A Deep-Learning-Based, Fully Automated Program to Segment and Quantify Major Spinal Components on Axial Lumbar Spine Magnetic Resonance Images

Paraspinal muscle endurance and morphology (PMEM) score: a new method for prediction of postoperative mechanical complications after lumbar fusion

BACKGROUND CONTEXT

Although the relationships between paraspinal muscles and lumbar degenerative disorders have been acknowledged, paraspinal muscle evaluation has not been incorporated into clinical therapies.

PURPOSE

We aimed to establish a novel paraspinal muscle endurance and morphology (PMEM) score to better predict mechanical complications after lumbar fusion.

STUDY DESIGN

Prospective cohort study.

PATIENT SAMPLE

A total of 212 patients undergoing posterior lumbar interbody fusion with at least 1 year of follow-up were finally included.

OUTCOME MEASURES

Mechanical complications including screw loosening, pseudarthrosis and other complications like cage subsidence, and patient-reported outcomes were evaluated at last follow-up.

METHODS

The PMEM score comprised 1 functional muscular parameter (the performance time of the endurance test) and 2 imaging muscular parameters (relative functional cross-sectional area [rFCSA] of paraspinal extensor muscles [PEM] and psoas major [PS] on magnetic resonance imaging). The score was established based on a weighted scoring system created by rounding β regression coefficients to the nearest integer in univariate logistic regression. The diagnostic performance of the PMEM score was determined by binary logistic regression model and receiver operating characteristic (ROC) curve with the area under the curve (AUC). Additionally, pairwise comparisons of ROC curves were conducted to compare the diagnostic performance of the PMEM score with conventional methods based on a single muscular parameter. Moreover, differences of mechanical complications and patient-reported outcomes among the PMEM categories were analyzed using Chi-square test with Bonferroni correction.

RESULTS

The PMEM score, calculated by adding the scores for each parameter, ranges from 0 to 5 points. Patients with higher PMEM scores exhibited higher rates of mechanical complications (p<.001). Binary logistic regression revealed that the PMEM score was an independent factor of mechanical complications (p<.001, OR=2.002). Moreover, the AUC of the PMEM score (AUC=0.756) was significantly greater than those of the conventional methods including the endurance test (AUC=0.691, Z=2.036, p<.05), PEM rFCSA (AUC=.690, Z=2.016, p<.05) and PS rFCSA (AUC=0.640, Z=2.771, p<.01). In terms of the PMEM categories, a score of 0-1 was categorized as low-risk muscular state of mechanical complications; 2-3, as moderate; and 4-5, as high-risk state. Moving from the low-risk state to the high-risk state, there was a progressive increase in the rates of mechanical complications (13.8% vs. 32.1% vs. 72.7%; p<.001), and a decrease in the rates of clinically significant improvement of patient-reported outcomes (all p<.05).

CONCLUSIONS

The PMEM score might comprehensively evaluate paraspinal muscle degeneration and exhibit greater ability in predicting mechanical complications than the conventional evaluations after lumbar fusion. Surgeons might develop individualized treatment strategy tailored to different muscle degeneration statuses reflected by the PMEM score for decreasing the risk of mechanical complications.

The Quantitative Evaluation of Automatic Segmentation in Lumbar Magnetic Resonance Images

OBJECTIVE

This study aims to overcome challenges in lumbar spine imaging, particularly lumbar spinal stenosis, by developing an automated segmentation model using advanced techniques. Traditional manual measurement and lesion detection methods are limited by subjectivity and inefficiency. The objective is to create an accurate and automated segmentation model that identifies anatomical structures in lumbar spine magnetic resonance imaging scans.

METHODS

Leveraging a dataset of 539 lumbar spinal stenosis patients, the study utilizes the residual U-Net for semantic segmentation in sagittal and axial lumbar spine magnetic resonance images. The model, trained to recognize specific tissue categories, employs a geometry algorithm for anatomical structure quantification. Validation metrics, like Intersection over Union (IOU) and Dice coefficients, validate the residual U-Net's segmentation accuracy. A novel rotation matrix approach is introduced for detecting bulging discs, assessing dural sac compression, and measuring yellow ligament thickness.

RESULTS

The residual U-Net achieves high precision in segmenting lumbar spine structures, with mean IOU values ranging from 0.82 to 0.93 across various tissue categories and views. The automated quantification system provides measurements for intervertebral disc dimensions, dural sac diameter, yellow ligament thickness, and disc hydration. Consistency between training and testing datasets assures the robustness of automated measurements.

CONCLUSION

Automated lumbar spine segmentation with residual U-Net and deep learning exhibits high precision in identifying anatomical structures, facilitating efficient quantification in lumbar spinal stenosis cases. The introduction of a rotation matrix enhances lesion detection, promising improved diagnostic accuracy, and supporting treatment decisions for lumbar spinal stenosis patients.

Deep Learning Model for Grading and Localization of Lumbar Disc Herniation on Magnetic Resonance Imaging

BACKGROUND

Methods for grading and localization of lumbar disc herniation (LDH) on MRI are complex, time-consuming, and subjective. Utilizing deep learning (DL) models as assistance would mitigate such complexities.

PURPOSE

To develop an interpretable DL model capable of grading and localizing LDH.

STUDY TYPE

Retrospective.

SUBJECTS

1496 patients (M/F: 783/713) were evaluated, and randomly divided into training (70%), validation (10%), and test (20%) sets.

FIELD STRENGTH/SEQUENCE

1.5T MRI for axial T2-weighted sequences (spin echo).

ASSESSMENT

The training set was annotated by three spinal surgeons using the Michigan State University classification to train the DL model. The test set was annotated by a spinal surgery expert (as ground truth labels), and two spinal surgeons (comparison with the trained model). An external test set was employed to evaluate the generalizability of the DL model.

STATISTICAL TESTS

Calculated intersection over union (IoU) for detection consistency, utilized Gwet's AC1 to assess interobserver agreement, and evaluated model performance based on sensitivity and specificity, with statistical significance set at P < 0.05.

RESULTS

The DL model achieved high detection consistency in both the internal test dataset (grading: mean IoU 0.84, recall 99.6%; localization: IoU 0.82, recall 99.5%) and external test dataset (grading: 0.72, 98.0%; localization: 0.71, 97.6%). For internal testing, the DL model (grading: 0.81; localization: 0.76), Rater 1 (0.88; 0.82), and Rater 2 (0.86; 0.83) demonstrated results highly consistent with the ground truth labels. The overall sensitivity of the DL model was 87.0% for grading and 84.0% for localization, while the specificity was 95.5% and 94.4%. For external testing, the DL model showed an appreciable decrease in consistency (grading: 0.69; localization: 0.66), sensitivity (77.2%; 76.7%), and specificity (92.3%; 91.8%).

DATA CONCLUSION

The classification capabilities of the DL model closely resemble those of spinal surgeons. For future improvement, enriching the diversity of cases could enhance the model's generalization.

LEVEL OF EVIDENCE: 4

TECHNICAL EFFICACY

Stage 2.

Three-dimensional freehand ultrasonography to measure muscle volume of the lumbar multifidus: Reliability of processing technique and validity through comparison to magnetic resonance imaging

There is a growing interest in muscle characteristics of the lumbar multifidus related to low back pain, but findings between studies are inconsistent. One of the issues explaining these conflicting findings might be the use of two-dimensional measures of cross-sectional area and thickness of the lumbar multifidus in most studies, which might be a suboptimal representation of the entire muscle volume. A three-dimensional volumetric assessment, combined with standardized imaging and processing measurement protocols, is highly recommended to quantify spinal muscle morphology. Three-dimensional freehand ultrasonography is a technique with large potential for daily clinical practice. It is achieved by combining conventional two-dimensional ultrasound with a motion-tracking system, recording the position and orientation of the ultrasound transducer during acquisition, resulting in a three-dimensional reconstruction. This study investigates intra- and interprocessor reliability for the quantification of muscle volume of the lumbar multifidus based on three-dimensional freehand ultrasound and its validity, in 31 patients with low back pain and 20 healthy subjects. Two processors manually segmented the lumbar multifidus on three-dimensional freehand ultrasound images using Stradwin software following a well-defined method. We assessed the concurrent validity of the measurement of multifidus muscle volume using three-dimensional freehand ultrasound compared with magnetic resonance imaging in 10 patients with low back pain. Processing reliability and agreement were determined using intraclass correlation coefficients, Bland-Altman plots, and calculation of the standard error of measurement and minimal detectable change, while validity was defined based on correlation analysis. The processing of three-dimensional freehand ultrasound images to measure lumbar multifidus volume was reliable. Good to excellent intraclass correlation coefficients were found for intraprocessor reliability. For interprocessor reliability, the intraclass correlation coefficients were moderate to good, emphasizing the importance of processing guidelines and training. A single processor analysis is preferred in clinical studies or when small differences in muscle volume are expected. The correlation between magnetic resonance imaging and three-dimensional freehand ultrasound measurements of lumbar multifidus volume was moderate to good but with a systematically smaller multifidus volume measured on three-dimensional freehand ultrasound. These results provide opportunities for both researchers and clinicians to reliably assess muscle structure using three-dimensional freehand ultrasound in patients with low back pain and to monitor changes related to pathology or interventions. To allow implementation in both research and clinical settings, guidelines on three-dimensional freehand ultrasound processing and training were provided.

SpineQ: Unsupervised 3D Lumbar Quantitative Assessment

Most lumbar quantitative assessment methods can only analyze the image from one view and require laborious manual annotation. We aim to develop an unsupervised pipeline for 3D quantitative assessment of the lumbar spine that can assess the MRI with different views. We combine rule-based and deep learning methods to generate multi-tissue segmentation, and parameters can be measured from segmentation results using the anatomical and geometric prior. Preliminary testing demonstrates that our proposed method can generate accurate segmentation and measurement results.Clinical Relevance- The proposed unsupervised 3D lumbar quantitative assessment pipeline can significantly improve the efficiency and consistency of clinical diagnosis and surgical planning.

Deep learning-based automatic sella turcica segmentation and morphology measurement in X-ray images

BACKGROUND

Although the morphological changes of sella turcica have been drawing increasing attention, the acquirement of linear parameters of sella turcica relies on manual measurement. Manual measurement is laborious, time-consuming, and may introduce subjective bias. This paper aims to develop and evaluate a deep learning-based model for automatic segmentation and measurement of sella turcica in cephalometric radiographs.

METHODS

1129 images were used to develop a deep learning-based segmentation network for automatic sella turcica segmentation. Besides, 50 images were used to test the generalization ability of the model. The performance of the segmented network was evaluated by the dice coefficient. Images in the test datasets were segmented by the trained segmentation network, and the segmentation results were saved in binary images. Then the extremum points and corner points were detected by calling the function in the OpenCV library to obtain the coordinates of the four landmarks of the sella turcica. Finally, the length, diameter, and depth of the sella turcica can be obtained by calculating the distance between the two points and the distance from the point to the straight line. Meanwhile, images were measured manually using Digimizer. Intraclass correlation coefficients (ICCs) and Bland-Altman plots were used to analyze the consistency between automatic and manual measurements to evaluate the reliability of the proposed methodology.

RESULTS

The dice coefficient of the segmentation network is 92.84%. For the measurement of sella turcica, there is excellent agreement between the automatic measurement and the manual measurement. In Test1, the ICCs of length, diameter and depth are 0.954, 0.953, and 0.912, respectively. In Test2, ICCs of length, diameter and depth are 0.906, 0.921, and 0.915, respectively. In addition, Bland-Altman plots showed the excellent reliability of the automated measurement method, with the majority measurements differences falling within ± 1.96 SDs intervals around the mean difference and no bias was apparent.

CONCLUSIONS

Our experimental results indicated that the proposed methodology could complete the automatic segmentation of the sella turcica efficiently, and reliably predict the length, diameter, and depth of the sella turcica. Moreover, the proposed method has generalization ability according to its excellent performance on Test2.

Artificial intelligence in spine surgery

The continuous progress of research and clinical trials has offered a wide variety of information concerning the spine and the treatment of the different spinal pathologies that may occur. Planning the best therapy for each patient could be a very difficult and challenging task as it often requires thorough processing of the patient's history and individual characteristics by the clinician. Clinicians and researchers also face problems when it comes to data availability due to patients' personal information protection policies. Artificial intelligence refers to the reproduction of human intelligence via special programs and computers that are trained in a way that simulates human cognitive functions. Artificial intelligence implementations to daily clinical practice such as surgical robots that facilitate spine surgery and reduce radiation dosage to medical staff, special algorithms that can predict the possible outcomes of conservative versus surgical treatment in patients with low back pain and disk herniations, and systems that create artificial populations with great resemblance and similar characteristics to real patients are considered to be a novel breakthrough in modern medicine. To enhance the body of the related literature and inform the readers on the clinical applications of artificial intelligence, we performed this review to discuss the contribution of artificial intelligence in spine surgery and pathology.

An update on animal models of intervertebral disc degeneration and low back pain: Exploring the potential of artificial intelligence to improve research analysis and development of prospective therapeutics

Animal models have been invaluable in the identification of molecular events occurring in and contributing to intervertebral disc (IVD) degeneration and important therapeutic targets have been identified. Some outstanding animal models (murine, ovine, chondrodystrophoid canine) have been identified with their own strengths and weaknesses. The llama/alpaca, horse and kangaroo have emerged as new large species for IVD studies, and only time will tell if they will surpass the utility of existing models. The complexity of IVD degeneration poses difficulties in the selection of the most appropriate molecular target of many potential candidates, to focus on in the formulation of strategies to effect disc repair and regeneration. It may well be that many therapeutic objectives should be targeted simultaneously to effect a favorable outcome in human IVD degeneration. Use of animal models in isolation will not allow resolution of this complex issue and a paradigm shift and adoption of new methodologies is required to provide the next step forward in the determination of an effective repairative strategy for the IVD. AI has improved the accuracy and assessment of spinal imaging supporting clinical diagnostics and research efforts to better understand IVD degeneration and its treatment. Implementation of AI in the evaluation of histology data has improved the usefulness of a popular murine IVD model and could also be used in an ovine histopathological grading scheme that has been used to quantify degenerative IVD changes and stem cell mediated regeneration. These models are also attractive candidates for the evaluation of novel anti-oxidant compounds that counter inflammatory conditions in degenerate IVDs and promote IVD regeneration. Some of these compounds also have pain-relieving properties. AI has facilitated development of facial recognition pain assessment in animal IVD models offering the possibility of correlating the potential pain alleviating properties of some of these compounds with IVD regeneration.

Coarse X-ray Lumbar Vertebrae Pose Localization and Registration Using Triangulation Correspondence

Plain film X-ray scanners are indispensable for medical diagnostics and clinical procedures. This type of device typically produces two radiographic images of the human spine, including the anteroposterior and lateral views. However, these two photographs presented perspectives that were distinct. The proposed procedure consists of three fundamental steps. For automated cropping, the grayscale lumbar input image was initially projected vertically using its vertical pattern. Then, Delaunay triangulation was performed with the SURF features serving as the triangle nodes. The posture area of the vertebrae was calculated by utilizing the edge density of each node. The proposed method provided an automated estimation of the position of the human lumbar vertebrae, thereby decreasing the radiologist’s workload, computing time, and complexity in a variety of bone-clinical applications. Numerous applications can be supported by the results of the proposed method, including the segmentation of lumbar vertebrae pose, bone mineral density examination, and vertebral pose deformation. The proposed method can estimate the vertebral position with an accuracy of 80.32 percent, a recall rate of 85.37 percent, a precision rate of 82.36%, and a false-negative rate of 15.42 percent.

Convolutional Neural Networks in Spinal Magnetic Resonance Imaging: A Systematic Review

OBJECTIVE

Convolutional neural networks (CNNs) are being increasingly used in the medical field, especially for image recognition in high-resolution, large-volume data sets. The study represents the current state of research on the application of CNNs in image segmentation and pathology detection in spine magnetic resonance imaging.

METHODS

For this systematic literature review, the authors performed a systematic initial search of the PubMed/Medline and Web of Science (Core collection) databases for eligible investigations. The authors limited the search to observational studies. Outcome parameters were analyzed according to the inclusion criteria and assigned to 3 groups: 1) segmentation of anatomical structures, 2) segmentation and evaluation of pathologic structures, and 3) specific implementation of CNNs.

RESULTS

Twenty-four retrospectively designed articles met the inclusion criteria. Publication dates ranged from 2017 to 2021. In total, 14,065 patients with 113,110 analyzed images were included. Most authors trained their network with a training-to-testing ratio of 80/20, while all but 2 articles used 5- to 10-fold cross-validation. Nine articles compared their performance results with other neural networks and algorithms, and all 24 articles described outcomes as positive.

CONCLUSIONS

State-of-the-art CNNs can detect and segment-specific anatomical landmarks and pathologies across a wide range, comparable to the skills of radiologists and experienced clinicians. With rapidly evolving network architectures and growing medical image databases, the future is likely to show growth in the development and refinement of these capable networks. However, the aid of automated segmentation and classification by neural networks cannot and should not be expected to replace clinical experts.

Convolutional neural networks for the automatic segmentation of lumbar paraspinal muscles in people with low back pain

The size, shape, and composition of paraspinal muscles have been widely reported in disorders of the cervical and lumbar spine. Measures of size, shape, and composition have required time-consuming and rater-dependent manual segmentation techniques. Convolutional neural networks (CNNs) provide alternate timesaving, state-of-the-art performance measures, which could realise clinical translation. Here we trained a CNN for the automatic segmentation of lumbar paraspinal muscles and determined the impact of CNN architecture and training choices on segmentation performance. T₂-weighted MRI axial images from 76 participants (46 female; age (SD): 45.6 (12.8) years) with low back pain were used to train CNN models to segment the multifidus, erector spinae, and psoas major muscles (left and right segmented separately). Using cross-validation, we compared 2D and 3D CNNs with and without data augmentation. Segmentation accuracy was compared between the models using the Sørensen-Dice index as the primary outcome measure. The effect of increasing network depth on segmentation accuracy was also investigated. Each model showed high segmentation accuracy (Sørensen-Dice index ≥ 0.885) and excellent reliability (ICC_2,1 ≥ 0.941). Overall, across all muscles, 2D models performed better than 3D models (p = 0.012), and training without data augmentation outperformed training with data augmentation (p < 0.001). The 2D model trained without data augmentation demonstrated the highest average segmentation accuracy. Increasing network depth did not improve accuracy (p = 0.771). All trained CNN models demonstrated high accuracy and excellent reliability for segmenting lumbar paraspinal muscles. CNNs can be used to efficiently and accurately extract measures of paraspinal muscle health from MRI.

An externally validated deep learning model for the accurate segmentation of the lumbar paravertebral muscles

PURPOSE

Imaging studies about the relevance of muscles in spinal disorders, and sarcopenia in general, require the segmentation of the muscles in the images which is very labour-intensive if performed manually and poses a practical limit to the number of investigated subjects. This study aimed at developing a deep learning-based tool able to fully automatically perform an accurate segmentation of the lumbar muscles in axial MRI scans, and at validating the new tool on an external dataset.

METHODS

A set of 60 axial MRI images of the lumbar spine was retrospectively collected from a clinical database. Psoas major, quadratus lumborum, erector spinae, and multifidus were manually segmented in all available slices. The dataset was used to train and validate a deep neural network able to segment muscles automatically. Subsequently, the network was externally validated on images purposely acquired from 22 healthy volunteers.

RESULTS

The median Jaccard index for the individual muscles calculated for the 22 subjects of the external validation set ranged between 0.862 and 0.935, demonstrating a generally excellent performance of the network, although occasional failures were noted. Cross-sectional area and fat fraction of the muscles were in agreement with published data.

CONCLUSIONS

The externally validated deep neural network was able to perform the segmentation of the paravertebral muscles in an accurate and fully automated manner, although it is not without limitations. The model is therefore a suitable research tool to perform large-scale studies in the field of spinal disorders and sarcopenia, overcoming the limitations of non-automated methods.

Artificial Intelligence and Computer Vision in Low Back Pain: A Systematic Review

Chronic Low Back Pain (LBP) is a symptom that may be caused by several diseases, and it is currently the leading cause of disability worldwide. The increased amount of digital images in orthopaedics has led to the development of methods related to artificial intelligence, and to computer vision in particular, which aim to improve diagnosis and treatment of LBP. In this manuscript, we have systematically reviewed the available literature on the use of computer vision in the diagnosis and treatment of LBP. A systematic research of PubMed electronic database was performed. The search strategy was set as the combinations of the following keywords: "Artificial Intelligence", "Feature Extraction", "Segmentation", "Computer Vision", "Machine Learning", "Deep Learning", "Neural Network", "Low Back Pain", "Lumbar". Results: The search returned a total of 558 articles. After careful evaluation of the abstracts, 358 were excluded, whereas 124 papers were excluded after full-text examination, taking the number of eligible articles to 76. The main applications of computer vision in LBP include feature extraction and segmentation, which are usually followed by further tasks. Most recent methods use deep learning models rather than digital image processing techniques. The best performing methods for segmentation of vertebrae, intervertebral discs, spinal canal and lumbar muscles achieve Sørensen-Dice scores greater than 90%, whereas studies focusing on localization and identification of structures collectively showed an accuracy greater than 80%. Future advances in artificial intelligence are expected to increase systems' autonomy and reliability, thus providing even more effective tools for the diagnosis and treatment of LBP.

Multi-muscle deep learning segmentation to automate the quantification of muscle fat infiltration in cervical spine conditions

Muscle fat infiltration (MFI) has been widely reported across cervical spine disorders. The quantification of MFI requires time-consuming and rater-dependent manual segmentation techniques. A convolutional neural network (CNN) model was trained to segment seven cervical spine muscle groups (left and right muscles segmented separately, 14 muscles total) from Dixon MRI scans (n = 17, 17 scans < 2 weeks post motor vehicle collision (MVC), and 17 scans 12 months post MVC). The CNN MFI measures demonstrated high test reliability and accuracy in an independent testing dataset (n = 18, 9 scans < 2 weeks post MVC, and 9 scans 12 months post MVC). Using the CNN in 84 participants with scans < 2 weeks post MVC (61 females, 23 males, age = 34.2 ± 10.7 years) differences in MFI between the muscle groups and relationships between MFI and sex, age, and body mass index (BMI) were explored. Averaging across all muscles, females had significantly higher MFI than males (p = 0.026). The deep cervical muscles demonstrated significantly greater MFI than the more superficial muscles (p < 0.001), and only MFI within the deep cervical muscles was moderately correlated to age (r > 0.300, p ≤ 0.001). CNN's allow for the accurate and rapid, quantitative assessment of the composition of the architecturally complex muscles traversing the cervical spine. Acknowledging the wider reports of MFI in cervical spine disorders and the time required to manually segment the individual muscles, this CNN may have diagnostic, prognostic, and predictive value in disorders of the cervical spine.

New MR-based measures for the evaluation of age-related lumbar paraspinal muscle degeneration

PURPOSE

Although signal intensity on T2W axial images is sensitive in detection of fatty infiltration to assess paraspinal muscle degeneration, it is affected by inhomogeneities of magnetic fields and individual variabilities. The purpose of this study was to propose reference adjusted signal measures on T2W axial images and determine their capacities in reflecting age-related lumbar paraspinal muscle degeneration.

METHODS

Lumbar MR images of 421 population-based subjects (177 men and 244 women, mean age 53.1 years, range 19.8-87.9 years) were studied. A custom software Spine Explore (Tulong 2.0) was used to automatically obtain paraspinal measurements of multifidus, erector spinae and psoas major. FCSA/TCSA was defined as functional cross-sectional area relative to total cross-sectional area of paraspinal muscle. Two new signal measures were canal-adjusted and cerebrospinal fluid (CSF)-adjusted signal, defined as the ratio between mean signal measurements and the mean signal of the canal and CSF.

RESULTS

The raw signal measurements of the paraspinal muscles were weakly correlated to age (r = 0.28-0.39, P < 0.001). When the signal of canal (r = 0.43-0.59, P < 0.001) or CSF (r = 0.45-0.61, P < 0.001) was used as reference, the correlations substantially increased. Signal measurements of three paraspinal muscles, adjusted or not, were strongly associated with Goutallier score (ρ = 0.60-0.65, P < 0.001) and FCSA/TCSA (r = -0.64 to -0.82, P < 0.001). Greater Goutallier score was associated with greater age (r = 0.38-0.60, P < 0.001), while Lumbar indentation value (LIV) not.

CONCLUSION

On routine T2W axial MR images the adjusted signal measurements using an internal reference of CSF or canal can better reflect age-related degenerative changes in the paraspinal muscles.