Automatic Modic Changes Classification in Spinal MRI

BACKGROUND Modic changes (MC) are abnormal bone signals under the vertebral endplates on magnetic resonance imaging (MRI) of the spine. Three types of MC may be seen on MRI as abnormal bone signals of the vertebral endplate and adjacent bone marrow. This retrospective study of 203 patients with symptomatic cervical ossification of the posterior longitudinal ligament (OPLL) aimed to evaluate factors associated with MC on cervical spine MRI. MATERIAL AND METHODS We included 203 patients with symptomatic cervical ossification of the OPLL. All patients underwent MRI with T1 and T2 sequences to assess the presence and type of MC. Univariate and multivariate logistic regression analyses were used to identify the risk factors for MC. RESULTS The prevalence of MC in patients with symptomatic cervical OPLL was 21.18%. Type 2 MC accounted for 88.64% of the cases. Local type and MC share the same segment (100.00%), followed by segmental type (77.27%), mixed type (75.00%), and continuous type (75.00%). Age (OR=1.05, 95% CI: 1.01-1.09, P=0.013) and neck pain (OR=2.67, 95% CI: 1.04-6.83, P=0.041) were significantly correlated with MC. Further receiver operating characteristic (ROC) analysis displayed an area under the curve (AUC) of 0.657 (P=0.002) for age, and the optimal cutoff was 60.50 years (sensitivity 0.67, specificity 0.68). CONCLUSIONS The results of this study found that older age and neck pain were significantly associated with development of cervical spine MC. Patients ≥60.5 years with symptomatic cervical OPLL had a higher probability of developing MC.

Retrospective cohort study. To examine the relationship between preoperative Modic change (MC) and postoperative clinical assessment scores for patients receiving lumbar discectomy or transforaminal lumbar interbody fusion for lumbar disk herniation. Lumbar disk herniation is a risk factor for MC development. MC on spinal magnetic resonance imaging (MRI) has been associated with worse preoperative and postoperative clinical assessment scores. We reviewed data for 285 primary single-level surgeries. Preoperative and 12-month postoperative assessment scores were recorded using the visual analog scale leg pain, Oswestry Disability Index, and Short Form-36 Physical Component Summary. MC subgroup on preoperative MRI was recorded by a single neuroradiologist. One hundred seventy-nine patients (female, 56%; age-53±13 y) with preoperative MRI were included. Age and sex were similar across MC subgroups. The sample prevalence of MC on preoperative MRI was 62%, and MC2 was the most common subgroup (35%). No differences in preoperative assessment scores were identified, regardless of presence or absence of MC. For the overall cohort, improvement in assessment scores were observed: Short Form-36 improved an average of 8.2 points [95% CI (95% CI), 5.8-10.7], Oswestry Disability Index by 11.3 points (95% CI, 8.7-14.0), and visual analog scale by 2.8 points (95% CI, 2.1-3.5). In nearly all cases, MCID values were met, even when stratifying by MC subgroup. Few differences in postoperative assessment scores were identified when comparing across MC1, MC2, or no MC groups. Statistically and clinically significant improvement in postoperative clinical assessment scores was observed for both lumbar discectomy and transforaminal lumbar interbody fusion groups. MC on preoperative MRI was not associated with worse preoperative or postoperative clinical assessment scores. Level III.

Longitudinal cohort study of twins representative of the general population. To assess the relationship between Modic change (MC) and severe, disabling low back pain (LBP), features of intervertebral disc degeneration (DD) and incident MC during 10-year follow-up. MC describes vertebral endplate and bone marrow lesions visible on magnetic resonance imaging (MRI). MC has been associated with DD. It remains unclear whether MC causes LBP independently or through association with DD. Moreover, association of MC with severe, disabling LBP is uncertain. Volunteers were recruited from the TwinsUK register to MRI and interview between 1996 and 2000 with a subset attending for follow-up a decade later. MC, DD (evaluated by loss of disc height and signal intensity, presence of disc bulge and anterior osteophytes) and Schmorl's nodes (SN) were determined on T2-weighted lumbar MR scans. Complete data were available for 823 subjects at baseline and 429 at follow-up. Mean age at baseline was 54.0 years (range 32-70) with 96% females. The prevalence of MC was 32.2% at baseline and 48.7% at follow-up. Subjects with MC were older (P < 0.001) and more overweight (BMI: P = 0.026, weight: P < 0.001). At both baseline and follow-up, more subjects reporting severe LBP demonstrated MC (subjects with MC vs. without MC: 35.0% vs. 16.4% respectively, P < 0.001 at baseline; and 35.1% vs. 20.0% respectively, P < 0.001 at follow-up). In multivariable analyses, MC remained significantly associated with episodes of severe, disabling LBP (OR 1.58; 95% CI 1.04-2.41) after adjustment for age, BMI, DD, and SN at baseline. Loss of disc height and disc signal intensity were independently associated with prevalent MC at baseline, and disc height and disc bulge with incident MC during follow-up. MC is an independent risk factor for episodes of severe and disabling LBP in middle-aged women. These observations support further work aimed at identifying the precise histology underlying MC. 2.

A cross-sectional observational study. To investigate whether there is a difference in findings of lumbar Modic changes in low-field (0.3 T) magnetic resonance imaging (MRI) compared with high-field (1.5 T). It is a challenge to give patients with low back pain a specific diagnosis. Modic changes as seen on MRI have been reported to be a possible source of pain. However, it is unclear whether the diagnosis is independent on the field strength. Twenty patients with Modic changes, 11 women and 9 men (mean age, 53.6 yr; range, 29-81 yr), with or without sciatica, seen in a Danish outpatient low back pain clinic were included. All patients obtained MRI scans on both a high-field and a low-field MRI scanner. Two radiologists evaluated all lumbar endplates independently, using a standardized evaluation protocol. Kappa statistics were used to analyze the interobserver reproducibility. We used paired t test to analyze the difference between low- and high-field MRI. The total number of Modic changes diagnosed with high-field MRI was significantly higher than that with low-field MRI. However, 3 to 4 times as many Modic type 1 changes were found with low-field MRI compared with high-field MRI. Contrarily, with high-field MRI type 2 changes were diagnosed twice as often. There was a significant difference between low- and high-field MRI regarding the overall prevalence of any Modic change, but this had opposite directions for types 1 and 2: type 2 dominated in high field and conversely in high field [corrected]. The type of MRI unit should be taken into consideration when diagnosing patients with Modic changes

Background: Modic changes (MC) and Schmorl’s nodes (SN) are commonly observed spinal abnormalities in lumbar magnetic resonance imaging (MRI) studies. MC is categorized into three types: type I (inflammatory), type II (fatty degeneration), and type III (sclerotic). SN refers to the herniation of the nucleus pulposus into adjacent vertebrae, often resembling an “ice-cream scoop” on imaging. Materials and Methods: This study analyzed MRI scans of the lumbar spine from 500 patients to evaluate the presence of SN and MC alongside other spinal conditions. Participants were divided into four groups: •Group A: Control group with no SN, MC, or other spinal pathologies. •Group B: Patients without SN or MC presenting conditions like spinal straightening or altered lordosis. •Group C: Individuals with MC, classified by type (type I, II, and III) and location (superior, inferior, right, left, and multilevel). •Group D: Patients with SN, categorized by shape (round, sharp, irregular, and small) and location, including cases with multiple nodes across levels. Results: Type II MC was the most prevalent subtype observed. Gender differences were noted: males were more commonly affected in the right superior region, whereas females showed a higher incidence in the right inferior region. MC was significantly associated with aging (P &lt; 0.05). SN was predominantly spherical and centrally located. Females demonstrated a higher prevalence of SN (8.5%) than males (6.5%), with a significant age-related correlation (P &lt; 0.05). Notably, 24% of females exhibited both SN and MC in the same region. Conclusion: The study highlights gender-specific patterns in MC and SN. It reveals that type 2 MC is most commonly observed at the L4 and L5 vertebrae, particularly in the “LEFT” and “SUPERIOR” regions. SN is more frequently found in females, with a higher incidence of multilevel occurrences. Notably, the co-occurrence of MC and SN at the same vertebral sites was observed in 24% of females. Type 2 MC was especially prominent in these cases, suggesting a potential relationship between the two conditions.

Modic changes (MC) are signs of vertebral pathology visible on magnetic resonance (MR) images that have been associated with low back pain (LBP) and disc degeneration in people. Multiple breeds of dogs also develop MCs and coincident back pain. However, the association between breed, MC, and spinal pathologies has yet to be fully elucidated. This study aimed to identify the prevalence of MC that occur spontaneously in the lumbar vertebral column of dogs diagnosed with intervertebral disc disease (IVDD) and examine their association with demographic criteria and the disc width index (DWI). Medical records and lumbar vertebral column MR images were examined from 104 dogs (831 intervertebral disc spaces and adjacent vertebrae), which were divided into three groups: chondrodystrophic dogs (CD; n =54) and non-chondrodystrophic dogs (NCD; n =30) with IVDD as the primary diagnosis, and control dogs (n =20) with other spinal diseases as their primary diagnosis. Increasing age and a diagnosis of IVDD were significantly associated with MC in dogs (p < 0.001 and p = 0.0062, respectively). In CD dogs with IVDD, Type 2 MC were most prevalent, whereas, in NCD dogs, Type 3 MC were the most prevalent type. Type 2 MC were distributed nearly equally across the lumbar vertebral column, while Type 3 MC were primarily detected at the level of L7-S1. This study demonstrated that MC developed spontaneously in dogs, are common in dogs diagnosed with IVDD, and the type observed varies by breed. Further research is needed to understand the pathogenesis of MC; however, the increased presence of Type 2 MC in CD dogs, similar to what is found in people with disc degeneration, suggests that CD dogs could serve as models for MC in people.

ObjectiveThe aim of this study was to investigate the clinicopathologic features of lumbar disc herniation (LDH) with endplate degeneration and the association between cartilaginous fragments and inflammatory response to the herniated disc.Summary of background dataLDH often involves hyaline cartilage fragments pulled from the vertebral endplates. Modic changes are closely associated with LDH that contains hyaline cartilage, and cartilaginous endplates seem to affect resorption of the herniated disc.MethodsA total of 78 patients who underwent microscopic discectomy between 9 and 16 weeks after an occurrence of LDH were reviewed. Modic changes, disc degeneration, high-intensity zone, and vertebral corner defect were evaluated using magnetic resonance imaging (MRI). Histopathological observations of cartilaginous endplates and inflamed granulation tissue in the herniated disc were made. In cases with inflamed granulation tissue, neovascularization and macrophage infiltration were also evaluated using immunohistochemical analysis.ResultsModic changes were observed in approximately one-third of the patients (26 cases: type 1, 7; type 2, 17; and type 3, 2). Cartilaginous endplates were observed in 32 cases (41%) and in the majority of cases with Modic changes compared with cases without Modic changes (65%, p = 0.001). Although inflamed granulation tissue was observed in 60 cases (77%), no significant differences were detected in patient age and the composition of the herniated material. Immunohistochemical analysis showed that fewer CD34-positive capillaries and CD68-positive cells were found in cases with cartilaginous fragments compared with those without cartilaginous fragments (p < 0.001). In addition, a higher immunoreactivity to CD34 and CD68 was found in herniated discs <25% of whose area was occupied by cartilaginous endplates compared with discs whose area was occupied at 25% or more (p < 0.001).ConclusionThere is an association between LDH with endplate degeneration and cartilaginous herniation, with Modic type 2 predominating. Furthermore, neovascularization and macrophage infiltration, especially if the amount of cartilage is high, are likely to be less frequent in cartilaginous herniation, leading to failure in the spontaneous remission of clinical symptoms.

Modic changes (MC) on magnetic resonance imaging (MRI) have been associated with the development and severity of low back pain (LBP). The etiology of MC remains elusive, but it has been suggested that altered metabolism may be a risk factor. As such, this study aimed to identify metabolomic biomarkers for MC phenotypes of the lumbar spine via a combined metabolomic-genomic approach. A population cohort of 3,584 southern Chinese underwent lumbar spine MRI. Blood samples were genotyped with single-nucleotide polymorphisms (SNP) arrays (n = 2,482) and serum metabolomics profiling using magnetic resonance spectroscopy (n = 757), covering 130 metabolites representing three molecular windows, were assessed. Genome-wide association studies (GWAS) were performed on each metabolite, to construct polygenic scores for predicting metabolite levels in subjects who had GWAS but not metabolomic data. Associations between predicted metabolite levels and MC phenotypes were assessed using linear/logistic regression and least absolute shrinkage and selection operator (LASSO). Two-sample Mendelian randomization analysis tested for causal relationships between metabolic biomarkers and MC. 20.4% had MC (10.6% type 1, 67.2% type 2, 22.2% mixed types). Significant MC metabolomic biomarkers were mean diameter of very-low-density lipoprotein (VLDL)/low-density lipoprotein (LDL) particles and cholesterol esters/phospholipids in large LDL. Mendelian randomization indicated that decreased VLDL mean diameter may lead to MC. This large-scale study is the first to address metabolomics in subject with/without lumbar MC. Causality studies implicate VLDL related to MC, noting a metabolic etiology. Our study substantiates the field of "spino-metabolomics" and illustrates the power of integrating metabolomics-genomics-imaging phenotypes to discover biomarkers for spinal disorders, paving the way for more personalized spine care for patients.

The impact of endplate changes, such as Modic changes (MCs), on symptoms following discectomy remains uncertain. This study aimed to investigate the postoperative progression of MCs and bony endplate defects after discectomy for lumbar disc herniation (LDH) and to elucidate its association with clinical outcomes. A total of 247 patients who underwent microscopic discectomy were included. The natural progression of MCs and bony endplate defects at the operated level were assessed using magnetic resonance imaging. Clinical outcomes were evaluated using a visual analog scale (VAS) and the Roland-Morris Disability Questionnaire. The study examined the progression of endplate changes and their impact on clinical outcomes, with a focus on the type of postoperative MC. The natural progression of postoperative MCs was varied, with MCs coexisting with bony endplate defects (p < 0.001). Persistent or converted type 1-related MCs were more likely to exacerbate pre-existing signal changes at one year and were associated with higher VAS scores for low back pain (LBP) compared to type 2 MCs (p < 0.001). Patients with type 1-related MCs (β: 0.429, p < 0.001) and preoperative LBP (β: 0.215, p = 0.025) were closely associated with residual LBP at one year or more. In patients with baseline signal changes in subchondral bone marrow, the presence of type 1 MCs was more likely to promote the increase of MCs than type 2. Furthermore, the presence of type 1 MCs may predict residual postoperative LBP, suggesting that bone marrow signal changes are more strongly associated with clinical symptoms than morphological changes in bony endplates in patients with LDH.

BackgroundBony fusion rate was significantly lower in patients with type 3 Modic change than patients with normal endplates. It is not known whether there are relevant differences in fusion efficiency among patients with type 2 sclerotic Modic change or non-sclerotic Modic change, or no Modic change.MethodsA retrospective study contained 196 lumbar segments in 123 subjects undergoing posterior lumbar interbody fusion (PLIF) with pedicle screw instrumentation (PSI) to assess the effect of type 2 sclerotic Modic change on fusion efficiency. These endplates were allocated into groups A, B, and C, according to their Modic changes. Group A had endplates with type 2 Modic change and endplate sclerosis. Group B had type 2 Modic change without endplate sclerosis. Group C had neither Modic change nor endplate sclerosis. The presence of Modic change was determined by magnetic resonance imaging (MRI). Endplate sclerosis in type 2 Modic change was detected by computed tomography (CT) before the operation. We collected CT data 3 months to more than 24 months after operation in patients to assess bony fusion.ResultsIncidences of bony fusion were 58.8% in group A, 95.0% in group B, 94.3% in group C. The bony fusion rate was significantly lower in group A than in either group B or C. There was no significant difference between groups B and C. Thus, endplates with type 2 sclerotic Modic change had a lower fusion rate in patients undergoing PLIF with PSI.ConclusionType 2 sclerotic Modic change could be an important factor that affects solid bony fusion in patients undergoing PLIF with PSI. CT may help diagnose endplate sclerosis in patients with type 2 change and inform the choice of the best site for spinal fusion.

Introduction Low back pain is the world's most disabling condition. Modic changes (MC) are associated with low back pain. These changes are spinal phenotypes that represent vertebral endplate and adjacent marrow changes on MRI. MC are classified into three main types (type I, type II, and type III) and mixed types (type I/II and II/III). Due to methodological biases in previous studies, the morphology, involvement of MC, and their association with other spinal phenotypes remain speculative. As such, the aim of this study was to evaluate the relationship of MC with other spinal MRI phenotypes in a large-scale population-based study. Materials and Methods Based on the Hong Kong Disc Degeneration Cohort of Southern Chinese, we assessed the T1- and T2-weighted MRIs of 1,604 subjects (62.4% females; mean age: 49 years) from L1 to S1. The MC assessment included the presence, type, vertical height, and axial area of MC. MC were evaluated as type I, type I/II, type II, type II/III, and type III. Types were regrouped in the analyses as “type I” (types I and I/II) and “type II” (types II and II/III). Very small MC, such as MC in only one sagittal plane, were excluded. Additional imaging phenotype findings were assessed (disc bulges/extrusions, Schmorl nodes, disc degeneration). Disc degeneration was based on the Pfirrmann classification. A degenerative disc disease (DDD) score was tabulated, which represented the global severity of disc degeneration of the lumbar spine. The lumbar spine was further stratified to upper (L1-L4) and lower (L4-S1) regions. Results The prevalence of MC was 24.7% (“type I”: 6.3%, “type II”: 15.5%). Of all MC, 77% were at L4-S1. Subjects with MC were older (mean age: 53 vs. 48 years, p &lt; 0.001) and had higher DDD scores ( p &lt; 0.001). “Type I” MC were more common at lower lumbar levels ( p = 0.021), were less likely to be located only in the anterior region ( p = 0.017), and were more associated with disc bulges/extrusions ( p &lt; 0.001) in comparison to “type II” MC. MC of the lower lumbar levels were not commonly noted only in the anterior region, involved more likely only the left or right endplate and had a higher prevalence of disc bulges/extrusions and disc degeneration in comparison to upper lumbar levels ( p &lt; 0.001). Large MC (≥2/3 of the axial area) were more likely located at lower lumbar levels (83 vs. 73%, p = 0.001) and had a higher prevalence of disc bulge/extrusion (83 vs. 72%, p = 0.001) and Schmorl node at the affected level (52 vs. 39%, p &lt; 0.001) compared with smaller MC. Conclusion Based on one of the largest population-based studies, our findings strengthen the belief that MC are clearly associated with disc pathology, such as disc degeneration, disc bulges/extrusions, and endplate abnormalities (e.g., Schmorl nodes). MC type- and level-related findings in relation to additional MRI phenotypes were also identified. This study further refines the phenotypic classification of MC, which if standardized can have immense utility in studies assessing the role of biomarkers (e.g., genetics) in relation to clinically relevant spinal changes. Acknowledgments The study has been supported by AOSpine Research Network Exchange Award. Disclosure of Interest None declared

Cervical spondylotic myelopathy (CSM) represents the most commonly acquired cause of spinal cord dysfunction among individuals over 55 years old. The pathophysiology of the condition involves mechanical factors, which result to injury of the cervical spinal cord. In T-2 weighted magnetic resonance (MR) images of the spine the site of injury is depicted as a region of high intensity signal within the cervical spine cord. The present study aims to investigate whether texture analysis of MR signal in CSM could provide novel quantitative prognostic factors, rendering possible the prognostic estimation of the outcome of a therapeutic surgical intervention for CSM. The sample of the study comprised 12 MR images of the cervical spine, corresponding to 6 CSM patients, who had undergone surgical intervention with anterior cervical discectomy and spinal canal decompression. Following a specific MR imaging protocol a pair of T2-weighted sagittal images of the spine, corresponding to pre- and post-operative MR scans, were obtained for each of the patients. Employing custom developed software, the region of high intensity signal, associated to CSM, was automatically segmented from each MR image. Utilizing custom developed algorithms a number of textural features were extracted from the segmented ROIs and employed in the design of a classification system, based on the Quadratic classifier. The latter was used for the discrimination between pre-operative and post-operative MR images. Statistical analysis revealed the existence of statistically significant differences between textural features, corresponding to pre- and post-operative CSM MR signals. The quadratic classifier characterized correctly all the pre- and post-operative MR images (100% classification accuracy). The results of the present study indicate that textural features, generated from MR images of the spine, may serve as prognostic factors regarding the prediction of the post-operative outcome of CSM patients.

ObjectivesTo evaluate the changes of Hounsfield units (HU) value in different types of Modic changes (MCs) and to analyze the correlation between the change of HU value and area ratio of MCs region, bone mineral density (BMD), and degree of intervertebral disc degeneration.MethodsOne hundred fifty-eight endplates with MCs were included and analyzed. HU values of MCs regions and adjacent vertebral corresponding regions without MCs were measured. The area ratio of MCs region was defined as the area of MCs divided by the area of endplate or the vertebral sagittal plane. BMD was measured by Dual-energy x-ray absorptiometry (DXA). Degree of intervertebral disc degeneration was evaluated based on Pfirrmann classification. According to the types of variables, descriptive statistics, Kolmogorove-Smirnov test, paired t-test, Wilcoxon signed-rank test, Independent-Samples T Test, and Pearson correlation analysis were used.ResultsThe HU values in any types of MCs are significantly higher than that of adjacent vertebral corresponding regions without MCs (P < 0.001). The HU value of the type III MCs is higher than that of the type I and type II MCs. HU value was positively correlated with BMD. In the levels with Grade V disc degeneration, the area ratio of MCs region was significant increased.ConclusionsHU values of the vertebral endplate and bone marrow were increased in most MCs regions with all types of MCs. HU value of endplates had a significantly positive correlation with BMD. Higher area ratio of MCs region is associated with more severe intervertebral disc degeneration.

The purpose was to investigate if high intensity zones (HIZ), Modic (MC), and endplate changes (EPC) display different behaviors measured with quantitative magnetic resonance imaging (MRI) with and without loading of the spine and if there is a simultaneous presence of these features in the same motion segment. 130 motion segments in patients with chronic low back pain (n = 26, 25-69 year, mean 38 year, 11 males) were examined. HIZs, MCs, and EPCs (i.e.,structural findings, reflecting calcifications, erosions, and fissures) were determined with standardized MRI. Different T2-values with and without loading for these features were then determined with the quantitative MRI method T2-mapping. Significantly different behaviors were found in the spinal tissues with associated HIZs, MC, and EPC (p < 0.004). HIZ (62% of patients, 1-2/patient) was associated with EPC (100% of patients, 1-7/patient) (p = 0.0003 and 0.0004 for upper and lower EPs), with an occurrence of 91% for upper and 71% for lower endplates adjacent to discs with HIZ. MC (81% of patients, 1-3/patient) were associated with EPC (p < 0.0001) with an occurrence of 87% for endplates adjacent to vertebrae with MC. The occurrence of both HIZ and MC was 43% (p = 0.0001) for upper and 29% (p = 0.003) for lower vertebrae. HIZ was associated with simultaneous presence of both MC and EPC in the same motion segment. T2-mapping was found to objectively reflect changes in the spinal tissues associated with HIZs, MC, and EPC.

Modic changes of the lumbar spine: prevalence, risk factors, and association with disc degeneration and low back pain in a large-scale population-based cohort