MR‑based proton density
fat fractions of the vertebral bone marrow and paraspinal muscle are associated
with BMD from QCT in patients with LBP
Ze Li1, Junrong Chen2, and Huilou Liang3 1Chengdu Sport University, Chengdu, China, 2Sichuan Provincial Orthopedics Hospital, Chengdu, China, 3GE HealthCare MR Research, Beijing, China
Synopsis
Keywords: Bone, Bone, Fat
Motivation: The relationship between vertebral bone mineral density (BMD) and fat composition of adjacent vertebral body and paraspinal muscle in patients with low back pain (LBP) remains unclear.
Goal(s): To investigate the association between vertebral BMD and the fat fractions of vertebral bone marrow and paraspinal muscle in LBP patients.
Approach: A retrospective study was conducted on LBP patients who underwent both quantitative computed tomography (QCT) and chemical shift-encoded MRI examinations.
Results: Age, fat fractions of bone marrow and psoas major are independent factors that influence vertebral bone mineral density.
Impact: The
fat fraction of vertebral bone marrow and paraspinal muscles independently
influences bone mineral density. The MR-based fat fractions may potentially predict
osteoporosis and osteoporotic fractures without radiation, providing a safer
way to diagnose osteoporotic vertebral fractures and associated complications.
Introduction
As a common occurrence in the
middle-aged and elderly population, osteoporosis leads to increased bone
fragility in the vertebral body, and causes the development of osteoporotic
vertebral fractures1. Osteoporotic fractures are
common but difficult to detect. Low back pain (LBP) is a common musculoskeletal
disorder that affects a large proportion of the population worldwide. One of the causes of LBP is osteoporotic vertebral fractures,
which can lead to pain and disability2. Therefore, it’s crucial to
identify patients at high risk for these fractures, which is often reflected by
bone mineral density (BMD) from quantitative computed tomography (QCT)3. Recent studies demonstrated
that fatty tissue may play a role in bone metabolism4. MRI is better than CT at
distinguishing fat from non-fat components5. Chemical shift-encoded MRI
technique provides a radiation-free way to accurately measure the proton
density fat fraction (PDFF) of muscle and bone6,7. Several studies have
investigated the relationship between vertebral bone marrow fat, BMD, and
vertebral fractures8. However, there is still a
need to explore the correlation between bone and fat composition in LBP
patients. Therefore, this study aimed to investigate the association between
vertebral volumetric BMD (vBMD) and PDFF of vertebral bone marrow (BM) and paraspinal
muscle (PSM) in LBP patients.
Methods
Patients: After IRB-approved written
informed consent was obtained, LBP patients were scanned on 3.0 T MRI (SIGNA
Architect, GE Healthcare) and CT (Siemens Somatom Definition AS+) to measure
PDFF of BM and PSM, and vertebral vBMD, respectively. Finally, 509 subjects
(24-76 years, 267 females) were enrolled in this retrospective study. MR imaging parameters: The MR scan included axial T2
FSE (0.5×0.8×3.0mm3, TR=3224ms, TE=120ms), axial
IDEAL-IQ (2.0×2.0×2.0mm3, TR=8.0ms, TE=3.6ms, number of TEs=6, number
of shots=2).
Data processing: CT data were transmitted to
the QCT Pro workstation to measure the mean vBMD of two vertebral bodies at L1
and L2 levels (Figure 1A). The IDEAL-IQ images were processed in AW4.7
workstation to calculate PDFF map. ROIs of the L1-L5 vertebrae were obtained in
sagittal plane, avoiding cortical bone and outlining the entire vertebral cancellous
bone (Figure 1B). The mean PDFF values of bilateral paraspinal muscles
including multifidus (MF), erector spinae (ES) and psoas major (PS) were
obtained on a ROIs basis at the central level of L1/2 to L5/S1 (Figure 1C),
respectively. Statistical analysis: All analyses were performed
using SPSS 22.0 software. All patients were divided into normal bone density
(vBM>120mg/cm3), osteopenia (vBMD 80 to 120mg/cm3) and
osteoporosis groups (vBMD<80mg/cm3) as recommended by the ISCD in
20079. The normality of data was
analyzed by P-P Chart. The differences of PDFF, age and BMI
among the three vBMD groups were tested by one-way ANOVA with post hoc analysis
(LSD). The relationship between PDFF and vBMD, age, BMI were analyzed using
Pearson correlation coefficients. The relationship between vBMD and PDFF was
further tested using multiple linear regression with sex, age and BMI
variables. Differences were considered statistically significant at p<0.
05.
Results
As shown in Table 1 and Figure 2, vBMD, age,
PDFF were significantly different among three groups (p<0.001). As
shown in Table 2, correlation analysis showed that vBMD was moderately to
severely negatively correlated with PDFF (p<0.01), and was especially
pronounced with BM PDFF (r=-0.655; p<0.01). Age was moderately
positively correlated with PDFF (p<0.01). However, there was no
significant correlation between BMI and PDFF (p>0.05). As shown in
Table 3, the R2 and adjusted R2 of multiple linear regression model were 0.559
and 0.553, respectively (p<0.05). The model revealed that age (β=-0.367, p<0.001), BM PDFF (β=-0.424, p<0.001)
and PS PDFF (β=-0.075, p=0.039) were independent factors of vBMD. However, MF
PDFF (p=0.397) and ES PDFF (p=0.328) were not associated with
vBMD.
Discussion
The
innovative aspect of this study was to investigate the relationship between
lumbar vBMD and fat fractions of vertebral bone marrow and paraspinal muscle in
LBP patients. Age, BM PDFF and PS PDFF are independent factors of vBMD. PDFF
maybe a potential imaging biomarker for assessing the risk of osteoporosis,
which provides a safer way to predict and prevent fractures in advance.
Conclusion
Our
results demonstrated that in patients with LBP, PDFF of bone marrow and
paraspinal muscle increased with decreasing vBMD. Age, bone marrow fat
fractions and psoas major fat fractions are independent factors of vBMD. In
the future, it is possible that predicting osteoporosis through vertebral bone
marrow fat fraction may improve the detection rate of osteoporotic fractures,
which is important for reducing osteoporotic vertebral fractures and
complications.
Acknowledgements
No acknowledgement found.
References
1. van den Bergh JP, Szulc P, Cheung AM,
Bouxsein M, Engelke K, Chapurlat R. The clinical application of high-resolution
peripheral computed tomography (HR-pQCT) in adults: state of the art and future
directions. Osteoporos Int. 2021;32(8):1465-1485.
doi:10.1007/s00198-021-05999-z
2. Wáng YXJ, Du EZ, Gong J, Cheng X.
Interpretation of osteoporotic vertebral deformity on frontal view radiographs
of the chest and abdomen: a pictorial review. Quant Imaging Med Surg.
2021;11(1):423-442. doi:10.21037/qims-2020-28
3. Mao SS, Li D, Syed YS, et al. Thoracic
Quantitative Computed Tomography (QCT) Can Sensitively Monitor Bone Mineral
Metabolism: Comparison of Thoracic QCT vs Lumbar QCT and Dual-energy X-ray
Absorptiometry in Detection of Age-relative Change in Bone Mineral Density. Acad
Radiol. 2017;24(12):1582-1587. doi:10.1016/j.acra.2017.06.013
4. Goetz TG, Nair N, Shiau S, et al. In
premenopausal women with idiopathic osteoporosis, lower bone formation rate is
associated with higher body fat and higher IGF-1. Osteoporos Int.
2022;33(3):659-672. doi:10.1007/s00198-021-06196-8
6. Syväri J, Junker D, Patzelt L,
et al. Longitudinal
changes on liver proton density fat fraction differ between liver segments. Quant
Imaging Med Surg. 2021;11(5):1701-1709. doi:10.21037/qims-20-873
7. Gersing AS, Holwein C, Suchowierski J, et
al. Cartilage T2 Relaxation Times and Subchondral Trabecular Bone Parameters
Predict Morphological Outcome After Matrix-Associated Autologous Chondrocyte
Implantation With Autologous Bone Grafting. Am J Sports Med.
2020;48(14):3573-3585. doi:10.1177/0363546520965987
8. Gassert FT, Kufner A, Gassert FG, et al.
MR-based proton density fat fraction (PDFF) of the vertebral bone marrow
differentiates between patients with and without osteoporotic vertebral
fractures. Osteoporos Int. 2022;33(2):487-496.
doi:10.1007/s00198-021-06147-3
9. Baim S, Binkley N, Bilezikian JP, et al.
Official Positions of the International Society for Clinical Densitometry and
executive summary of the 2007 ISCD Position Development Conference. J Clin
Densitom. 2008;11(1):75-91. doi:10.1016/j.jocd.2007.12.007
Figures
Figure 1. Procedure for calculating vBMD and PDFF of BM
and PSM in lower lumbar spine. vBMD measurement of L1 and L2 on the QCT pro
workstation. ROIs were drawn automatically in the axial, sagittal and coronal
images (A). MRI of IDAEL-IQ sequence showing segmentation for PDFF analysis.
ROIs of the BM were drawn at L1-L5 vertebrae in the sagittal images (B). ROIs
of PSM were manually drawn at the central level of L1/2 to L5/S1, respectively
(C). ROI, regions of interest; vBMD,
volumetric bone mineral density; PDFF, the proton density fat fraction; BM,
bone marrow; PSM, paraspinal muscle.
Table 1. The clinical characteristics and PDFF of
BM and PSM measurements among the normal bone density, osteopenia, and
osteoporosis groups. The p values are from one-way ANOVA. The double
asterisks (**) indicate p<0.01. BMI, body mass index; vBMD,
volumetric bone mineral density; PDFF, the proton density fat fractions; BM,
bone marrow; PSM, paraspinal muscle; MF, multifidus; ES, erector spinae; PS,
psoas major; SD, standard deviation.
Figure 2. PDFF of BM and PSM measurements among
normal bone density, osteopenia, and osteoporosis groups. The p values
are from post hoc analysis. The single asterisk (*) indicates p<0.05.
PDFF, proton density fat fraction; BM, bone marrow; PSM, paraspinal muscle; MF,
multifidus; ES, erector spinae; PS, psoas major.
Table 2. Correlation coefficients between the
PDFF of bone marrow and paraspinal muscle versus vBMD, age and BMI. Pearson
correlation coefficients were used to analyze the correlations among continuous
variables. The double asterisks (**) indicates p<0.01. BMI, body mass
index; vBMD, volumetric bone mineral density; PDFF, the proton density fat
fractions; BM, bone marrow; MF, multifidus; ES, erector spinae; PS, psoas
major.
Table 3. Multiple linear regression analysis
was performed the independent factors related to vBMD. vBMD is the dependent
variable. The single asterisk (*) and double asterisks (**) indicate p<0.05
and p<0.01, respectively. BMI, body mass index; PDFF, the proton
density fat fractions; BM, bone marrow; MF, multifidus; ES, erector spinae; PS,
psoas major; β, Bate; SE standard error; t, t value; VIF, variance
inflation factor.