Novel Application of the Pfirrmann Disc Degeneration Grading System to 9.4T MRI: Higher Reliability Compared to 3T MRI

A fully automatic MRI-guided decision support system for lumbar disc herniation using machine learning

Background

Normalized decision support system for lumbar disc herniation (LDH) will improve reproducibility compared with subjective clinical diagnosis and treatment. Magnetic resonance imaging (MRI) plays an essential role in the evaluation of LDH. This study aimed to develop an MRI-based decision support system for LDH, which evaluates lumbar discs in a reproducible, consistent, and reliable manner.

Methods

The research team proposed a system based on machine learning that was trained and tested by a large, manually labeled data set comprising 217 patients' MRI scans (3255 lumbar discs). The system analyzes the radiological features of identified discs to diagnose herniation and classifies discs by Pfirrmann grade and MSU classification. Based on the assessment, the system provides clinical advice.

Results

Eventually, the accuracy of the diagnosis process reached 95.83%. An 83.5% agreement was observed between the system's prediction and the ground-truth in the Pfirrmann grade. In the case of MSU classification, 95.0% precision was achieved. With the assistance of this system, the accuracy, interpretation efficiency and interrater agreement among surgeons were improved substantially.

Conclusion

This system showed considerable accuracy and efficiency, and therefore could serve as an objective reference for the diagnosis and treatment procedure in clinical practice.

Tracing the disc: The novel qualitative morphometric MRI based disc degeneration classification system

Background

This study aimed to develop a classification system for lumbar disc degeneration using routine magnetic resonance images (MRIs) that is easily applicable and unaffected by existing classifications' limitations, and to compare its reliability, reproducibility, and discriminative power to the widely used Pfirrmann classification.

Methods

Five features were graded. This new classification system has eight grades, with at least one of these five features altering each grade. The T2-weighted sagittal images were acquired using a rapid spin-echo sequence with a repetition time of 2680 to 4900 milliseconds, an echo time of 100 to 109 milliseconds, and an echo train length of 17. Slice thick was 4 mm and the display field of view was 32 × 32 cm. The new classification system used five features: signal intensity, disc height, disc boundary regularity, and nucleus annulus separation. Increased signal intensity, decreased height, decreased regularity, and decreased nucleus-annulus separation indicated degeneration. Four raters classified 400 discs from 80 patients using the Pfirrmann and Novel systems. Statistical analyses were conducted to investigate reliability and correlation.

Results

The overall ICC and kappa values were found to be higher in the novel classification. (0.988 indicating excellent agreement for ICC and 0.76/0.94 indicating good-very good agreement for kappa). The Kendall tau c value, which shows the correlation between the two classifications and indicates the validity of the new classification, was 0.872, which is very strong. Through the use of cross-tabulations, the discriminatory power of the two newly added classification criteria was determined.

Conclusions

This study demonstrates the intra-rater and inter-rater reliability of an easy-to-use, discriminative novel morphometric MRI based classification system for lumbar disc degeneration. The differentiation of grades based on five distinct criteria may generate novel hypotheses regarding treatment selection and response monitoring, as well as new insights into the study of disc degeneration.

The influence of geometry on intervertebral disc stiffness

Geometry plays an important role in intervertebral disc (IVD) mechanics. Previous computational studies have found a link between IVD geometry and stiffness. However, few experimental studies have investigated this link, possibly due to difficulties in non-destructively quantifying internal geometric features. Recent advances in ultra-high resolution MRI provides the opportunity to visualise IVD features in unprecedented detail. This study aimed to quantify 3D human IVD geometries using 9.4 T MRIs and to investigate correlations between geometric variations and IVD stiffness. Thirty human lumbar motion segments (fourteen non-degenerate and sixteen degenerate) were scanned using a 9.4 T MRI and geometric parameters were measured. A 1kN compressive load was applied to each motion segment and stiffness was calculated. Degeneration caused a reduction (p < 0.05) in IVD height, a decreased nucleus-annulus area ratio, and a 1.6 ± 3.0 mm inward collapse of the inner annulus. The IVD height, anteroposterior (AP) width, lateral width, cross-sectional area, nucleus-annulus boundary curvature, and nucleus-annulus area ratio had a significant (p < 0.05) influence on IVD stiffness. Linear relationships (p < 0.05, r > 0.47) were observed between these geometric features and IVD compressive stiffness and a multivariate regression model was generated to enable stiffness to be predicted from features observable on clinical imaging (stiffness, N/mm = 6062 - (61.2 × AP width, mm) - (169.2 × IVD height, mm)). This study advances our understanding of disc structure-function relationships and how these change with degeneration, which can be used to both generate and validate more realistic computational models.

Comparison of magnetic resonance imaging findings among sciatica patients classified as centralizers or non-centralizers

OBJECTIVE

To compare if the degenerative findings from MRI differ between the sciatica patients classified as centralizers (CEN) and non-centralizers (Non-CEN) according to the McKenzie Method of mechanical diagnosis and therapy.

STUDY DESIGN

A cross-sectional study.

METHODS

Patients (N = 100) referred to a spine clinic of a single tertiary hospital for specialist consultation for sciatica. The McKenzie-based assessment was performed by the mechanical diagnosis and therapy-trained physiotherapists. Clinical data and prevalence of lumbar MRI findings were compared between the groups.

RESULTS

There was no significant difference in leg pain intensity between the groups. The Non-CEN had significantly more intense back pain, mean 56 (SD 30) and were more disabled 44 (SD 15) compared to the CEN mean 41 (SD 25) and mean 31 (11), measured with a visual analogue scale (0-100), and the Oswestry Disability Index (0-100), respectively. The CEN had more severe degenerative findings on MRI than the Non-CEN: vertebral end-plate changes were 63% and 43%; mean Pfirrmann's disc degeneration lumbar summary score was 12.8, and 10.6; and severity score of total damage was 12.0 and 10.1, respectively. There were differences neither in disc contour changes nor nerve root stenosis on MRI.

CONCLUSIONS

Sciatica patients classified as non-centralizers had significantly more severe back pain, and were significantly more disabled than centralizers, who instead had more severe degenerative findings on MRI. Thus, classification to non-centralizers by the McKenzie method seems not predict higher incidence of degenerative findings on MRI compared to centralizers.

An in vitro comparison of three nucleus pulposus removal techniques for partial intervertebral disc replacement: An ultra-high resolution MRI study

Background

Nuclectomy, also known as nucleotomy, is a percutaneous surgical procedure performed to remove nucleus material from the center of the disc. Multiple techniques have been considered to perform a nuclectomy, however, the advantages and disadvantages of each are not well understood.

Aims

This in vitro biomechanical investigation on human cadaveric specimens aimed to quantitatively compare three nuclectomy techniques performed using an automated shaver, rongeurs, and laser.

Material & Methods

Comparisons were made in terms of mass, volume and location of material removal, changes in disc height, and stiffness. Fifteen vertebra-disc-vertebra lumbar specimens were acquired from six donors (40 ± 13 years) and split into three groups. Before and after nucleotomy axial mechanical tests were performed and T2-weighted 9.4T MRIs were acquired for each specimen.

Results

When using the automated shaver and rongeurs similar volumes of disc material were removed (2.51 ± 1.10% and 2.76 ± 1.39% of the total disc volume, respectively), while considerably less material was removed using the laser (0.12 ± 0.07%). Nuclectomy using the automated shaver and rongeurs significantly reduced the toe-region stiffness (p = 0.036), while the reduction in the linear region stiffness was significant only for the rongeurs group (p = 0.011). Post-nuclectomy, 60% of the rongeurs group specimens showed changes in the endplate profile while 40% from the laser group showed subchondral marrow changes.

Discussion

From the MRIs, homogeneous cavities were seen in the center of the disc when using the automated shaver. When using rongeurs, material was removed non-homogeneously both from the nucleus and annulus regions. Laser ablation formed small and localized cavities suggesting that the technique is not suitable to remove large volumes of material unless it is developed and optimized for this application.

Conclusion

The results demonstrate that both rongeurs and automated shavers can be used to remove large volumes of NP material but the reduced risk of collateral damage to surrounding tissues suggests that the automated shaver may be more suitable.

Application of single and cooperative different delivery systems for the treatment of intervertebral disc degeneration

Intervertebral disc (IVD) degeneration (IDD) is the most universal pathogenesis of low back pain (LBP), a prevalent and costly medical problem across the world. Persistent low back pain can seriously affect a patient's quality of life and even lead to disability. Furthermore, the corresponding medical expenses create a serious economic burden to both individuals and society. Intervertebral disc degeneration is commonly thought to be related to age, injury, obesity, genetic susceptibility, and other risk factors. Nonetheless, its specific pathological process has not been completely elucidated; the current mainstream view considers that this condition arises from the interaction of multiple mechanisms. With the development of medical concepts and technology, clinicians and scientists tend to intervene in the early or middle stages of intervertebral disc degeneration to avoid further aggravation. However, with the aid of modern delivery systems, it is now possible to intervene in the process of intervertebral disc at the cellular and molecular levels. This review aims to provide an overview of the main mechanisms associated with intervertebral disc degeneration and the delivery systems that can help us to improve the efficacy of intervertebral disc degeneration treatment.

Imaging Evaluation of Intervertebral Disc Degeneration and Painful Discs-Advances and Challenges in Quantitative MRI

In recent years, various quantitative and functional magnetic resonance imaging (MRI) sequences have been developed and used in clinical practice for the diagnosis of patients with low back pain (LBP). Until now, T2-weighted imaging (T2WI), a visual qualitative evaluation method, has been used to diagnose intervertebral disc (IVD) degeneration. However, this method has limitations in terms of reproducibility and inter-observer agreement. Moreover, T2WI observations do not directly relate with LBP. Therefore, new sequences such as T2 mapping, T1ρ mapping, and MR spectroscopy have been developed as alternative quantitative evaluation methods. These new quantitative MRIs can evaluate the anatomical and physiological changes of IVD degeneration in more detail than conventional T2WI. However, the values obtained from these quantitative MRIs still do not directly correlate with LBP, and there is a need for more widespread use of techniques that are more specific to clinical symptoms such as pain. In this paper, we review the state-of-the-art methodologies and future challenges of quantitative MRI as an imaging diagnostic tool for IVD degeneration and painful discs.

Injectable hydrogels for articular cartilage and nucleus pulposus repair: Status quo and prospects

Osteoarthritis (OA) and chronic low back pain due to degenerative (intervertebral) disc disease (DDD) are two of the major causes of disabilities worldwide, affecting hundreds of millions of people and leading to a high socioeconomic burden. Although OA occurs in synovial joints and DDD occurs in cartilaginous joints, the similarities are striking, with both joints showing commonalities in the nature of the tissues and in the degenerative processes during disease. Consequently, repair strategies for articular cartilage (AC) and nucleus pulposus (NP), the core of the intervertebral disc, in the context of OA and DDD share common aspects. One of such tissue engineering approaches is the use of injectable hydrogels for AC and NP repair. In this review, the state-of-the-art and recent developments in injectable hydrogels for repairing, restoring, and regenerating AC tissue suffering from OA and NP tissue in DDD are summarized focusing on cell-free approaches. The various biomaterial strategies exploited for repair of both tissues are compared, and the synergies that could be gained by translating experiences from one tissue to the other are identified.

A comprehensive tool box for large animal studies of intervertebral disc degeneration

Preclinical studies involving large animal models aim to recapitulate the clinical situation as much as possible and bridge the gap from benchtop to bedside. To date, studies investigating intervertebral disc (IVD) degeneration and regeneration in large animal models have utilized a wide spectrum of methodologies for outcome evaluation. This paper aims to consolidate available knowledge, expertise, and experience in large animal preclinical models of IVD degeneration to create a comprehensive tool box of anatomical and functional outcomes. Herein, we present a Large Animal IVD Scoring Algorithm based on three scales: macroscopic (gross morphology, imaging, and biomechanics), microscopic (histological, biochemical, and biomolecular analyses), and clinical (neurologic state, mobility, and pain). The proposed algorithm encompasses a stepwise evaluation on all three scales, including spinal pain assessment, and relevant structural and functional components of IVD health and disease. This comprehensive tool box was designed for four commonly used preclinical large animal models (dog, pig, goat, and sheep) in order to facilitate standardization and applicability. Furthermore, it is intended to facilitate comparison across studies while discerning relevant differences between species within the context of outcomes with the goal to enhance veterinary clinical relevance as well. Current major challenges in pre-clinical large animal models for IVD regeneration are highlighted and insights into future directions that may improve the understanding of the underlying pathologies are discussed. As such, the IVD research community can deepen its exploration of the molecular, cellular, structural, and biomechanical changes that occur with IVD degeneration and regeneration, paving the path for clinically relevant therapeutic strategies.

Application study of 3D LAVA-Flex on lumbar intervertebral disc degeneration

Background

Degeneration of the intervertebral discs are very common diseases, indicating the specific or malignant changes in intervertebral disc component, structure and function. Imaging examination is currently used to evaluate the severity of lumbar intervertebral disc degeneration. This study was designed to investigate the diagnostic value of 3D LAVA-Flex in lumbar intervertebral disc degeneration.

Material and methods

Sagittal 3D LAVA-Flex and T2WI scans were performed in 45 patients with lumbar intervertebral disc degeneration. On T2WI, the degenerated intervertebral disc in every patient was evaluated using Pfirrmann grade. Then, the patients were re-evaluated using 3D LAVA-Flex with considerations of the distinction of nucleus pulposus and annulus fibrosus, hypointense signal of intervertebral disc and height of intervertebral disc. The evaluation results were compared between 3D LAVA-Flex and T2WI. Virtual endoscopy was also performed to evaluate the degenerated intervertebral disc.

Results

The intermediate–intense signal of nucleus pulposus and complete ring-shaped hyperintense signal of annulus fibrosus were found and the distinction of nucleus pulposus and annulus fibrosus was clear in the normal intervertebral disc on 3D LAVA-Flex. The incidence of linear hypointensity of narrowed intervertebral space (65/91) was higher than that of normal intervertebral space (4/134) (P = 0.000). A good consistency was shown between the LAVA-Flex grade and T2WI-based Pfirrmann grade. Virtual endoscopy based on 3D LAVA-Flex could help clearly show the anatomic relationship between the degenerated disc and intervertebral foramen.

Conclusions

3D LAVA-Flex and T2WI show similar efficacy in evaluating lumbar intervertebral disc degeneration. 3D LAVA-Flex-based virtual endoscopy possesses great potential in the study of intervertebral disc abnormalities.

Feasibility of Deep Learning Algorithms for Reporting in Routine Spine Magnetic Resonance Imaging

BACKGROUND

Artificial intelligence is gaining traction in automated medical imaging analysis. Development of more accurate magnetic resonance imaging (MRI) predictors of successful clinical outcomes is necessary to better define indications for surgery, improve clinical outcomes with targeted minimally invasive and endoscopic procedures, and realize cost savings by avoiding more invasive spine care.

OBJECTIVE

To demonstrate the ability for deep learning neural network models to identify features in MRI DICOM datasets that represent varying intensities or severities of common spinal pathologies and injuries and to demonstrate the feasibility of generating automated verbal MRI reports comparable to those produced by reading radiologists.

METHODS

A 3-dimensional (3D) anatomical model of the lumbar spine was fitted to each of the patient's MRIs by a team of technicians. MRI T1, T2, sagittal, axial, and transverse reconstruction image series were used to train segmentation models by the intersection of the 3D model through these image sequences. Class definitions were extracted from the radiologist report for the central canal: (0) no disc bulge/protrusion/canal stenosis, (1) disc bulge without canal stenosis, (2) disc bulge resulting in canal stenosis, and (3) disc herniation/protrusion/extrusion resulting in canal stenosis. Both the left and right neural foramina were assessed with either (0) neural foraminal stenosis absent, or (1) neural foramina stenosis present. Reporting criteria for the pathologies at each disc level and, when available, the grading of severity were extracted, and a natural language processing model was used to generate a verbal and written report. These data were then used to train a set of very deep convolutional neural network models, optimizing for minimal binary cross-entropy for each classification.

RESULTS

The initial prediction validation of the implemented deep learning algorithm was done on 20% of the dataset, which was not used for artificial intelligence training. Of the 17,800 total disc locations for which MRI images and radiology reports were available, 14,720 were used to train the model, and 3560 were used to validate against. The convergence of validation accuracy achieved with the deep learning algorithm for the foraminal stenosis detector was 81% (sensitivity = 72.4.4%, specificity = 83.1%) after 25 complete iterations through the entire training dataset (epoch). The accuracy was 86.2% (sensitivity = 91.1%, specificity = 82.5%) for the central stenosis detector and 85.2% (sensitivity = 81.8%, specificity = 87.4%) for the disc herniation detector.

CONCLUSIONS

Deep learning algorithms may be used for routine reporting in spine MRI. There was a minimal disparity among accuracy, sensitivity, and specificity, indicating that the data were not overfitted to the training set. We concluded that variability in the training data tends to reduce overfitting and overtraining as the deep neural network models learn to focus on the common pathologies. Future studies should demonstrate the accuracy of deep neural network models and the predictive value of favorable clinical outcomes with intervention and surgery.

LEVEL OF EVIDENCE

3.

CLINICAL RELEVANCE

Feasibility, clinical teaching, and evaluation study.

Lumbar Intervertebral Disc Degeneration as a Common Incidental Finding in Young Pregnant Women as Observed on Prenatal Magnetic Resonance Imaging

Background: Obstetric imaging, subserving fetal evaluation, may yield incidental maternal findings. Based on prenatal magnetic resonance (MR) imaging, this study aims to investigate incidental intervertebral disc degeneration and displacement in young, pregnant women. Methods: This retrospective study included the sagittal 1.5 Tesla, T2-weighted lumbar spine images of 943 pregnant Central Europeans (age range, 18-47 years), who initially had undergone MR imaging because of sonographically suspected fetal abnormalities. Qualitatively, 4715 lumbar intervertebral discs were evaluated for degeneration using a modified Pfirrmann MR classification (nondegenerated, low-grade, moderate, and high-grade degeneration), as well as for displacement. In addition to descriptive statistics, an ordinal regression analysis was performed to analyze the relationship between degeneration and the women's age, and body weight. Results: With regard to the highest degree of degeneration in each woman, 578 (61.3%) showed low-grade, 211 (22.4%) moderate, and 154 (16.3%) high-grade degeneration, and no woman had entirely nondegenerated discs. For the span from 18 to 47 years of age, moderate and high-grade degeneration increased from 6.7% to 36.7% and from 13.3% to 22.4%, respectively. Of 943 women, 57 (6%) had disc displacements, of which 97% were in conjunction with high-grade degeneration. There was a statistically significant relationship (p < 0.001) between degeneration and age, and between degeneration and body weight. Conclusions: In young pregnant women, lumbar intervertebral disc degeneration is a ubiquitous, incidental finding, increasing from the late second decade of life onward, which may be part of physiological aging, as opposed to a small percentage of incidental disc displacements.

Gene-Hydrogel Microenvironment Regulates Extracellular Matrix Metabolism Balance in Nucleus Pulposus

Gene therapy provides an ideal potential treatment for intervertebral disk degeneration by delivering synthetic microRNAs (miRNAs) to regulate the gene expression levels. However, it is very challenging to deliver miRNAs directly, which leads to inactivation, low transfection efficiency, and short half-life. Here, Agomir is loaded in hydrogel to construct a gene-hydrogel microenvironment for regulating the synthesis/catabolism balance of the tissue extracellular matrix (ECM) to treat degenerative diseases. Agomir is a cholesterol-, methylation-, and phosphorothioate-modified miRNA, which can mimic the function of miRNA to regulate the expression of the target gene. Agomir874 that mimics miRNA874 is synthesized to down regulate the expression of matrix metalloproteinases (MMPs) in nucleus pulposus (NP). At the same time, a polyethylene glycol (PEG) hydrogel is synthesized through Ag-S coordination of 4-arm PEG-SH and silver ion solution, which has injectable, self-healing, antimicrobial, degradable, and superabsorbent properties and matches perfectly with the mechanism of intervertebral disk. By delivering Agomir-loaded PEG-hydrogel to a degenerative intervertebral disk, a gene-hydrogel microenvironment is constructed in situ, which reduces the expression of MMPs, regulates the synthesis/catabolism balance of ECM in the NP of the intervertebral disk, and improves the tissue microenvironment regeneration.