL model's recall for cord signal abnormality (92.3%) was similar to all readers (range: 92.3-100.0%). Nearly perfect agreement was demonstrated for binary classification (grades 0/1 vs 2/3) (κ=0.95, p<.001 for spinal canal; κ=0.90, p<.001 for neural foramina). External testing showed substantial agreement using all classes (κ=0.76, p<.001 for spinal canal; κ=0.66, p<.001 for neural foramina) and high recall for cord signal abnormality (91.9%). The DL model demonstrated high sensitivities (range:83.7%-92.4%) and specificities (range:87.8%-98.3%) on both internal and external datasets for spinal canal and neural foramina classification.

CONCLUSIONS

Our DL model for degenerative cervical spondylosis on MRI showed good performance, demonstrating substantial agreement with the reference standard. This tool could assist radiologists in improving the efficiency and consistency of MRI cervical spondylosis assessments in clinical practice.

Artificial Intelligence Detection of Cervical Spine Fractures Using Convolutional Neural Network Models

OBJECTIVE

To develop and evaluate a technique using convolutional neural networks (CNNs) for the computer-assisted diagnosis of cervical spine fractures from radiographic x-ray images. By leveraging deep learning techniques, the study might potentially lead to improved patient outcomes and clinical decision-making.

METHODS

This study obtained 500 lateral radiographic cervical spine x-ray images from standard open-source dataset repositories to develop a classification model using CNNs. All the images contained diagnostic information, including normal cervical radiographic images (n=250) and fracture images of the cervical spine fracture (n=250). The model would classify whether the patient had a cervical spine fracture or not. Seventy percent of the images were training data sets used for model training, and 30% were for testing. Konstanz Information Miner (KNIME)'s graphic user interface-based programming enabled class label annotation, data preprocessing, CNNs model training, and performance evaluation.

RESULTS

The performance evaluation of a model for detecting cervical spine fractures presents compelling results across various metrics. This model exhibits high sensitivity (recall) values of 0.886 for fractures and 0.957 for normal cases, indicating its proficiency in identifying true positives. Precision values of 0.954 for fractures and 0.893 for normal cases highlight the model's ability to minimize false positives. With specificity values of 0.957 for fractures and 0.886 for normal cases, the model effectively identifies true negatives. The overall accuracy of 92.14% highlights its reliability in correctly classifying cases by the area under the receiver operating characteristic curve.

CONCLUSION

We successfully used deep learning models for computer-assisted diagnosis of cervical spine fractures from radiographic x-ray images. This approach can assist the radiologist in screening, detecting, and diagnosing cervical spine fractures.

Accuracy of Artificial Intelligence for Cervical Vertebral Maturation Assessment-A Systematic Review

Background/Objectives: To systematically review and summarize the existing scientific evidence on the diagnostic performance of artificial intelligence (AI) in assessing cervical vertebral maturation (CVM). This review aimed to evaluate the accuracy and reliability of AI algorithms in comparison to those of experienced clinicians. Methods: Comprehensive searches were conducted across multiple databases, including PubMed, Scopus, Web of Science, and Embase, using a combination of Boolean operators and MeSH terms. The inclusion criteria were cross-sectional studies with neural network research, reporting diagnostic accuracy, and involving human subjects. Data extraction and quality assessment were performed independently by two reviewers, with a third reviewer resolving any disagreements. The Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-2 tool was used for bias assessment. Results: Eighteen studies met the inclusion criteria, predominantly employing supervised learning techniques, especially convolutional neural networks (CNNs). The diagnostic accuracy of AI models for CVM assessment varied widely, ranging from 57% to 95%. The factors influencing accuracy included the type of AI model, training data, and study methods. Geographic concentration and variability in the experience of radiograph readers also impacted the results. Conclusions: AI has considerable potential for enhancing the accuracy and reliability of CVM assessments in orthodontics. However, the variability in AI performance and the limited number of high-quality studies suggest the need for further research.

Machine learning models for classifying non-specific neck pain using craniocervical posture and movement

OBJECTIVE

Physical therapists and clinicians commonly confirm craniocervical posture (CCP), cervical retraction, and craniocervical flexion as screening tests because they contribute to non-specific neck pain (NSNP). We compared the predictive performance of statistical machine learning (ML) models for classifying individuals with and without NSNP using datasets containing CCP and cervical kinematics during pro- and retraction (CKdPR).

DESIGN

Exploratory, cross-sectional design.

SETTING AND PARTICIPANTS

In total, 773 public service office workers (PSOWs) were screened for eligibility (NSNP, 441; without NSNP, 332).

METHODS

We set up five datasets (CCP, cervical kinematics during the protraction, cervical kinematics during the retraction, CKdPR and combination of the CCP and CKdPR). Four ML algorithms-random forest, logistic regression, Extreme Gradient boosting, and support vector machine-were trained.

MAIN OUTCOME MEASURES

Model performance were assessed using area under the curve (AUC), accuracy, precision, recall and F1-score. To interpret the predictions, we used Feature permutation importance and SHapley Additive explanation values.

RESULTS

The random forest model in the CKdPR dataset classified PSOWs with and without NSNP and achieved the best AUC among the five datasets using the test data (AUC, 0.892 [good]; F1, 0.832). The random forest model in the CCP dataset had the worst AUC among the five datasets using the test data [AUC, 0.738 (fair); F1, 0.715].

CONCLUSION

ML performance was higher for the CKdPR dataset than for the CCP dataset, suggesting that ML algorithms are more suitable than classical statistical methods for developing robust models for classifying PSOWs with and without NSNP.