Using Solid-State MRI and a Double-Tuned RF Coil to Quantify Bone Matrix and Mineral Densities in Rat Bones

Abstract

Quantitative information on the composition of bone, specifically the content of calcium phosphate mineral and organic matrix, is essential for accurate diagnosis of metabolic bone diseases such as osteoporosis, osteomalacia, and renal osteodystrophy, as well as for differentiating among these conditions. Conventional MRI fails to provide this information because these substances are solid and, therefore, yield no signal in conventional MRI scans, which typically employ spin or gradient echoes. In this report, we show how phosphorus and proton solid-state MRI yield the desired compositional information in bone specimens with ZTE and WASPI pulse sequences, respectively, coupled with the use of a two-port double-tuned solenoidal RF coil.

Electrical network simulations and construction details of the RF coil are detailed. Electrical performance was simulated using QUCS software to find the circuit component values that minimize reflected power and maximize interport isolation. Phantoms of known composition, as well as ex vivo femurs from normal, low bone density, and vitamin D-deficient rats, were included in the study. A simple correction for B₁ inhomogeneity was applied to achieve quantitative accuracy in the image intensity values.

Bone matrix and mineral densities derived from MRI strongly correlated (R² = 0.84) with chemical analysis, demonstrating the ability to measure compositional differences relevant to osteoporosis and osteomalacia.

1. Introduction

Metabolic bone diseases—including osteoporosis, osteomalacia, and renal osteodystrophy (ROD)—impact more than 55 million Americans, contributing to skeletal fragility, chronic pain, deformities, and increased fracture susceptibility [1-3]. These conditions represent a major public health burden, with estimated healthcare costs exceeding $22 billion annually [4].

Metabolic bone diseases include conditions with underlying abnormalities in bone mineralization, formation, resorption, or a combination [5]. Osteoporosis and related metabolic bone diseases have become increasingly prevalent with age and are a current healthcare concern among postmenopausal women and the elderly population, with a global prevalence of 18.3% [6]. On the other hand, chronic kidney disease (CKD) is also associated with metabolic bone diseases, afflicting approximately 10% of the global population. This condition burdens patients with mineral metabolism alterations, resulting in bone abnormalities [7, 8].

Specifically, chronic kidney disease-mineral bone disorder (CKD-MBD) is a systemic complication of CKD that includes bone turnover, mineralization, volume, linear growth or strength abnormalities, as well as vascular disorders [9]. CKD-MBD involves multiple components in bone metabolism alterations, collectively called renal osteodystrophy [9-11]. ROD can occur in different patterns according to the bone formation process and the cell lines affected. Osteocytes, which represent 95% of all bone cells, occupy the lacunar space and are surrounded by an unmineralized osteoid matrix, becoming important precursors that detect mechanical and metabolic signals to activate either osteoblasts (promoting bone formation) or osteoclasts (promoting bone resorption). In adults, bone tissue requires dynamic remodeling to maintain adequate biomechanical properties, being replaced with new tissue in an organized process orchestrated by multicellular units [12-16]. Hormonal factors, particularly parathyroid hormone (PTH), are critical in regulating calcium, phosphate, vitamin D, and bone metabolism [17]. Extremely high or low PTH levels lead to abnormal bone turnover, causing adynamic bone disease (ABD), osteomalacia, or osteitis fibrosa, all linked to higher mortality and fracture risk [9].

Given the fundamental differences between disorders of bone density (e.g., osteoporosis) and mineralization defects (e.g., osteomalacia, ROD), accurate discrimination between these conditions is crucial for effective clinical management. Currently, diagnosis and screening rely primarily on serum biomarkers and dual-energy X-ray absorptiometry (DXA), yet both modalities present significant limitations [8, 18-25]. The World Health Organization (WHO) defines fracture risk based on areal bone mineral density (aBMD, g·cm⁻²), as measured by DXA and referenced to a young adult norm. However, aBMD is not a reliable predictor of fracture risk and lacks the specificity to distinguish between different types of bone pathology [26-30].

Magnetic resonance imaging (MRI) offers the potential to image both the mineral and matrix phases of bone quantitatively. The obvious choice of MR-active isotopes is ³¹P and ¹H to measure bone mineral and matrix components, respectively [31-33]. Since both phases are solid materials, the MR signals will be characterized by very short T₂. Therefore, solid-state short-T₂ MRI techniques offer the potential for measuring the mineral and matrix quantitatively and noninvasively to diagnose these diseases and monitor treatment progression [34, 35]. While MRI is well established for soft tissue (long T₂s) as a standard diagnostic imaging modality in clinical MRI, bone MRI is at a relatively early development phase due to a lack of suitable techniques for hard tissue (very short T₂s), owing to intrinsic short spin-spin relaxation times (T₂) resulting in hardware limitations [35, 36]. In bone MRI, the signals are derived from the bone substance—matrix or mineral, where the T_2s are unfavorable. Conventional MRI pulse sequences generally employ spin or gradient echoes and require a multiplicity of switched gradients during the signal readout. However, with their very short T2 signals, solid materials preclude magnetic field gradient switching during the signal readout because of the limited speed at which the gradients may be switched. As a result, pulse sequences that do not employ gradient switching during the signal readout are required.

Our team and others have demonstrated the potential of bone matrix imaging, commonly employing a variety of zero echo time radial k-space acquisition techniques such as Water And fat Suppressed Projection Imaging (WASPI), Ultrashort Echo time (UTE), and Zero Echo Time (ZTE) imaging. WASPI in particular was designed to selectively image the short T₂ solid-state proton signals in the collagen fibers of the matrix by suppressing free water and fat signals within the tissue with various water and fat suppression techniques [31, 37-39]. Wehrli has likewise used ZTE sequences with long-T2 suppression in a study of cortical bone water [40].

Similarly, solid-state ³¹P MRI specifically targets and quantifies phosphorus (³¹P) in bone mineral, and has been tested in preclinical and clinical settings [32, 40-51]. Both ¹H and ³¹P solid-state imaging techniques operate with a single broadband excitation followed by immediate signal acquisition under a static gradient, with successive acquisitions covering k-space in a 3D radial trajectory to detect the short T₂ signals in the solid materials. In the case of ¹H, the broadband excitation is preceded by chemical shift selective liquid water and fat suppression pulses to suppress these substances while retaining the solid matrix signal. The standard Fourier-domain encoding and reconstruction methods are unsuitable for signals acquired under a static gradient. Alternatively, the radial acquisitions in k-space are remapped onto Cartesian coordinates for Fourier transformation, allowing over-sampling for central k-space and providing sufficient spatial resolution for accurate image reconstruction. Integrating high spatial resolution 3D conventional MRI into the protocol provides detailed analyses of cortical and trabecular structures [52-75], aiding in evaluating bone strength and fracture risk when subjected to impact configurations such as falls [65-69, 76, 77].

Advanced MRI techniques are emerging as powerful tools for evaluating bone health, offering isotope-specific insights into both mineral and organic matrix composition [31, 32, 37-51], where our group has demonstrated the utility of solid-state ¹H (proton) and ³¹P (phosphate) MRI in quantifying bone matrix (MAT) and mineral (MIN) densities, providing precise data on tissue composition. These measurements enable the calculation of the extent of bone mineralization (EBM)—the ratio of MIN to MAT—a critical parameter for distinguishing osteomalacia and ROD from osteoporosis. Importantly, EBM and matrix density are not measurable by DXA, CT, or serum biomarkers, and matrix assessment otherwise requires invasive bone biopsy.

To acquire proton and phosphorus signals, various RF coil designs have been used in multinuclear MR imaging and spectroscopy. For measurements that require the recording of signals from two isotopes from the same specimen or body part, the optimum design choice is for the same coil to excite and acquire the signals of both isotopes to guarantee that the imaged volumes and spatial dependence of the RF fields match at both frequencies.

In this study, we report the performance of a solenoidal double-tuned single RF coil developed for small rat bone specimens at 7T field strength in a preclinical magnet. We used femurs from control, ovariectomized, and partially nephrectomized plus vitamin D-deficient diet Sprague-Dawley rats. The latter two models were used to simulate osteoporosis and osteomalacia.

2. Methods

MRI experiments were performed on a preclinical Bruker 7T, 30 cm horizontal bore MRI scanner (Bruker Billerica, MA, USA) equipped with a 450 mT/m gradient system and Paravision 5.1 and 6.0.1 with approval from the BIDMC Institutional Animal Care and Use Committee. Image post-processing and data analysis were done using an in-house user interface package developed in MATLAB [78].

2.1. Probe-Head Design

Figure 1a shows the mechanical design of the DTSC, based on the design by Cross et al. [79]. The tuning and matching capacitors are adjustable while the coil is in place in the magnet, and their settings are easily reproduced using lockable multi-turn dials with digital readouts.

Figure 1.

a) Mechanical design of the coil. b) QUCS schematic for simulation of the coil tuning circuit. The proton port is on the left (V1), and the phosphorus port is on the right (V2).

2.2. Circuit and S-Parameter Simulations with QUCS

Figure 1b shows the schematic diagram of the tuning and matching circuit for the QUCS (Quite Universal Circuit Simulator) [80] software package that was used to simulate the electrical performance of the coil. The purpose of the simulation was to find a compromise between optimizing the coil performance by maximizing its current, minimizing the reflected power at both ports, and maximizing the inter-port isolation by varying the component values. Equal weight was given to the target optimizations. The resonance frequencies of ¹H and ³¹P were 300.131 MHz and 121.123 MHz, respectively. The C1 and C3 are matching and tuning capacitors (0.8pF - 38pF) of ω_HF; C7 and C9 are tuning and matching capacitors (0.8pF - 38pF) of ω_LF. The C4, C5 (3.7pF) and C2, C6, C8 (5.6 pF) are fixed capacitors in the symmetric and lumped circuits, respectively, balancing and isolating the two channels, and L2, L3, and L4 are three inductor coils (50 nH) in the three trap circuits.

Figure 2 shows a scattering parameters plot simulated with the QUCS for the double-tuned single coil with excellent isolation between the two frequencies. In the QUCS, the ¹H and ³¹P channels were set to the respective resonance frequencies of 300.131 MHz and 121.123 MHz and matched to 50Ω impedance where S11 = 33.5 and S22 = 34. The blue peaks represent ³¹P, and the red peaks represent ¹H resonance frequencies at 7T. We used commercially available fixed and variable non-magnetic high ‘Q’ capacitors. The values of individual parameters have been chosen depending on their availability. Figure 2 shows excellent isolation between ¹H (33.5 dB) and ³¹P (34dB) resonance frequencies and successful tuning and matching with the following values: C1=1.15 pF, C3=6.8 pF, C7=8.8 pF, and C9=23 pF (Figure 1b).

Figure 2.

The S-parameters plot obtained using the QUCS simulation program shows excellent isolation between ¹H (33.5 dB) and ³¹P (34dB) resonance frequencies. The blue peaks represent ³¹P, and the red peaks represent ¹H resonance frequencies at 7T.

2.3. Design of PCB and Solenoid Coil

Figure 3 shows the PCB layout designed with the EAGLE freeware (a) and the physical PCB developed in the laboratory (b).

Figure 3.

a) The PCB layout for a 7T double-tuned single solenoid coil designed by EAGLE, and b) The PCB was fabricated in the laboratory with all the electronic components.

Given the size of the rat femurs, the dimensions of the solenoid coil, L1 (20 mm length, 10 mm diameter, and four turns), were chosen to accommodate both cortical and trabecular bones with a better filling factor (distal epiphysis of the rat femurs fills 10 mm NMR tube, and the tube fits tightly in the RF coil). The coil was designed with 16-gauge copper wire for mechanical stability and to withstand high-power RF pulses. The coil's inductance (110 nH) was chosen for better isolation between ¹H and ³¹P ports (Figure 1b). All parameters, particularly length, inner and outer diameters of the solenoid coil, the gauge of the copper wire (diameter), and the number of turns, including the coil pitch, were chosen for the best filling factor to achieve optimum SNR in the solid-state ¹H and ³¹P short-T₂ rat bone imaging.

2.4. Double-Tuned RF Probe-Head

The probe head was constructed with non-magnetic variable and fixed ceramic, high-Q, low heat, low-noise, and non-magnetic capacitors (Voltronics Corporation, Denville, NJ, USA). The PCB layout was simulated using Easily Applicable Graphical Layout Editor (EAGLE, CadSoft Computer, Germany) [81], fabricated and secured to the body of the probe head with non-magnetic and non-metallic screws in adjustable slots for iso-centering the RF coil in x- and y- directions. The variable tuning and matching capacitors were connected with lockable 4-digit dials to tune and match ¹H and ³¹P channels with two different resonance frequencies (Figure 4b). The ¹H and ³¹P channels coupled to the solenoid coil were connected with high-precision semi-rigid coaxial cables (Huber+Suhner Inc., Herisau, Switzerland) for maximum transmission and reception of RF and MR signals to reduce parasitic capacitance and minimize radiation tendency from the RF coil. The RF coil has detuning circuits with three traps for isolating and enabling transmit-receive operations in the ¹H and ³¹P channels. The design ensures maximum RF delivery to rat femurs without reflected residual power [82].

Figure 4.

a) Double-tuned single solenoid coil assembly with PCB,¹H/³¹P channels, solenoid coil (b). Lockable 4-digit dials.

2.5. Linearity, Resolution, and B₁ Maps with Phantoms

Two sets of three equal-density (1:1:1) polyethylene glycol (PEG), (¹H) and hydroxyapatite (HA), (³¹P) pellets were prepared to evaluate the B₁ field of the DTSC and the partial volume effects in ¹H and ³¹P imaging. Another two sets of pellets, one with polyethylene glycol (PEG, representing the organic matrix component of the bone) and the second with hydroxyapatite (HA, representing the mineral component of the bone) in varying mass densities in 1:2:4 ratios, were prepared to check PEG (¹H) and HA (³¹P) pellet image intensities proportional to their physical mass densities. Proton (¹H) and phosphorus (³¹P) B₁ maps were obtained with pure water (H₂O) and pure hydroxyapatite (HA) phantoms using the ZTE pulse sequences used in rat femur imaging to evaluate the homogeneity of the B₁ field. Figure 5 shows coronal, sagittal, and axial profiles of the middle slice of the B1 maps of pure water ¹H (top) and pure hydroxyapatite ³¹P (bottom) phantom images before and after B₁ correction applied with B₁ correction matrices.

Figure 5.

Coronal, sagittal, and axial images of homogeneous water (top) and hydroxyapatite (bottom) phantoms before and after B1 corrections.

2.6. Bone Specimens for Testing and Validation

Rat femurs from skeletally mature female Sprague Dawley (SD) strain (250–275g, 15 weeks old, Charlestown, MA, USA) were chosen and divided into three equally sized groups: control (CON), ovariectomized (OVX), and partially nephrectomized with vitamin D deficient diet (Vit-D). The SD strain is commonly used in musculoskeletal research, and female rats were chosen as osteoporosis is more prevalent in the female sex. The control (CON) group animals were not subjected to surgical or dietary interventions. The animals allocated to the ovariectomized (OVX) group underwent ovariectomy to induce a state of low bone mass and micro-architectural deterioration (similar to osteoporosis). The vitamin D-deficient (Vit-D) group animals underwent a 5/6 nephrectomy. Then, they were placed on a modified diet of 0.4% calcium and 0% vitamin D (modified Basal Diet 5755, TestDiet, Richmond, IN, USA) to induce inadequate bone mineralization. All animals were euthanized via CO₂ inhalation at 8, 10, and 12 weeks for CON, OVX, and Vit-D groups, respectively. These time points were chosen to ensure a state of reduced bone density and mineralization in the two disease groups, respectively [83, 84].

2.7. ¹H and ³¹P MR Spectroscopy and Imaging on Phantoms and Rat Femurs

¹H and ³¹P MR spectroscopy and imaging studies were conducted on liquid (water and phosphoric acid) and solid (PEG and HA pellets) samples to test DTSC performance. Subsequently, proton (¹H) and phosphorus (³¹P) T₁ measurements were carried out on rat femurs with a single pulse (acquisition of free induction decays (FIDs)), a series of progressive saturation, and ZTE imaging sequences (3D radial coverage of k-space) on the 7T preclinical scanner [32, 43, 85]. The signals were averaged over 64 scans for ¹H and 128 scans for ³¹P, with 64 dummy scans in ¹H and ³¹P measurements. Initially, ¹H ZTE imaging was performed on rat femurs covering cortical and trabecular bones with water + fat suppression using VAPOR saturation elements added to the ¹H ZTE sequence [86]. The short-T₂ bone matrix (¹H) signals were acquired with optimized TR values computed from the Ernst angle calculations [87]. Rat femur ¹H images were obtained with SW = 250,000 Hz, TR = 20 ms, radial projections = 13,030, matrix size = 64 × 64 × 64 and FOV = 30 × 30 × 30, No. of averages (NA) = 8 under a fixed gradient of 170 mT/m with 0.468 mm resolution in 35 minutes. Immediately after completing ¹H ZTE imaging, the long-T₁ bone mineral (³¹P) images were acquired with ³¹P ZTE sequence on the same rat femur with TR = 1000 ms, number of averages (NA) = 2, and the rest of the parameters were the same as ¹H ZTE imaging with 0.468 mm resolution, obtained in 7 hours. For quantitative results, an internal dual calibration phantom (DP) was kept in all ¹H and ³¹P rat femur imaging measurements. ¹H and ³¹P ZTE imaging was carried out on cortical and trabecular bones together in all control (CON), ovariectomized (OVX), and vitamin-D deficient (Vit-D) rat femurs to compute bone matrix and mineral densities [88].

2.8. Density Calculations with 3-Density Pellets and Dual Calibration Phantom

The physical and MRI-derived densities of the three-density PEG and HA phantoms and the dual calibration phantom pellets were plotted together, which confirmed linear correlations ($r_{\text{PG}}^2$ = 0.99 and $r_{\text{HA}}^2$ = 0.97) [78, 86, 88, 89]. Each rat femur specimen's bone matrix density (BMD_MAT) was determined by dividing the total sum of the ¹H MRI voxel intensities by the total bone volume (TV). Similarly, bone mineral densities (BMD) were computed by dividing the total sum of ³¹P MRI voxel intensities by the total volume (TV) of the same rat bone. The MRI-derived bone matrix and mineral densities were converted to DP-PEG-HA phantom equivalent mass densities using linear regressions and converted to true bone matrix (MAT) and mineral (MIN) densities [32, 86, 89] to calculate EBM.

2.9. Chemical Analysis

Gold standard chemical analysis was performed on all cortical and trabecular rat bone specimens. Bone tissue volumes were determined using a gas pycnometer (Accupyc 1330, Micromeritics, Norcross, GA, USA). The dry bone mass of each specimen was determined by drying it in an oven at 75 °C for 24 hours. Ash mass from each specimen was obtained after they were turned to ash at 600° C to burn off the organic matrix over 96 hours until no change in mass was observed. Tissue mineral density was calculated as the mass of ash divided by the volume of bone tissue (TV) measured from pycnometer studies. Tissue matrix density was calculated as (dry bone mass–ash mass)/TV.

2.10. Statistical Analysis

Student’s t-test was used to compare bone matrix and mineral densities from cortical and trabecular bones. Regression analysis and Blan-Altman plotting demonstrated the correlation of EBM values measured from MRI and chemical analysis. Statistical analysis was performed using GraphPad Prism (version 9.3.1 for Windows, GraphPad Software, San Diego, CA, USA). Two-tailed p-values less than 0.05 were considered significant.

3. Results

3.1. ¹H and ³¹P images with B₁ Corrections

The MR image intensities of the three-density PEG and HA phantom pellets were improved with B₁ correction, which showed linearity proportional to the weight fraction ratio 1:2:4. Figure 6 shows ¹H (a) and ³¹P (b) 3-density PEG and HA phantom images before (top) and after (bottom) B₁ corrections. After B₁ correction, PEG (¹H) and HA (³¹P) 3-density pellet image intensities show improved linearity with histograms of the pellet volumes within each ROI before and after B₁ corrections, respectively. The B₁ corrections improved the linearity of the 3-density phantom image intensities, and the histograms highlighted the B1 effect of enhanced linearity.

Figure 6.

Impact of B₁ correction on ₁H and ₃₁P images. The 3-density PEG (¹H) (a) and HA (³¹P) (b) pellet images and histograms.

3.2. Linearity Tests with Density Pellets

The 3-density PEG (¹H) and HA (³¹P) pellet MRI image intensities acquired by DTSC showed linearity with their physical densities, which confirmed the linearity proportional to the physical densities and the homogeneity of the DTSC B₁ field. The MRI-derived densities (D_CMi) of the three-density PEG and HA pellets and their physical densities (D_CPi) demonstrated significant linear relationships with a strong positive correlation (r² = 0.99) [90, 91]. In addition, the B₁ profiles of the B₁ maps obtained with homogenous PEG and HA phantoms confirmed the B1 field homogeneity of the double-tuned solenoid coil [86, 89].

3.3. Co-registration of ¹H and ³¹P Bone Images

Figure 7 shows representative rat femur ¹H images with water+fat suppression (left) and ³¹P images (right) acquired sequentially in a single imaging session at 7T. The two-color images in the middle of the bottom row are rat femur ¹H (left) and ³¹P (right) images of the same transverse slice (ROI) acquired by DTSC with 0.468mm resolution. Notably, DTSC sequentially acquired ¹H and ³¹P rat femur images with exact FOV, matrix size, and spatial resolution to co-register them precisely at any specific 3D spatial location (ROI) (Figure 7).

Figure 7.

Top row: ¹H images of rat femur with water + fat suppression (left) and ³¹P images of rat femur (right): coronal and sagittal views. Bottom row: ¹H (left) and ³¹P (right) transverse images of rat femur. The two (color) images in the middle of the bottom row are ¹H (left) and ³¹P (right) images of the same slice (ROI) acquired by the double-tuned single RF coil.

3.4. Quantitative Bone Matrix and Mineral Densities

Rat bone matrix and mineral densities were measured by MRI and chemical analysis. The DTSC facilitated the acquisition of bone matrix and mineral densities from cortical and trabecular bones, providing EBM on any 3D spatial location (ROI) (Figure 7). Bone matrix and mineral densities obtained by MRI and the gold-standard chemical analysis were similar across all femurs and ROIs (p=0.11), demonstrating good agreement with negligible (3-6%) variation. The regression analysis (R²=0.84) and the Bland-Altman plot confirmed a strong correlation between the EBMs derived from MRI and chemical analysis, showcasing DTSC efficacy (Figure 8), while demonstrating data grouping along the regression line.

Figure 8.

Regression (top) and Bland-Altman plot (bottom) of rat femur bone Extent of bone mineralization (EBM), measured by chemical analysis and MRI for CON, OVX, and Vit-D specimens.

4. Discussion

This study presents a versatile DTSC probe head for preclinical magnetic resonance imaging at 7T, designed to measure rat bone matrix and mineral densities with the same FOV and matrix size in a single session without interrupting the imaging study. Our results demonstrate the DTSC’s capabilities to provide high spatial resolution with optimum SNR, enabling simultaneous and sequential acquisition of cortical and trabecular bone matrix (¹H) and mineral (³¹P) imaging in a single session. The findings highlight the utility of this technique in advancing non-invasive and quantitative diagnosis of bone health.

The DTSC was engineered with a solenoid coil architecture with high B₁ homogeneity optimized for short-T₂ rat femur imaging. The lockable 4-digit dials integrated with variable capacitors provide precise tuning and matching for the resonance frequencies. Individual shielding traps in the ¹H and ³¹P channels avoid possible cable resonance without additional decoupling mechanisms. DTSC has excellent isolation of 30dB between the ¹H and ³¹P channels. The isolation of approximately 30 dB ensures minimal crosstalk and maximizes RF transmission efficiency, which is critical for accurately imagining short-T₂ and long-T₁ ¹H and ³¹P MR signals in hard tissue. Unlike conventional dual-coil systems, this double-tuned single-coil approach eliminates discrepancies in field-of-view (FOV) and matrix size, providing precise co-registration of multinuclear (¹H/³¹P or ¹H/²³Na) images with exact ROI (Figure 7).

Linearity tests using PEG and HA phantoms confirmed the probe head’s sensitivity and efficiency to measure mass density ratios accurately (Figure 6). MR Imaging experiments showed strong correlations between MRI-derived and chemically determined bone matrix and mineral densities, with negligible variance. This accuracy validates the DTSC for quantitative imaging and extends its potential, differentiating between metabolic bone diseases such as osteoporosis and osteomalacia.

By allowing sequential acquisition of solid-state ¹H and ³¹P MRI from the exact same volume and spatial resolution, this system facilitates precise co-registration of bone matrix and mineral signals—an essential advance for studying metabolic bone diseases. For the first time, this standardizable RF coil design supports accurate, quantitative measurements of both bone mineral and organic matrix densities from cortical and trabecular regions in the same imaging session, without repositioning the specimen or compromising signal integrity. This technological integration provides a means to compute the EBM, a composite parameter critical for distinguishing between osteomalacia and osteoporosis. In doing so, it introduces a capability to non-invasively assess the biological state of bone tissue composition. Furthermore, because the probe is engineered for optimal B₁ homogeneity, signal linearity, and field matching at both resonance frequencies, it eliminates spatial mismatches that traditionally limited dual-nucleus studies. This improves diagnostic precision in preclinical models and lays the foundation for personalized assessment of skeletal disease progression and response to treatment. The standardized and reproducible RF platform enables high-throughput studies across experimental models, extending the biological utility of solid-state MRI beyond basic research to translational and potentially clinical domains. This unified coil architecture fundamentally redefines how bone health and mineralization disorders can be noninvasively studied and characterized in vivo.

The DTSC’s capability to measure bone matrix and mineral densities from the same region of interest enhances diagnostic precision. By acquiring ¹H and ³¹P images sequentially in the same scanning session and the RF probe head without needing to change or reposition samples, the DTSC reduces scan time and improves workflow efficiency.

While DTSC demonstrated excellent performance in preclinical settings, challenges remain in adapting this technology for in vivo studies. The solenoid coil’s orientation requirements limit its use in superconducting MRI systems. Birdcage or array coils are required for the larger volumes of live animal and human imaging.

5. Conclusions

This study establishes the feasibility of solid-state MRI for quantitative bone imaging of small specimens, offering significant advancements in preclinical and potentially clinical diagnostics. By providing precise measurements of bone matrix and mineral densities, the technology represents a transformative tool for understanding bone health and addressing the global burden of metabolic bone diseases.

Abbreviations

¹H

Hydrogen or proton

³¹P

Phosphorus

2D

Two Dimensional

3D

Three Dimensional

B₀

Main Magnetic Field

B₁

Magnetic field generated by RF coil

BMD

Bone Mineral Density

C

Calibration /Cortical

⁰C

Centigrade (temperature)

CF

Conversion Factor

cm

Centimeter

CO₂

Carbone dioxide

CON

Control

CON-Cort

Control-Cortical (Bone)

CON-Trab

Control-Trabecular (Bone)

cROI

cortical Region Of Interest

D

Density

DXA

Dual-energy X-ray absorptiometry

EBM

Extent of bone mineralization

C

Chemical

HA

Hydroxyapatite

kHz

Kilohertz

MATLAB

Matrix Laboratory software program

mg

Milligram

mm

Millimeter

MRI

Magnetic Resonance Imaging

ms

milliseconds

mn

mineral

mx

matrix

nH

nano henry (capacitance)

μCT

Micro-computed tomography

μs

Microseconds

OM

Osteomalacia

OP

Osteoporosis

OVX

Ovariectomized

OVX-Cort

Ovariectomized-Cortical (Bone)

OVX-Trab

Ovariectomized-Trabecular (Bone)

PEG

Polyethylene glycol

pF

pico-farad

RF

Radio Frequency

ROI

Region of Interest

SNR

Signal-to-noise ratio

T

Tesla (magnetic field strength)/True

T₁

Spin-lattice relaxation time

T₂

Spin-Spin relaxation time

TV

Tissue Volume (bone)

VAPOR

VAriable Power RF pulses with Optimized Relaxation delays

Vit-D

Vitamin D deficient

Vit-D-Cort

Vitamin D deficient-Cortical (Bone)

Vit-D-Trab

Vitamin D deficient-Trabecular (Bone)

ZTE

Zero Time Echo/Zero Echo Time