Magnetic resonance imaging-based bone and muscle quality parameters for predicting clinical subsequent vertebral fractures after percutaneous vertebral augmentation

Introduction

Subsequent vertebral fracture (SVF) is a common complication following percutaneous vertebral augmentation (PVA) surgery in patients with an osteoporotic vertebral compression fracture (OVCF) (1). Clinical SVF can lead to recurrent pain and exacerbated kyphotic deformity and thus often necessitates a second surgery, thereby increasing patient pain and financial burden (2,3). The risk factors associated with SVF have been widely discussed in previous research, with osteoporosis and sarcopenia being identified as critical contributors (4,5). Therefore, assessing bone quality and the degree of muscle atrophy may assist in predicting the likelihood of SVF.

According to the World Health Organization (WHO), dual-energy X-ray absorptiometry (DEXA) is the gold standard for diagnosing osteoporosis WHO (6). However, it has many limitations, such as potentially yielding erroneously elevated bone mineral density (BMD) measurements in patients with concurrent osteophytes, aortic calcification, or obesity (7). Additionally, not all patients with OVCF routinely undergo DEXA scanning. Unless contraindicated, magnetic resonance imaging (MRI) is essential for diagnosing OVCF and SVF. Recently, a novel method known as the MRI-based vertebral bone quality (VBQ) score has been introduced as an alternative for the evaluation of osteoporosis (8). Meanwhile, measurement of the cross-sectional area (CSA) and the degree of fat infiltration (DFI) of the paravertebral muscles on MRI provides a good assessment of the extent of sarcopenia (9). Related studies have shown that these MRI-based parameters demonstrate potential as a predictor for identifying fragile fractures and various post-spinal surgery complications, such as pedicle screw loosening, cage subsidence, and proximal junctional kyphosis (10-13). However, the correlation between these parameters and SVF remains undefined.

Therefore, the purpose of this study was to evaluate the utility of these MRI-based bone and muscle quality parameters in predicting clinical SVF after PVA. Additionally, we aimed to investigate the correlation between the VBQ score and both the CSA and DFI of the paravertebral muscles. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-24-712/rc).

Methods

Study design, ethics, and population

This cross-sectional study included 341 consecutive patients with OVCF treated with PVA at Xuzhou Central Hospital between January 1, 2017 and December 31, 2020. The study focused on OVCF cases in the thoracic and lumbar regions, as these are the most frequently affected areas in osteoporotic fractures. This study was conducted in accordance with the ethical principles of the Declaration of Helsinki (as revised in 2013) and received approval from the Ethics Committee of Xuzhou Central Hospital (No. XZXY-LK-20210310-0134). The requirement for individual patient consent was waived by the ethics committee, as the research was retrospective in nature, used pre-existing data, and did not involve any additional interventions or risks to the patients.

This study included male and female patients aged 65 years or older who were diagnosed with single-segment OVCF based on symptoms, clinical signs, and MRI findings and who had undergone PVA treatment. Patients were excluded if they had two or more segmental OVCFs; a history of interim spinal trauma, spinal surgery, severe spinal deformities (defined as scoliosis with Cobb angle >40° or thoracolumbar kyphosis >70°), spinal tumors or infections, or neuromuscular diseases, as these conditions could have affected the measurements of bone and muscle quality parameters; or insufficient medical records or unusable imaging data. After applying the inclusion and exclusion criteria, 289 patients were included in the final analysis, and they were divided into two groups depending on whether SVF occurred within a 3-year follow-up period: the SVF group and the non-SVF group. Demographic parameters used in the statistical analysis were collected from the electronic patient records. Detailed definitions of patient characteristics are provided in the Appendix 1.

Sample size calculation

To ensure the reliability of our results, we conducted a post hoc power analysis using G*Power software version 3.1.9.7. Based on an estimated effect size of 0.5, a significance level (α) of 0.05, and a desired power (1−β) of 0.80, we calculated the required sample size for our study. Considering the allocation ratio between the non-SVF and SVF groups, which was approximately 6.05 (248/41), the analysis indicated that a minimum total sample size of 206 patients was necessary to achieve adequate statistical power. Our actual sample size comprised 289 patients, exceeding the requirement and indicating that our study possessed sufficient statistical power to support the validity of our findings.

Definition of SVF

In previous studies, SVF has been categorized into clinical and radiological types (14-16). Radiological SVF is identified by a decrease in body height of more than 20% on lateral X-ray. Clinical SVF includes these radiological criteria along with the presence of bone marrow edema in new segments on MRI and recurrent back pain. In this study, we focused on clinical SVF because it directly correlates with patient symptoms and quality of life. SVF was defined as a new vertebral fracture occurring within 3 years after the initial PVA procedure. Patients who presented with new-onset back pain underwent MRI as soon as possible after symptom onset to confirm the diagnosis. The presence of bone marrow edema on MRI was used as a definitive sign of an acute fracture.

Imaging procedures

The MRI scanning was conducted using a 3-Tesla Discovery MR750w system (GE HealthCare, Chicago, IL, USA). The imaging protocol included sagittal T1- and axial T2-weighted sequences. The sagittal T1-weighted imaging included a T1 fast spin-echo pulse sequence with a repetition time (TR) of 770 ms, an echo time (TE) of 7 ms, a matrix of 320×256, 2 excitations, a slice thickness of 4.0 mm, and a slice space of 0.4 mm; meanwhile, axial T2-weighted imaging included a T2 fast spin-echo pulse sequence with a TR of 3,600 ms, a TE of 100 ms, a matrix of 384×288, 2 excitations, a slice thickness of 4.0 mm, and a slice space of 0.4 mm. Axial slices were acquired angled along each individual disc plane to ensure precise anatomical alignment and accurate measurements.

Measurement of VBQ score

The VBQ score was calculated on lumbar noncontrast T1-weighted MRI according to the method introduced by Ehresman et al. (8). In a midsagittal slice of the spine, regions of interest (ROIs) were placed in the medullary portions of the L1–4 vertebrae and the cerebrospinal fluid (CSF) space at the L3 level (Figure 1). The mean signal intensity (SI) for each ROI was captured and used to determine the VBQ score, which was calculated by dividing the median SI from the L1–4 vertebrae by the SI of the CSF as follows:

VBQscore=SIL1−L4SICSF[1]

Figure 1 Non-contrast T1-weighted MRI showing the signal intensity of the L1–4 and CSF using regions of interest (yellow circles). *, the L5 fracture. MRI, magnetic resonance imaging; CSF, cerebrospinal fluid.

When factors such as scoliotic changes, venous plexus, or hemangioma impede precise ROI assessments on midsagittal slices, measurements could be conducted on parasagittal planes. The vertebral body was omitted if any abnormality spanned across all its sagittal sections. In certain instances, the ROI for the CSF space at L2 or L4 was used as an alternative to precisely represent the SI of the CSF due to the space at L3 being occupied by descending roots. The VBQ score was independently assessed by two spine surgeons, with the final score derived from the average of both evaluations.

Measurement of paravertebral muscle quality parameters

The CSA and DFI of the paravertebral muscles (with a focus on multifidus and erector spinae muscles) were measured using ImageJ software (US National Institutes of Health, Bethesda, MD, USA) on axial T2-weighted MRI (Figure 2). Middle images of the disc space at the L4–5 and L5–S1 levels were selected by locating lines on the sagittal slice. For each disc level, CSA was measured by meticulously outlining the muscle contours based on the method described by Crawford et al., which uses the facet joint as a landmark and excludes epimuscular fat compartments to focus solely on the muscle tissue (17). The DFI was determined using the circle method as described by Fortin et al. (18). This method involves selecting six ROIs from the visible areas of the muscle tissue with the least visual fatty infiltration in the multifidus and erector spinae muscles. The maximum intensity value obtained from these ROIs was used as the upper threshold to distinguish muscle tissue from fat, while the lower limit was set uniformly at 0 to minimize errors. After thresholds were applied, fatty tissue was highlighted in red, and the DFI was determined by calculating the proportion of red areas within the overall muscle space. The final CSA and DFI values were calculated as the average of measurements from both the L4–5 and L5–S1 levels, which were obtained bilaterally and independently assessed by two spine surgeons to ensure accuracy and reduce bias.

Figure 2 Assessment of paravertebral muscle quality using axial T2-weighted MRI with ImageJ software. (A) The total cross-sectional area. (B) The degree of fat infiltration. MRI, magnetic resonance imaging; MF, multifidus; ES, erector spinae; PM, psoas major.

Blinding methods

To minimize potential bias in this study, all imaging analyses were conducted in a blinded manner. The assessors responsible for measuring the VBQ score and paravertebral muscle quality parameters were blinded to the clinical information, including the visual analog scale (VAS) pain scores and the occurrence of SVF. Patient data were anonymized, ensuring that no identifying information was accessible during the analysis process. This blinding approach was maintained throughout the study to ensure the objectivity and reliability of the imaging assessments.

Statistical analysis

Based on the occurrence of SVF within a 3-year follow-up period, all the patients were divided into the SVF group or the non-SVF group. The normality of all parameters was tested via the Kolmogorov-Smirnov test. Two independent samples t tests were applied for parameters following a normal distribution. On the other hand, for parameters that do not conform to a normal distribution, the Mann-Whitney test was employed for intergroup comparisons, and these parameters are expressed as the interquartile range (IQR). Categorical variables are expressed as numbers and frequencies and were statistically analyzed with chi-square or Fisher exact test.

A univariable logistic regression analysis was used for all factors. Age, sex, body mass index (BMI), and other factors with a P value less than 0.1 in the univariable analysis were selected for inclusion in the multivariable logistic regression analysis. Finally, the receiver operating characteristic (ROC) curve was plotted, and the area under the curve (AUC) was calculated. The optimal cutoff values for predicting SVFs were determined with the Youden index. To compare the predictive performance between the different prediction models, the DeLong test was used. Pearson correlation analysis was used to assess the relationship between the VBQ score and both CSA and DFI. Statistical analysis was performed using R-Studio version 2023.12.0 (Posit Software, Boston, MA, USA). P<0.05 indicated a statistically significant difference.

Results

Participants and grouping

The participant demographics are provided in Table 1. A total of 289 patients (69 male; average age 72.1±9.8 years) were included in this study. Based on the occurrence of SVF within a 3-year follow-up period, the patients were divided into an SVF group (41 patients, 14.19%) or a non-SVF group (248 patients, 85.81%).

Comparison of bone and muscle quality parameters between the two groups

Detailed participant clinical characteristics are given in Table 1. Significant differences were noted in the occurrence of intervertebral cement leakage between the SVF and non-SVF groups (78.05% vs. 46.37%; P<0.001). Of note, compared with the non-SVF group, the SVF group had a significantly higher VBQ score (P=0.0017) and DFI (P<0.001). However, there was no significant difference in CSA between the SVF group and the non-SVF group (P=0.829).

Prediction of SVF via bone and muscle quality parameters

The multivariable regression analysis revealed that SVF was associated with a higher VBQ score [odds ratio (OR) =3.66; P<0.001] and a higher DFI (OR =3.72; P<0.001). However, there was no significant relationship between the SVF and CSA (Table 2). The ROC curve demonstrated that the AUC value of the VBQ score alone was 0.863, with the Youden index indicating an optimal cutoff value at 3.49 (sensitivity 85.4%; specificity 73.4%). Similarly, the AUC value for DFI alone stood at 0.851, with an optimal cutoff value at 48.2% (sensitivity 90.2%; specificity 71.0%). Most notably, the combined predictive model incorporating both VBQ score and DFI significantly enhanced the predictive accuracy, demonstrated by an AUC value of 0.925 (Figure 3). According to the Delong test, the AUC value of the VBQ score and that of DFI were comparable (0.863 vs. 0.851; P=0.8437); the AUC value of combined model was higher than that of the VBQ score alone (0.925 vs. 0.863; P=0.0389) and that of the DFI alone (0.925 vs. 0.851; P=0.0254).

Figure 3 ROC curve analysis illustrating the predictive performance for subsequent vertebral fracture of the (A) VBQ score, (B) DFI, and (C) their combination. ROC, receiver operating characteristic; AUC, area under the curve; VBQ, vertebral bone quality; DFI, degree of fat infiltration.

The correlation between VBQ score and CSA, DFI

There was a linear correlation between VBQ score and DFI (Figure 4). The Pearson correlation coefficient showed that VBQ score was positively correlated with DFI (r=0.647; P<0.001) while no significant correlation was apparent between VBQ score and CSA (r=−0.039; P=0.7495).

Figure 4 Correlation between VBQ score and muscle quality parameters. (A) A positive correlation between VBQ score and DFI, with a coefficient of 0.647 (P<0.001). (B) No significant correlation between VBQ score and CSA (P=0.7495). VBQ, vertebral bone quality; CSA, cross-sectional area; DFI, degree of fat infiltration.

Discussion

Studies have confirmed that OVCF is strongly correlated with VBQ score, as well as with the CSA and DFI of paravertebral muscles (9,10,19,20). These MRI-based bone and muscle quality parameters are a particularly amenable to measurement and have greater applicability than does DEXA. However, the relationship between these parameters and SVF remains unclear. Our study discovered that VBQ score and DFI were independent risk factors for predicting SVF, and their combined predictive value was significantly enhanced. Meanwhile, CSA was not found to be an independent risk factor for SVF.

The incidence of SVF after PVA remains controversial, which can be attributed to the varying follow-up periods and definitions adopted by different researchers. SVF can be categorized into clinical and radiological subsequent fracture. In a meta-analysis that included 32 prospective studies, the occurrence of clinical subsequent fracture after PVA was 16.02%, while the rate of radiographic fracture was notably higher, at 20.91% (16). In our study, the clinical SVF was adopted, and the incidence rate (14.19%) was consistent with previous studies (3,15,16,21). This high rate emphasizes the importance of the early identification of patients at high risk for SVF following PVA treatment.

Since Ehresman et al. first proposed the concept of the VBQ score in 2020, numerous studies have demonstrated that it is an effective indicator of bone quality, showing a strong correlation with DEXA-based T-scores, which are widely used to assess BMD (7,8,22-26). The VBQ score was derived from MRI-based assessments that quantify the fat content within the bone marrow. Higher VBQ scores indicate greater fat infiltration in the bone marrow, which is associated with lower bone density and an increased risk of fractures, similar to lower T-scores in DEXA. Özmen et al. further validated the VBQ score and identified an optimal threshold value of 2.7 for distinguishing between normal bone quality and osteoporosis (23). This threshold allows the VBQ score to serve as a reliable tool for opportunistic screening, with the ability to rule out osteoporosis and identify patients who may need further evaluation. The VBQ score’s ease of use and high reproducibility make it an attractive alternative to DEXA, particularly in settings where MRI is already being used for other diagnostic purposes.

In addition, recent studies have demonstrated that the VBQ score exhibits strong performance in predicting post-spinal surgery complications in patients with osteoporosis (11-13). Contrary to a previous study by Li et al., which examined the characteristics of VBQ score in older adult patients with vertebral fragility fractures, our study indicated that the VBQ score was a significant independent predictor of SVF according to the multivariable logistic regression analysis (10). In the ROC analysis, VBQ score demonstrated good predictive ability for SVF (AUC =0.863). According to the Youden index, the optimal cutoff value of VBQ was 3.49. The odds of experiencing SVF in patients with a VBQ score higher than the threshold are 3.66 times higher compared to those with lower scores.

MRI not only allows for the measurement of VBQ score to assess bone quality but can also simultaneously measure CSA and DFI to accurately evaluate the condition of paravertebral muscles (27). As a crucial structure in providing stability for the vertebral column, the paravertebral muscles mainly include the multifidus muscle and erector spinae muscle. The degeneration of paravertebral muscles, also known as sarcopenia, is characterized by a decrease in muscle volume, muscle weakness, and infiltration of fat tissues (28). The results showed that the DFI was significantly elevated in the SVF group (P<0.05), whereas there was no statistical difference in CSA between the two groups. Logistic regression analysis also revealed that the DFI was an independent factor associated with the occurrence of SVF. However, the CSA did not show significant statistical significance in the multivariable analysis. The reason for this observation is that the paravertebral muscles undergo a transformation where muscle tissue is progressively infiltrated by adipose tissue as degeneration occurs (29). Consequently, although some patients may experience almost complete replacement of muscle by fat, the overall CSA remains largely unchanged. These findings align with those of Xiong et al., who also identified the degeneration of paraspinal muscles, particularly through increased fatty infiltration, as a critical predictor of SVF (30).

To further investigate the parameters that affect SVF, a Pearson correlation coefficient was used to assess the linear association between VBQ score and both CSA and DFI. The VBQ score was positively correlated with DFI but showed no significant correlation with CSA. These findings align with emerging evidence from recent systematic reviews that highlight the close interplay between bone health and muscle fat infiltration in aging populations (31). This relationship reflects the phenomenon of osteosarcopenia, where bone and muscle tissue undergo concurrent degeneration (32). In patients with osteoporosis, the substitution of vertebral bone marrow with fat tissue is detectable on T1-weighted MRI as increased SI, which in turn elevates VBQ score. Concurrently, the atrophy and fat infiltration of paravertebral muscles increase DFI. This fat infiltration in bone marrow and muscles suggests a shared mechanism driving both bone and muscle deterioration. Evidence indicates that osteosarcopenia is associated with factors such as genetics, age, inflammation, obesity, and various biochemical pathways (33). This overlapping pathophysiology further supports the correlations and highlights the importance of considering both bone and muscle quality as a whole in assessing fracture risk.

The risk of SVF is significantly higher when osteoporosis and sarcopenia coexist. In our ROC analysis, incorporating both VBQ score and DFI yielded a higher AUC (AUC =0.925) compared to that of VBQ score alone (AUC =0.863; P<0.05). Although a specific cutoff value could not be determined for this combined model, its clinical significance lies in its enhanced predictive accuracy, which may enable the more precise identification of high-risk patients. In practice, this model can be integrated into risk stratification tools such as nomograms or risk calculators, allowing clinicians to provide individualized assessments and tailor interventions accordingly. For example, patients identified as high risk may benefit from intensified osteoporosis management and targeted muscle-strengthening programs, which can reduce the incidence of SVF. This comprehensive approach holds the potential to not only prevent subsequent fractures more effectively but also improve the overall quality of life for the aging population.

Limitations

This study involved several limitations which should be addressed. First, we employed a cross-sectional, retrospective analysis, and the sample size of the SVF group was relatively small. Therefore, a comprehensive and sizable prospective investigation is necessary to confirm our findings. Additionally, previous studies have identified severe low back pain, vertebral body height restoration, and the use of soft braces as contributing factors to SVF (1,2). However, these factors were not included as covariates in our study, which could have potentially introduced bias to the results. Furthermore, in this study, we only identified clinical SVF in patients who presented with new-onset back pain rather than all cases of SVF (radiological and clinical SVF) as confirmed by regular imaging follow-up, which means asymptomatic SVF might have been missed, potentially leading to an underestimation of the actual SVF incidence.

Conclusions

The VBQ score and DFI were identified as independent predictors for clinical SVF after percutaneous vertebral augmentation. The combination of VBQ score and DFI significantly improves the predictive accuracy for SVF. Higher VBQ score and DFI level were associated with an increased risk of experiencing SVF, highlighting the importance of these MRI-based parameters in clinical assessment. Additionally, our findings indicated a significant positive correlation between VBQ score and DFI, suggesting a concurrent deterioration of bone and muscle quality in patients at risk for SVF.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-24-712/rc

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-24-712/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the Ethics Committee of Xuzhou Central Hospital (No. XZXY-LK-20210310-0134). Individual consent for this retrospective analysis was waived.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Inose H, Kato T, Ichimura S, Nakamura H, Hoshino M, Togawa D, et al. Risk factors for subsequent vertebral fracture after acute osteoporotic vertebral fractures. Eur Spine J 2021;30:2698-707. [Crossref] [PubMed]
2. Li YA, Lin CL, Chang MC, Liu CL, Chen TH, Lai SC. Subsequent vertebral fracture after vertebroplasty: incidence and analysis of risk factors. Spine (Phila Pa 1976) 2012;37:179-83. [Crossref] [PubMed]
3. Zou J, Mei X, Zhu X, Shi Q, Yang H. The long-term incidence of subsequent vertebral body fracture after vertebral augmentation therapy: a systemic review and meta-analysis. Pain Physician 2012;15:E515-22.
4. Chen Z, Song C, Chen M, Li H, Ye Y, Liu W. What are risk factors for subsequent fracture after vertebral augmentation in patients with thoracolumbar osteoporotic vertebral fractures. BMC Musculoskelet Disord 2021;22:1040. [Crossref] [PubMed]
5. Chen Z, Yao Z, Wu C, Wang G, Liu W. Assessment of clinical, imaging, surgical risk factors for subsequent fracture following vertebral augmentation in osteoporotic patients. Skeletal Radiol 2022;51:1623-30. [Crossref] [PubMed]
6. Kanis JA. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report. WHO Study Group. Osteoporos Int 1994;4:368-81. [Crossref] [PubMed]
7. Chen Z, Lei F, Ye F, Yuan H, Li S, Feng D. MRI-based vertebral bone quality score for the assessment of osteoporosis in patients undergoing surgery for lumbar degenerative diseases. J Orthop Surg Res 2023;18:257. [Crossref] [PubMed]
8. Ehresman J, Pennington Z, Schilling A, Lubelski D, Ahmed AK, Cottrill E, Khan M, Sciubba DM. Novel MRI-based score for assessment of bone density in operative spine patients. Spine J 2020;20:556-62. [Crossref] [PubMed]
9. Jeon I, Kim SW, Yu D. Paraspinal muscle fatty degeneration as a predictor of progressive vertebral collapse in osteoporotic vertebral compression fractures. Spine J 2022;22:313-20. [Crossref] [PubMed]
10. Li W, Zhu H, Liu J, Tian H, Li J, Wang L. Characteristics of MRI-based vertebral bone quality scores in elderly patients with vertebral fragility fractures. Eur Spine J 2023;32:2588-93. [Crossref] [PubMed]
11. Gao Y, Ye W, Ge X, Wang H, Xiong J, Zhu Y, Wang Z, Wang J, Tang P, Liu W, Cai W. Assessing the utility of MRI-based vertebral bone quality (VBQ) for predicting lumbar pedicle screw loosening. Eur Spine J 2024;33:289-97. [Crossref] [PubMed]
12. Huang Y, Chen Q, Liu L, Feng G. Vertebral bone quality score to predict cage subsidence following oblique lumbar interbody fusion. J Orthop Surg Res 2023;18:258. [Crossref] [PubMed]
13. Kuo CC, Soliman MAR, Aguirre AO, Ruggiero N, Kruk M, Khan A, Ghannam MM, Almeida ND, Jowdy PK, Smolar DE, Pollina J, Mullin JP. Vertebral Bone Quality Score Independently Predicts Proximal Junctional Kyphosis and/or Failure After Adult Spinal Deformity Surgery. Neurosurgery 2023;92:945-54. [Crossref] [PubMed]
14. Kasperk C, Grafe IA, Schmitt S, Nöldge G, Weiss C, Da Fonseca K, Hillmeier J, Libicher M, Sommer U, Rudofsky G, Meeder PJ, Nawroth P. Three-year outcomes after kyphoplasty in patients with osteoporosis with painful vertebral fractures. J Vasc Interv Radiol 2010;21:701-9. [Crossref] [PubMed]
15. Yi X, Lu H, Tian F, Wang Y, Li C, Liu H, Liu X, Li H. Recompression in new levels after percutaneous vertebroplasty and kyphoplasty compared with conservative treatment. Arch Orthop Trauma Surg 2014;134:21-30. [Crossref] [PubMed]
16. Sun HB, Jing XS, Tang H, Hai Y, Li JJ, Shan JL, Wang DC. Clinical and radiological subsequent fractures after vertebral augmentation for treating osteoporotic vertebral compression fractures: a meta-analysis. Eur Spine J 2020;29:2576-90. [Crossref] [PubMed]
17. Crawford RJ, Cornwall J, Abbott R, Elliott JM. Manually defining regions of interest when quantifying paravertebral muscles fatty infiltration from axial magnetic resonance imaging: a proposed method for the lumbar spine with anatomical cross-reference. BMC Musculoskelet Disord 2017;18:25. [Crossref] [PubMed]
18. Fortin M, Battié MC. Quantitative paraspinal muscle measurements: inter-software reliability and agreement using OsiriX and ImageJ. Phys Ther 2012;92:853-64. [Crossref] [PubMed]
19. Ehresman J, Schilling A, Yang X, Pennington Z, Ahmed AK, Cottrill E, Lubelski D, Khan M, Moseley KF, Sciubba DM. Vertebral bone quality score predicts fragility fractures independently of bone mineral density. Spine J 2021;21:20-7. [Crossref] [PubMed]
20. Huang W, Cai XH, Li YR, Xu F, Jiang XH, Wang D, Tu M. The association between paraspinal muscle degeneration and osteoporotic vertebral compression fracture severity in postmenopausal women. J Back Musculoskelet Rehabil 2023;36:323-9. [Crossref] [PubMed]
21. Park JS, Park YS. Survival analysis and risk factors of new vertebral fracture after vertebroplasty for osteoporotic vertebral compression fracture. Spine J 2021;21:1355-61. [Crossref] [PubMed]
22. Huang W, Gong Z, Wang H, Zheng C, Chen Y, Xia X, Ma X, Jiang J. Use of MRI-based vertebral bone quality score (VBQ) of S1 body in bone mineral density assessment for patients with lumbar degenerative diseases. Eur Spine J 2023;32:1553-60. [Crossref] [PubMed]
23. Özmen E, Biçer O, Meriç E, Circi E, Barış A, Yüksel S. Vertebral bone quality score for opportunistic osteoporosis screening: a correlation and optimal threshold analysis. Eur Spine J 2023;32:3906-11. [Crossref] [PubMed]
24. Kadri A, Binkley N, Hernando D, Anderson PA. Opportunistic Use of Lumbar Magnetic Resonance Imaging for Osteoporosis Screening. Osteoporos Int 2022;33:861-9. [Crossref] [PubMed]
25. Li R, Yin Y, Ji W, Wu X, Jiang H, Chen J, Zhu Q. MRI-based vertebral bone quality score effectively reflects bone quality in patients with osteoporotic vertebral compressive fractures. Eur Spine J 2022;31:1131-7. [Crossref] [PubMed]
26. Lin W, He C, Xie F, Chen T, Zheng G, Yin H, Chen H, Wang Z. Assessment of bone density using the 1.5 T or 3.0 T MRI-based vertebral bone quality score in older patients undergoing spine surgery: does field strength matter? Spine J 2023;23:1172-81. [Crossref] [PubMed]
27. Liu S, Schmidt H, Ziegeler K, Zhang T, Yang D, Taheri N, Pumberger M, Becker L. Inter-software and inter-threshold reliability of quantitative paraspinal muscle segmentation. Eur Spine J 2024;33:369-78. [Crossref] [PubMed]
28. Schönnagel L, Zhu J, Camino-Willhuber G, Guven AE, Tani S, Caffard T, Haffer H, Muellner M, Chiapparelli E, Arzani A, Amoroso K, Moser M, Shue J, Tan ET, Carrino JA, Sama AA, Cammisa FP, Girardi FP, Hughes AP. Relationship between lumbar spinal stenosis and axial muscle wasting. Spine J 2024;24:231-8. [Crossref] [PubMed]
29. Chen Z, Shi T, Li W, Sun J, Yao Z, Liu W. Role of paraspinal muscle degeneration in the occurrence and recurrence of osteoporotic vertebral fracture: A meta-analysis. Front Endocrinol (Lausanne) 2022;13:1073013. [Crossref] [PubMed]
30. Xiong Y, Zhang C, Chen X, Wu L, Liang S, Zhang Y, Huang J, Guo W, Zeng X, Xu F. Prediction of Subsequent Vertebral Fracture After Acute Osteoporotic Fractures from Clinical and Paraspinal Muscle Features. Calcif Tissue Int 2024;114:614-24. [Crossref] [PubMed]
31. Mallio CA, Volterrani C, Bernetti C, Stiffi M, Greco F, Beomonte Zobel B. Exploring the interplay between paraspinal muscular status and bone health in osteoporosis and fracture risk: a comprehensive literature review on computed tomography (CT) and magnetic resonance imaging (MRI) studies. Quant Imaging Med Surg 2024;14:4189-201. [Crossref] [PubMed]
32. Li W, Wang F, Chen J, Zhu H, Tian H, Wang L. MRI-based vertebral bone quality score is a comprehensive index reflecting the quality of bone and paravertebral muscle. Spine J 2024;24:472-8. [Crossref] [PubMed]
33. Clynes MA, Gregson CL, Bruyère O, Cooper C, Dennison EM. Osteosarcopenia: where osteoporosis and sarcopenia collide. Rheumatology (Oxford) 2021;60:529-37. [Crossref] [PubMed]