Quantification of vertebral bone marrow fat fraction using time-interleaved multi-echo gradient-echo water-fat MRI: preliminary experience in children

Stefan Ruschke¹, Amber Pokorney², Holger Eggers³, Jan S. Kirschke⁴, Thomas Baum⁵, Dimitrios C. Karampinos¹, and Houchun Harry Hu²

¹Diagnostic and Interventional Radiology, Technische Universität München, Munich, Germany, ²Radiology, Phoenix Children's Hospital, Phoenix, AZ, United States, ³Philips Research, Hamburg, Germany, ⁴Section of Diagnostic and Interventional Neuroradiology, Technische Universität München, Munich, Germany, ⁵Department of Diagnostic and Interventional Radiology, Technische Universität München, Munich, Germany

Synopsis

In this work, we describe our preliminary clinical experience using a previously reported time-interleaved six-echo gradient-echo (TIMGRE) acquisition for water-fat chemical-shift encoded MRI. The acquisition scheme involved two interleaves that acquired three echoes each with fly-back gradients. The pulse sequence was used to quantify vertebral bone marrow fat fraction in a pilot cohort of 12 pediatric patients (age range: 1-13 years) at 3T with 1.2-1.6 mm in-plane resolution and 1.2-3 mm slices. The knowledge on bone marrow fat fraction may provide insight into adverse effects on bone health later in life, given that there is clinical relevance of vertebral osteoporotic fractures in adults.

Introduction

In recent years, there has been a significant increase in the utilization of MR imaging and spectroscopy to quantify vertebral bone marrow fat content [1-4]. Multiple independent groups have demonstrated that vertebral bone marrow fat content is a useful imaging biomarker in the characterization of bone health [5-6], metabolic disorders [7], and cancer [8-9]. While a majority of the works thus far has been focused in adults and older aging populations, similar vertebral bone marrow fat content data in children remain limited. The capability to robustly and accurately quantify vertebral bone marrow fat content in children may help better understanding the bone development process at an early age [10] and provide insight into potential effects on adverse bone health later in life. However, pediatric imaging usually requires smaller voxel sizes than adult imaging, leading to increased echo time steps in water-fat imaging. The objective of this work was to demonstrate the feasibility of a previously reported time-interleaved multi-echo gradient-echo (TIMGRE) water-fat “Dixon” pulse sequence [11], allowing the flexible selection of echo times independently of voxel size for robust water-fat decomposition in pediatric imaging. Furthermore, a complex-based multi-peak water-fat model with single T2* estimation in TIMGRE ensures robust proton-density fat fraction determination.

Materials and Methods

Figure 1 depicts the TIMGRE sequence. The 3D acquisition scheme acquires six echoes using fly-back gradients in two interleaves with three echoes each. The interleaf offset time was chosen to achieve a constant echo spacing between consecutive echoes. A previously described phase correction procedure [11] was used. It corrects phase inconsistencies from concomitant fields between the two interleaves and linear phase variations along the readout direction.

All studies were performed on two 3T Ingenia MR whole-body systems (Philips Healthcare). The built-in posterior coil array in the scanner table was used for signal reception. The addition of the TIMGRE water-fat pulse sequence was approved by the institutional review board. The cohort of pediatric patients studied thus far (4M, 8F, age range: 1-13 years, 5.1+/-3.2 years) were receiving routine spine MRI examinations for clinically indicated reasons and general anesthesia was used in all cases in accordance with institutional protocol. Typical TIMGRE imaging parameters used in this study were: 16-26 sagittal slices, slice thickness: 1.2 to 3 mm, S/I and A/P FOV: 360 to 500 mm and 140 to 240 mm, in-plane voxel resolution: 1.2-1.6 mm, no SENSE acceleration and no partial Fourier sampling, flip angle: 3 degrees, TR: 8.8-12.9 ms, effective TE spacing 1.2-1.3 ms, first TE: 1.5-1.7 ms, total scan time: 2-3 min.

Results

Figure 2 illustrates representative whole-spine examples from five of the 12 patients, using a color scale from 0-100% for the proton-density fat fraction. Figure 3 plots representative average fat fraction values measured from regions-of-interests drawn along the five lumbar vertebra sections in the 12 patients. The data is further split into two groups, one with patients (n=9) that showed no spinal abnormalities on their exam (black symbols), and one with patients (n=3) that exhibited clinically relevant pathology (red symbols). Particularly in the former group, the data suggests an age dependence in the vertebral bone marrow fat fraction (r²= 0.58 linear fit, r² = 0.68 exponential fit) that is visible by MR imaging and manifests early in childhood development.

Discussion and Conclusion

In this pilot work, we have demonstrated preliminary feasibility of a high spatial resolution TIMGRE water-fat pulse sequence for vertebral bone marrow fat quantification in a small cohort of pediatric patients. Our study is ongoing as we continue to accumulate more pediatric data to assess age- and gender- dependence on vertebral bone marrow fat content [2]. The age dependence of the bone marrow fat fraction has been previously studied extensively in adults [2, 12], but not in children. In adults, it is known that age-related increase in vertebral marrow adiposity is associated with bone loss and reduction in bone mineral density [5]. In children, there is limited literature on the red to yellow marrow conversion in long bones [13], but the association between bone formation and marrow conversion in the spine remains largely unknown. A non-invasive method to monitor age-related changes in vertebral bone marrow fat fraction in children will help in understanding the relationship between vertebral bone acquisition and marrow adipogenesis. The knowledge gained in children will provide insight into potential adverse effects on bone health later in life, given that there is a strong clinical relevance of vertebral osteoporotic fractures in adults.

Acknowledgements

The authors acknowledge Philips Healthcare for research support and the German Academic Exchange Service (DAAD) for support through the “Doktorandenstipendium”.

References

[1] Bredella MA, et al., Marrow adipose tissue quantification of the lumbar spine by using dual-energy CT and single-voxel (1)H MR spectroscopy: A feasibility study. Radiology 2015; 227(1):230-235.

[2] Baum T, et al., Assessment of whole spine vertebral bone marrow fat using chemical shift-encoding based water-fat MRI. J Magn Reson Imaging 2015; 42(4):1018-1023.

[3] Martin J, et al., Rapid determination of vertebral fat fraction over a large range of vertebral bodies. J Med Imaging Radiat Oncol 2014; 58(2):155-163.

[4] Roldan-Valadez E, et al., Gender and age groups interactions in the quantification of bone marrow fat content in lumbar spine using 3T MR spectroscopy: a multivariate analysis of covariance (Mancova). Eur J Radiol 2013; 82(11):e697-702.

[5] Griffith JF, et al., Vertebral marrow fat content and diffusion and perfusion indexes in women with varying bone density: MR evaluation. Radiology 2006; 241(3):831-838.

[6] Karampinos D, et al., Association of MRS-based vertebral bone marrow fat fraction with bone strength in a human in vitro model. J Osteoporos 2015; 2015:152349. PMID: 25969766

[7] Mostoufi-Moab S, et al., Adverse fat depots and marrow adiposity are associated with skeletal deficits and insulin resistance in long-term survivors of pediatric hematopoietic stem cell transplantation. J Bone Miner Res 2015; 30(9):1657-1666.

[8] Schraml C, et al., Multiparametric analysis of bone marrow in cancer patients using simultaneous PET/MR imaging: correlation of fat fraction, diffusivity, metabolic activity, and anthropometric data. J Magn Reson Imaging 2015; 42(4):1048-1056.

[9] Kim YP, et al., Differentiation between focal malignant marrow-replacing lesions and benign red marrow deposition of the spine with T2*-corrected fat-signal fraction map using a three-echo volume interpolated breath-hold gradient echo Dixon sequence. Korean J Radiol 2014; 15(6):781-791.

[10] Rosen CJ, et al., Marrow fat and the bone microenvironment: developmental, functional, and pathological implications. Crit Rev Eukaryot Gene Expr 2009; 19(2):109-124.

[11] Ruschke S, et al., Addressing phase errors in quantitative water-fat imaging at 3T using a time-interleaved multi-echo gradient-echo acquisition. ISMRM 2015, abstract #3657.

[12] Griffith J, et al., Bone marrow fat content in the elderly: a reversal of sex difference seen in younger subjects. J Magn Reson Imaging 2012; 36(1):225-230.

[13] Waitches G, et al., Sequence and rate of bone marrow conversion in the femora of children as seen on MR imaging: are accepted standards accurate? AJR Am J Roentgenol 1994; 162(6):1399-1406.

Figures

Figure 1. Pulse sequence diagram of the TIMGRE sequence. The acquisition scheme used in this work involved two interleaves that acquired three echoes each with fly-back gradients. A constant echo spacing is achieved.

Figure 2. Whole-spine proton density fat-fraction maps in five patients. Age (years) and gender (M,F) are listed. Patients in (A,C) had no spine abnormalities, as noted by the reading radiologist. Patient in (B) has spina bifida and a tethered spinal cord. Patient in (D) had findings of fatty filum at L1-L3 and complaints of enuresis, the reason for the MR exam. Patient in (E) has levoconvex scoliosis. Note the low fat fraction of the vertebral bone marrow in (D,E) in contrast to (C).

Figure 3. Plot of average fat fraction at the levels of L1-L5 (different symbols) from the study cohort. Patients with no abnormal image findings of the spine (n=9, black symbols) are plotted separately from the those who had abnormal spine findings (n=3, red symbols, see B, D, and E in Figure 2). A linear correlation (solid line) and an exponential (dotted line) correlation fit is provided through the black data points.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)

4478