Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Feb 19.
Published in final edited form as: NMR Biomed. 2019 Feb 22;32(5):e4080. doi: 10.1002/nbm.4080

Imaging of the Region of the Osteochondral Junction (OCJ) Using a 3D Adiabatic Inversion Recovery Prepared Ultrashort Echo Time Cones (3D IR-UTE Cones) Sequence at 3T

Ya-Jun Ma 1, Saeed Jerban 1, Michael Carl 2, Lidi Wan 1, Tan Guo 1, Hyungseok Jang 1, Graeme M Bydder 1, Eric Y Chang 3,1, Jiang Du 1
PMCID: PMC7895481  NIHMSID: NIHMS1670015  PMID: 30794338

Abstract

The purpose of this study is to develop a 3D adiabatic inversion recovery ultrashort echo time cones (IR-UTE-Cones) sequence for high resolution and contrast imaging of the region of osteochondral junction (OCJ) of human knee joint using a clinical 3T scanner. A feasibility study on direct imaging of the OCJ region was performed on a human patellar cartilage sample and on eight cadaveric knee joints using T1-weighted, proton-density (PD) weighted, and short-T2 weighted 3D IR-UTE-Cones sequences. Contrast-to-noise ratio was measured to evaluate the effectiveness of the 3D IR-UTE-Cones sequences for selective imaging of the OCJ region. CT imaging was performed in parallel for the cadaveric knee joints. The optimized T1-weighted 3D IR-UTE-Cones sequence was used to image the knee joints of eight healthy volunteers and six patients with osteoarthritis (OA) to evaluate morphological changes in the OCJ region. Clinical PD- and T2- weighted FSE sequences were also performed for comparison. The T1-weighted 3D IR-UTE-Cones sequence showed high resolution and contrast bright band of the normal OCJ region in the cadaveric joints. Normal OCJ appearances were also seen in healthy volunteers. Abnormal OCJ regions, manifested as ill-defined, focal loss or non-visualization of the high intensity band adjacent to the subchondral bone plate, were observed in the knee joints of both ex vivo and in vivo OA patients. The 3D IR-UTE-Cones sequence can image OCJ regions ex vivo and in vivo, with abnormalities depicted with high resolution and contrast. The technique may be useful for demonstrating involvement of OCJ regions in early OA.

Keywords: Osteochondral junction, calcified cartilage, adiabatic inversion recovery, 3D ultrashort echo time

INTRODUCTION

Osteoarthritis (OA) affects at least 30 million adults in the United States and has a substantial impact on the country’s healthcare system, with a total annual cost estimated at $5,700 per patient (1). OA is considered a whole joint disease with pathologic changes that often involve all of the constituent joint tissues (2). Of increasing interest is the region of the osteochondral junction (OCJ), which encompasses the tissues between the deep uncalcified layers of cartilage and the marrow spaces of the trabecular bone. It includes the deep radial uncalcified cartilage, tidemark, calcified cartilage and subchondral bone plate (3). In the strictest sense, the OCJ is the region where calcified cartilage meets subchondral bone. However, by analogy to the enthesis, which is the region where tendons or ligaments meet bone, it is useful to consider the tissues between the superficial layers of articular cartilage and the subchondral bone marrow as a whole group, which, like the enthesis, is concerned with the dissipation of stress in a region between semirigid and rigid tissues. While these tissues are avascular in the normal joint, in OA, osteoclasts are activated and form channels through the subchondral bone plate, which allow blood vessels and nerves to extend from the marrow into deep cartilage (4, 5). This is associated with a cascade of abnormalities, including local inflammation and upregulation of metalloproteinase activity, extracellular matrix degradation, reduction of cartilage load-bearing capacity and degenerative change (3, 4, 6). These changes may be relevant in the pathogenesis of OA.

MRI is widely used to directly and noninvasively evaluate articular cartilage, and plays an important role in the clinical diagnosis and treatment of OA (7, 8). MRI of the OCJ region is difficult though, due to the short mean transverse relaxation times (i.e. short T2 or T2*) of deep radial uncalcified articular cartilage, calcified cartilage and subchondral bone, which results in little or no signal when conventional pulse sequences are used (9). Ultrashort echo time (UTE) sequences have been employed to overcome this limitation (10), and have been used to evaluate the OCJ region. Studies have focused on 2D acquisition techniques (9–14) and high contrast images of the OCJ region have been achieved by combining UTE acquisitions with long T2 signal suppression schemes including dual echo acquisition and echo subtraction, single adiabatic inversion recovery (IR) and dual adiabatic inversion recovery (DIR) pulses (9–14). Those techniques can effectively suppress signals from the more superficial layers of uncalcified cartilage (e.g., the superficial, middle and upper deep layers) which have longer T2s and T2*s, as well as marrow fat, leaving the OCJ to be selectively imaged. However, 2D UTE sequences are sensitive to eddy currents, and it is difficult to image the whole knee joint due to strong out-of-slice signal excitation especially for off-center slices (15). As a result, there has been relatively little clinical work published using 2D UTE imaging techniques to evaluate the OCJ region.

Compared to 2D UTE sequences, 3D UTE sequences have the advantage of reducing through plane partial volume effects, which is important for imaging the thin (i.e. 100–200 μm) and curved OCJ region (16). Multi-echo 3D UTE techniques with echo subtraction or T2* fitting have been used for this purpose, but there have been problems with low OCJ region contrast and the need for intensive post-processing (17–19). For clinical purposes, an optimal 3D UTE sequence should provide whole joint coverage, show the OCJ region with high contrast, be obtainable in clinically acceptable scan time and require little or no post-processing.

The purpose of this study was to develop 3D adiabatic inversion recovery prepared UTE Cones (3D IR-UTE-Cones) sequences for volumetric imaging of the OCJ region with high spatial resolution and contrast in clinically acceptable scan times. A patellar sample was used to demonstrate the feasibility of OCJ region imaging with the 3D IR-UTE-Cones sequence. Following this, eight cadaveric human knee joints were used to optimize the sequence parameters for high contrast imaging of the OCJ region in clinical feasible scan times. Finally, the 3D IR-UTE-Cones sequence was used in eight healthy volunteers and six patients with OA to assess the clinical feasibility of imaging the OCJ region in vivo.

MATERIALS AND METHODS

Specimens

A patellar sample was harvested from a normal knee joint (male, 31 years old). A transverse slab of ~8 mm thickness was cut and stored in phosphate buffered saline (PBS) soaked gauze at 4°C prior to MR and μCT imaging. Eight cadaveric human knee joints were obtained from eight donors (six males and two females, age range = 27–96 years old, mean ± standard deviation of 66 ± 28 years) from tissue banks approved by our Institutional Review Board. The specimens were stored frozen at −80 °C for about two months, and then taken out to thaw at room temperature for 36 hours before MR imaging.

In Vivo Subjects

Two groups of human subjects were included in this study. The first group consisted of eight healthy volunteers with a mean ± standard deviation age of 37 ± 13 years old (four males and four females). The second group consisted of six OA patients with mean ± standard deviation age of 58 ± 14 years old (two males and four females). The OA patients were diagnosed based on plain radiography before MR experiments. Informed consent was obtained from all subjects in accordance with our Institutional Review Board guidelines before MR imaging.

Pulse Sequence

The 3D IR-UTE-Cones pulse sequence was implemented on a 3T MR750 scanner (GE Healthcare Technologies, Milwaukee, WI). The sequence combined a basic 3D UTE-Cones sequence with an adiabatic IR preparation pulse (duration = 6 ms) (20). To improve the acquisition efficiency, a series of spokes (Nsp) was acquired after each IR pulse (Figure 1A). Fat signal was suppressed with a conventional chemical shift fat saturation (FS) module before the acquisition spoke trains. The time interval between two adjacent spokes was defined as τ. For each spoke, a short rectangular pulse (e.g., 50 μs) was used for non-selective signal excitation (Figure 1B), and this was followed by a spiral trajectory data acquisition with conical view ordering (Figure 1C). Data sampling began from the center of k-space and continued outwards. It was begun as soon as was practical after the RF excitation, with a minimal nominal delay time of 32 μs. Both radiofrequency (RF) and gradient spoiling were used to crush the residual transverse magnetizations. An 8-channel transmit/receive knee coil was used both for ex vivo and in vivo whole knee joint imaging.

Figure 1.

Figure 1

Open in a new tab

The 3D IR-UTE-Cones sequence. This employs an adiabatic inversion pulse for selective signal suppression, followed by 3D UTE-Cones data acquisition (A). In the basic 3D UTE-Cones sequence a short rectangular pulse is used for signal excitation and this is followed by 3D spiral sampling with a minimal nominal TE of 32 μs (B). The spiral trajectories are arranged with conical view ordering (C). To speed up data acquisition, multiple spokes can be sampled after each long T2 preparation (A). A fat saturation module is also used in front of the data acquisition spokes.

Contrast Mechanism

Figure 2A depicts the regions of different cartilage layers (i.e. superficial, middle, deep and calcified layers), subchondral bone and marrow fat in the patellar cartilage. Figure 2B illustrates the contrast mechanism for imaging of the OCJ region using the 3D IR-UTE-Cones sequence. Initially, the magnetizations of the more superficial (uncalcified) layers of articular cartilage including the superficial, middle and deep layers are inverted by the adiabatic IR pulse. After this the inverted magnetizations of these layers recover at rates based on their T1 relaxation mechanisms. Since uncalcified cartilage T1 values decrease from the superficial to deep layers (21), the longitudinal magnetization in the deeper zone recovers faster than those in the middle and superficial layers. As a result, the signal intensity is higher for deep uncalcified cartilage when 3D UTE data acquisitions start near the signal nulling point for the superficial layer cartilage. Tissues of the OCJ region have very short T2 components (much shorter than the duration of the adiabatic IR pulse), and their magnetizations are not inverted by the adiabatic IR pulse and are largely saturated (22). In addition, the OCJ region tissues have shorter T1s than the more superficial uncalcified layers of cartilage. As a result, the saturated magnetizations of tissues of the OCJ region recover faster than those of the superficial layers of articular cartilage. This leads to high signal intensity and high contrast for tissues of the OCJ region when the inversion time is close to the nulling point of the superficial layers of articular cartilage.

Figure 2.

Figure 2

Open in a new tab

The contrast mechanism for the OCJ region imaging using the 3D IR-UTE-Cones sequence. Panel A depicts the regions of different cartilage layers (i.e. superficial, middle, deep and calcified layers), subchondral bone and marrow fat in a patellar cartilage. The magnetizations of superficial, middle and deep layer uncalcified cartilage are inverted by the adiabatic IR pulse (B), while calcified layer cartilage and subchondral bone magnetizations are mostly saturated due to their short T2s. Thus, both the calcified layer cartilage and subchondral bone longitudinal magnetizations recover rapidly from zero after saturation. When TI is selected to suppress superficial zone cartilage, the calcified layers express a higher signal than the more superficial layer cartilages. Subchondral bone remains low signal due to its low proton density.

Feasibility Study on a Patellar Sample

The feasibility of OCJ region imaging with the 3D IR-UTE-Cones sequence was assessed on a patellar sample from a 31y male donor. To obtain high spatial resolution and signal to noise ratio (SNR) images of the sample, a small birdcage coil which fits in a 30 ml syringe was used for both RF transmission and reception. First, the T1 variation in different cartilage layers and subchondral bone was investigated. A recently reported 3D UTE-Cones actual flip angle imaging and variable flip angle (3D UTE-Cones AFI-VFA) sequence was used to measure the T1 of tissues in the patellar cartilage sample (23, 24). The sequence parameters were listed in Table 1. Then, three additional UTE protocols were evaluated for T1-weighted, proton density (PD) weighted and selective short-T2 imaging of the OCJ region. The T1-weighted 3D IR-UTE-Cones sequence was used to highlight the deep radial and calcified cartilage by selecting a TI to invert and suppress signals from the more superficial layers of articular cartilage. PD-weighted images were generated with the same IR-UTE-Cones sequence but with a much longer TI. Selective short-T2 imaging was performed with a short TR and short TI so that very short T2 signals were highlighted, with nearly complete suppression of longer T2 signals (20). The IR-UTE-Cones sequence parameters were: 1) T1-weighted: field of view (FOV) = 6×6×3 cm3, acquisition matrix = 384×384×100, voxel size = 156×156×300 μm3, receiver bandwidth (BW) = 125 kHz, TR/TI = 1200/450 ms, flip angle = 10°, Nsp = 17, τ = 8.3 ms and total scan time = 1h 11 min; 2) PD-weighted: same parameters as above T1-weighted UTE sequence except a longer TI = 1000 ms; 3) selective short-T2 imaging: FOV = 6×6×2.4 cm3, acquisition matrix = 256×256×40, voxel size = 234×234×600 μm3, BW = 125 kHz, TR/TI = 133/58 ms, flip angle = 20°, Nsp = 5, τ = 8.3 ms, nex = 2 and total scan time = 1h 16 min. For comparison, μCT was also performed with 18×18×18 μm3 resolution. To measure T2* of both calcified cartilage and subchondral bone, the same short T2 selective imaging sequence was used but with four longer TEs (i.e. 0.2, 0.4, 0.8 and 2.2 ms).

Table 1.

Sequence parameters for patellar cartilage sample, ex vivo knee joint specimen and in vivo knee joint studies.

3D UTE-Cones AFI (patellar cartilage sample) 3D IR-UTE-Cones (Ex vivo knee joint) 3D IR-UTE-Cones (In vivo knee joint)
FOV = 6×6×4 cm3,
Matrix = 360×360×100,
voxel size = 167×167×400μm3,
TR1/TR2 = 20/100 ms, flip angle = 45°, BW = 125 kHz, scan time = 1h 32 min		 FOV = 13×13×8 cm3,
Matrix = 256×256×40,
voxel size = 508×508×2000μm3,
TR/TI = 1200/540 ms, flip angle = 10°, Nsp = 9, τ = 9 ms, BW = 250 kHz, oversample factor = 1.2, scan time = 56 min		 FOV = 13×13×8 cm3,
Matrix = 256×256×40,
voxel size = 508×508×2000μm3,
TR/TI = 1200/600 ms, flip angle = 10°, Nsp = 21, τ = 5.2 ms, BW = 166 kHz, scan time = 10 min
3D UTE-Cones VFA (patellar cartilage sample) Clinical PD weighted FSE (Ex and in vivo knee joint) Clinical T2 weighted FSE (Ex and in vivo knee joint)
FOV = 6×6×4 cm3,
Matrix = 360×360×100,
voxel size = 167×167×400 μm3,
TR = 20 ms, flip angle = 2°, 4°, 6°, 8°, 10°, 15°, 20°, 25°, 30° and 35°, BW = 125 kHz, total scan time = 5 h 13 min		 FOV = 15 × 15 cm2,
Matrix = 352 × 256,
pixel size = 426×586 μm2,
TR = 3220 ms, TE = 27.8 ms, number of slices = 40, acceleration factor = 2, scan time = 2 min 30 sec		 FOV = 15 × 15 cm2,
Matrix = 352 × 256,
pixel size = 426×586 μm2,
TR = 7585 ms, TE = 71.5 ms, number of slices = 40, acceleration factor = 2, scan time = 2 min 32 sec
Open in a new tab

Ex vivo Study of Knee Joint Specimens

The observed OCJ contrast strongly depends on the combination of TR and TI. To optimize OCJ contrast, the ex vivo knee joint specimens were scanned with the 3D IR-UTE-Cones sequence using a series of TIs. The sequence parameters were: TR = 1200 ms, TI = 220, 260, 300, 340, 380, 420, 460, 500, 540, 560, 580, 600, 650, 700 and 800 ms, flip angle = 10°, FOV = 13×13×8.4 cm3, acquisition matrix = 256×256×28, voxel size = 508×508×3000 μm3, BW = 250 kHz, Nsp = 15, τ = 6 ms with scan time for each TI = 9 min 8 sec. The TIs with a higher contrast to noise ratio (CNR) between OCJ region and superficial cartilage, and between OCJ region and marrow fat were used.

A high resolution and high SNR IR-UTE-Cones sequence was used to image the OCJ region in all the cadaveric knee specimens. Its sequence parameters were listed in Table 1 and the total scan time was 56 min. To facilitate translational imaging of OCJ in vivo, a faster protocol with a clinically acceptable scan time (i.e. 10 min) was developed. Most of the sequence parameters were the same except for a higher Nsp of 21, a smaller τ of 5.2 ms and no oversampling. In addition, both clinical PD- and T2- weighted FSE sequences were used for comparison (sequence parameters shown in Table 1). Clinical CT scans were also performed.

In Vivo Study of Healthy Volunteers and Patients with OA

Since in vivo cartilage T1s at body temperature are typically longer than the T1s of cartilage specimen examined at room temperature, higher TIs of 500, 550, 600, 650, and 700 ms were tested on one healthy volunteer (male, 28 years old) to determine parameters for optimizing OCJ region contrast. The optimal TI was used for subsequent imaging of the OCJ region in healthy volunteers and patients with OA. Other sequence parameters were shown in Table 1. Clinical PD- and T2- weighted FSE sequences were performed for comparison with parameters the same as for the ex vivo knee joint study.

Image Analysis

The mean signal intensity inside a user drawn region of interest (ROI) within the OCJ region was used for signal calculation, with noise being calculated as the standard deviation of the signal in ROI placed in artifact free image background. The CNR between tissues of the OCJ region and the superficial cartilage (CNROCJ_SC) and marrow fat (CNROCJ_MF) were calculated as the respective signal differences divided by the background noise. T1 and T2* were also measured using previously reported methods (21, 23, 24). Rigid body registration was performed between μCT and UTE images for the patellar sample study. The Image co-registration was performed on a two dimensional basis via following steps: 1) The corresponding slices of the patella from the T1-weighted IR-UTE-Cones and μCT datasets were selected; 2) A set of four identical points were manually selected on the outer contour of the patella (surface of the cartilage and cut edges of the trabecular bone) on both μCT and MRI images; 3) Affine registration transform (linear mapping) was used to perform image registration by matching the four selected control points in the μCT and MRI images employing the Image processing toolbox in the MATLAB software; 4) The obtained transformation matrix was used to register μCT and all IR-UTE-Cones images. All analysis algorithms were written in MATLAB. The mean intensity within each of the ROIs was used for subsequent curve fitting.

RESULTS

Figure 3 shows high resolution 3D UTE-Cones images which clearly depict all the components of the patellar cartilage, including the superficial, middle and deep uncalcified layers, calcified cartilage and subchondral bone plate (dark band beneath the calcified cartilage and above marrow fat). The fitting curves and corresponding T1 values of different regions of cartilage and subchondral bone are also shown. As expected, the T1 values of uncalcified cartilage decrease from the superficial layer (963±23 ms) to middle layer (937±14 ms) and then deep layer (741±12 ms). The calcified cartilage has a much lower T1 of 405±16 ms, and the subchondral bone plate has an even shorter T1 of 285±12 ms. The significantly lower T1 values of the calcified cartilage make it possible to highlight the OCJ region with an IR sequence with an appropriate TI. In addition, the T2*s of the calcified cartilage and subchondral bone plate in the short T2 IR-UTE-Cones imaging were measured. Both tissues have extremely short T2* values (0.42±0.01 ms and 0.31±0.03 ms respectively, supplemental Figure 1).

Figure 3.

Figure 3

Open in a new tab

T1 measurements for cartilage and subchondral bone using a 3D UTE-AFI-VFA method for the patellar cartilage sample. The region of interests (i.e. red rectangles) are shown from the cartilage surface to the subchondral bone (superficial zone: A, middle zone: C, deep zone: E, calcified zone: G and subchondral bone: I). The corresponding T1 fitting curves and values are shown in B, D, F, H and J.

Figure 4 shows a comparison between UTE MRI and μCT images of the same patellar cartilage sample. UTE T1-weighted, PD-weighted and short-T2 images of the specimen are shown. ROIs including the high intensity band shown with the T1-weighted 3D IR-UTE-Cones sequence (Figure 4B) were located on the μCT and other UTE images. The high intensity band which lies beneath the superficial layers of cartilage was confirmed to be immediately above the subchondral bone plate on the μCT images (Figure 4A). It shows as high contrast with the T1-weighted UTE sequence but as low contrast with the PD-weighted UTE sequence (Figure 4C). The subchondral bone plate only shows high signal with the selective short-T2 UTE imaging sequence (Figure 4D), in which the signal from the superficial, middle and deep layers of uncalcified articular cartilage are suppressed. The deep radial and calcified cartilage (shown in the left half part in the red rectangle) were still visible in the short T2 image. Those results together suggest that the high intensity band in T1-weighted 3D IR-UTE-Cones imaging mainly comes from the calcified cartilage, with some contribution from the deep radial (uncalcified) layer, but no significant contribution from the subchondral bone plate due to its low proton density.

Figure 4.

Figure 4

Open in a new tab

Co-registered high-resolution μCT (A) and UTE-MR (B, C and D) images of the patellar cartilage sample. High contrast demonstration of the OCJ region is shown in B acquired with the T1-weighted IR-UTE-Cones sequence (TR/TI=1200/450ms). C shows an approximately PD weighted image acquired with a much longer TI using the IR-UTE-Cones sequence (TR/TI=1200/1000ms). D shows the pure short T2 components in the cartilage and bone acquired with a very short TR using the IR-UTE-Cones sequence (TR/TI=133/58ms). The same regions (i.e. in red rectangles) including part of deep layer, calcified layer and subchondral bone were selected for all images.

The OCJ region contrast depends on the choice of TI as demonstrated by Figure 5. The 3D IR-UTE-Cones images with TIs from 460 to 600 ms all show high contrast imaging of the OCJ region. CNROCJ_SC increased from −14.8±5.2 with a TI of 220 ms, to 20.1±7.4 with a TI of 540 ms, and then decreased to 15.7±6.6 with a TI of 800 ms. CNROCJ_MF increased from −10.0±3.0 with a TI of 220 ms, to 18.8±6.1 with a TI of 540 ms, and further increased to 25.0±9.6 with a TI of 800 ms. An optimal TI of 540 ms with a maximal CNROCJ_SC and good CNROCJ_MF was used for subsequent 3D IR-UTE-Cones imaging of the knee joint specimens. The OCJ contrast was well preserved in the highly accelerated acquisition with a Nsp of 21 (Supplemental Figure 2) with the total scan time reduced from 56 min to 10 min. Those results support the clinical feasibility of evaluating the OCJ in the whole knee joint in vivo.

Figure 5.

Figure 5

Open in a new tab

IR-UTE-Cones ex vivo knee images (from an 84 year-old male donor) with different TIs. The TR is fixed at 1.2s and TI varies from 0.22s to 0.8s. The contrast of the tissues in the OCJ region varies with the increasing TIs.

The OCJ regions in a normal and abnormal knee joint are displayed in Figures 6 and 7, respectively. The deep layers of articular cartilage, menisci, ligaments and tendons appear with little or no signal with conventional PD- and T2- weighted FSE sequences. The 3D FS-UTE-Cones sequence shows high signal for all the components of articular cartilage, menisci, ligaments and tendons (but not the subchondral bone plate). There is little contrast between the OCJ region and the more superficial layers of articular cartilage, which all show high signal. However, the 3D IR-UTE-Cones sequence shows well-defined bright band in the patellar, femoral and tibia cartilage of the knee joints. The abnormal knee joint specimen shows morphological changes in the OCJ region, including regions with ill-defined or focal absence of bright band adjacent to and paralleling the subchondral bone (arrows). These corresponded with abnormal regions on the CT images.

Figure 6.

Figure 6

Open in a new tab

OCJ region imaging of a normal ex vivo knee joint specimen from a 31-year-old donor (male). The clinical images (PD weighted FSE in first column and T2 weighted FSE in second column) are used for comparison with the T1-weighted IR-UTE-Cones images (third column, TR/TI=1200/540ms). The IR-UTE-Cones images show high OCJ contrasts (i.e. bright band), which can be seen more clearly in the zoomed images. The last column includes the conventional fat-saturated UTE-Cones images for comparison. These show signal from both uncalcified and calcified cartilage.

Figure 7.

Figure 7

Open in a new tab

Comparison of CT and MR imaging of an abnormal ex vivo knee specimen from a 57-year-old donor (male). T1-weighted IR-UTE-Cones images show darker and discontinuous regions of calcified cartilage bands, which on CT images correspond to an osteophyte at the posterior-most portion of the medial femoral condyle (top insert), small foci of calcium at the weight-bearing portion of the medial femoral condyle (middle insert) and eburnation at the patella (bottom insert).

High contrast imaging of the OCJ region in vivo can be obtained within a range of TIs (i.e. from 550 to 650 ms), with optimum CNR achieved with a TI of ~600 ms. Figure 8 shows OCJ region imaging of a normal knee joint. Continuous bright bands can be found in patellar, femoral and tibial cartilages, similar to those observed in the normal knee specimen (Figure 6). T1-weighted 3D IR-UTE-Cones imaging of four OA patients are shown in Figure 8. As with the OA knee specimen, morphological changes were observed in the OCJ region, including ill-defined focal loss and non-visualization of the high intensity band adjacent to the subchondral bone plate (Figure 9). The abnormal OCJ regions shown with the T1-weighted 3D IR-UTE-Cones imaging corresponded well with the abnormal superficial layers seen in the clinical PD- and T2-weighted FSE imaging.

Figure 8.

Figure 8

Open in a new tab

The OCJ region imaging of an in vivo knee from a 36-year-old normal volunteer (male). The clinical images (PD weighted FSE in first column and T2 weighted FSE in second column) are used for comparison with the T1-weighted IR-UTE-Cones images (third column, TR/TI=1200/600ms). The IR-UTE-Cones images show high OCJ contrasts (i.e. bright band), which can be seen more clearly on the zoomed in images.

Figure 9.

Figure 9

Open in a new tab

The OCJ imaging of four abnormal in vivo knees from four patients (ages of 54 (female), 49 (male), 54 (male) and 63 (female) years old). The clinical images (PD weighted FSE in first column and T2 weighted FSE in second column) are used for comparison with the T1-weighted IR-UTE-Cones images (third column, TR/TI=1200/600ms). The arrows show abnormal regions (zoomed in) on the images with reduction or loss of high intensity band in OCJ regions.

DICUSSION

In this study, we have demonstrated that 3D IR-UTE-Cones sequences can provide high contrast imaging of the OCJ ex vivo and in vivo. The 3D UTE-Cones sequence has a TE over 100 times shorter than that of conventional clinical sequences, and can directly detect signal from short T2 tissues or tissue components in the OCJ region including the calcified cartilage and subchondral bone. The major challenge is the high signal from the surrounding, more superficial layers of articular cartilage and marrow fat. Adiabatic IR preparation pulses provide a very effective method for suppressing signal from longer T2 cartilage, enabling direct demonstration of the OCJ region with high contrast. The much shorter T1 of the OCJ than that of the more superficial layers of articular cartilage further enables high resolution imaging of the OCJ when a highly T1-weighed 3D IR-UTE-Cones sequence is used. Joint fluid with a much longer T1 than both cartilage and marrow is almost completely nulled. The multi-spoke acquisition scheme provides very efficient method of reducing the total scan time without significant compromise in contrast (supplemental Figure 2), making it clinically feasible to image the OCJ region in vivo.

The signal sources for 3D IR-UTE-Cones imaging of the OCJ region largely include the calcified cartilage and part of the deep radial layers of articular cartilage, with little contribution from the subchondral bone plate due to its low proton density. The superficial and middle layers as well as part of the radial layer of articular cartilage have relatively long T2s and T1s, and can be well suppressed by adiabatic inversion pulses. The deep radial cartilage next to the calcified cartilage has much a lower T1 and T2, and is only partially suppressed by the adiabatic inversion pulse. Signal from the deep radial cartilage is lower than that from the calcified cartilage, partly because the calcified cartilage has a shorter T2* (thus less inversion and signal attenuation by the adiabatic inversion pulse), and a shorter T1 (thus faster signal recovery after the adiabatic inversion pulse). The subchondral bone plate shows as low signal with T1- and PD-weighted 3D IR-UTE-Cones imaging because of its low proton density, which is in the order of ~30% water by volume (25). This is much lower than the ~80% water by volume for articular cartilage. The bright band in the T1-weighed IR-UTE-Cones image is much wider than the typical calcified cartilage layer which may be caused by the short T2 blurring, partial volume effect (relatively thick slice of 2 mm and a curved surface) and contribution from deep radial layers which also have a short T1.

Imaging of the OCJ region is further complicated by the fact that articular cartilage has two forms of water, free water residing in the organic matrix and water bound to the macromolecules. The existence of bound and free water pools has also been confirmed by the cross-relaxation imaging experiments (26), results from recent MR spectroscopy studies (27), and our own studies (28). Our prior studies suggest that bound water fraction gradually increases from the articular surface (~15%) to the deep cartilage (~20%). Similar to the calcified cartilage, the bound water magnetizations are largely saturated by the adiabatic inversion pulse because of their short T2s, recovered during the TI time, and detected by UTE data acquisition. The spatial distribution of bound water also contributes to a spatial gradient of signal intensity from the superficial cartilage to the calcified cartilage. However, the short T2 signal from the calcified cartilage dominates in T1-weighted 3D IR-UTE-Cones imaging, as shown in Figure 4B.

The bright band region in the T1-weighted IR-UTE-Cones image (e.g. the left half part of the red rectangle in Figure 4B) was largely dark in the μCT image (i.e. Figure 4A). This means that there are no significant calcifications in the majority of the bright band. The calcified layer should appear bright in μCT. However, it is very thin and is tightly attached to or overlapped with the subchondral bone. A recent study concluded that μCT with a 2.8 μm isotropic resolution still could not separate calcified cartilage and subchondral bone due to their overlap (29) Moreover, the IR-UTE-Cones imaging suffered short T2 blurring, off-resonance artifacts and partial volume effect in this study. Thus, it is difficult to get perfect co-registration for MRI and μCT images to isolate the extremely thin calcified layers. A high-performance gradient system is necessary for the further verification.

The OCJ serves as a functional barrier between the synovial joint space and subchondral bone marrow, and may play an important role in the pathogenesis of OA (30–33). Although it is generally thought that the initial changes of OA occur in uncalcified cartilage with degeneration and erosion starting in superficial uncalcified cartilage (7, 8), there are a number of recent studies which are not explained by this model (4–6, 30–32). Calcified cartilage is a highly modified mineralized region with 10–100 times greater stiffness than uncalcified cartilage but an order of magnitude less stiffness than the subchondral bone plate. It functions as a transition zone of intermediate stiffness between uncalcified cartilage and subchondral bone (30). In OA, calcified cartilage may become reactive and extend to involve adjacent uncalcified cartilage, leading to thinning of the superficial uncalcified layers of cartilage (31). In a study of OA human femoral heads, the calcified cartilage layer was extremely hypermineralized and twice as hard as neighboring subchondral bone plate (32). In this situation, it may function as a hard grinding abrasive and accelerate wear rates. Changes in the calcified cartilage can compromise the uncalcified portion and cause it to degenerate (33–34). The hypermineralized calcified cartilage may have reduced T2* and T1 relaxation times; however, its water content may also be reduced (more mineral, less water). Therefore, the 3D IR-UTE signal intensity of hypermineralized calcified cartilage may be higher (if T1 shorting dominates) or lower (if water content reduction dominates) than that of the normal calcified cartilage. In this study, the signal loss of the bright band in the OA patients may be also caused by the degeneration of the cartilage (e.g. more free water and longer T1). Thus the OCJ region appeared dark in the T1-weighted IR-UTE-Cones sequence.

The OCJ region may play an important role in the initiation and/or progression of OA. MR imaging of the OCJ region may therefore be of critical importance in elucidating the structural and functional pathogenesis of OA including features associated with the internal layers of cartilage without loss of superficial layers. Clearly, more research is needed to investigate the morphology of normal vs. abnormal deep radial/calcified cartilage (including normal mineralized and hypermineralized calcified cartilage and degeneration of cartilage). Also, quantitative evaluation of the calcified cartilage, including measurement of MR relaxation times (T1, T2, T2* and T1ρ (35)) and proton density, is full of interest and will be investigated in future studies.

There are several limitations of this study. First, histology was not performed to confirm the signal source of the 3D IR-UTE-Cones imaging. However, μCT images supported the view that the high intensity band comes from the deep radial and calcified cartilage without contribution from the subchondral bone. Second, only a small number of knee joint specimens was studied. Abnormalities in the deep radial and calcified cartilage were observed, but without systematic investigation and especially correlation with histology. The 3D IR-UTE-Cones sequence may help with more accurate diagnosis of early OA by exploiting a completely different approach from the conventional MR sequences. The deep radial and calcified cartilage are involved in OA but have never been investigated with clinical MRI. Third, only a small number of healthy volunteers and OA patients was studied. The clinical significance of this technique remains to be further investigated. However, well preserved high intensity bands were observed in the knee joint of all healthy volunteers, while abnormalities in deep radial and calcified cartilage were only depicted in OA patients. These preliminary results clearly demonstrate the clinical potential of the 3D IR-UTE-Cones sequence for evaluating the involvement of deep radial and calcified cartilage in degeneration. Fourth, the scan time of 10 min is still relatively long for clinical scan. Fast acquisition with accelerated techniques such as parallel imaging or compressed sensing may be used in future studies (36). Fifth, both the proton density of the tissues in OCJ region and the distance between the bright bands in femur condyle and tibia may provide us more information to identify the abnormality related to different OA stages. Other quantitative biomarkers (such as T1, T2*, T2 and T1ρ) may also be included for more comprehensive evaluation. A grading system involving both morphological and quantitative evaluation of the OCJ region remains to be developed.

In conclusion, we have shown that the 3D IR-UTE-Cones sequence can be used to image the OCJ region in vivo. Abnormalities in the OCJ region (mainly the deep radial and calcified cartilage) can be depicted with high spatial resolution and high contrast. This technique could be used for larger scale clinical evaluation of the involvement of OCJ region in OA, especially in the early stages.

Supplementary Material

Supporting information

Supplemental Figure 1 Short T2 imaging and quantitative T2* measurement of calcified cartilage and subchondral bone. A high contrast OCJ region image (A) with the T1-weighted IR-UTE-Cones (TR/TI=1200/450ms) shown was used to define the position of the calcified cartilage and subchondral bone. Short T2 images (TE = 0.032, 0.2, 0.4, 0.8 and 2.2 ms) are shown in B to F. These were acquired with a short TR IR-UTE-Cones (TR/TI=133/58ms). The T2* of calcified cartilage (red line in G) and subchondral bone (red line in I) were 0.42 ± 0.01 ms (H) and 0.31 ± 0.03 ms (J), respectively. The bright line in E (arrows) is the same calcified cartilage which is shown in red in G.

Supplemental Figure 2 OCJ region imaging comparison between a long scan (56min) and a clinical feasible scan (10min) in a knee specimen study. The image quality of the 10min scan is comparable with that of the longer scan time image.

ACKNOWLEDGEMENTS

The authors acknowledge grant support from GE Healthcare, NIH (1R01 AR062581 and 1R01 AR068987) and the VA Clinical Science and Rehabilitation Research & Development Services (1I01CX001388 and I01RX002604).

Abbreviations used:

OA

osteoarthritis

OCJ

osteochondral junction

UTE

ultrashort echo time

IR

inversion recovery

DIR

dual adiabatic inversion recovery

3D IR-UTE-Cone

3D adiabatic inversion recovery prepared UTE Cones

PBS

phosphate buffered saline

FS

fat saturation

SNR

signal to noise ratio

AFI-VFA

actual flip angle imaging and variable flip angle

PD

proton density

FOV

field of view

BW

receiver bandwidth

CNR

contrast to noise ratio

REFERENCES

  • 1.Lawrence RC, Felson DT, Helmick CG, et al. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis Rheum. 2008;58(1):26–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Loeser RF, Goldring SR, Scanzello CR, Goldring MB. Osteoarthritis: a disease of the joint as an organ. Arthritis Rheum. 2012;64(6):1697–707. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Mapp PI, Walsh DA. Mechanisms and targets of angiogenesis and nerve growth in osteoarthritis. Nat Rev Rheumatol. 2012;8(7):390–8. [DOI] [PubMed] [Google Scholar]
  • 4.Walsh DA, McWilliams DF, Turley MJ, et al. Angiogenesis and nerve growth factor at the osteochondral junction in rheumatoid arthritis and osteoarthritis. Rheumatology. 2010;49(10):1852–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Suri S, Gill SE, Massena de Camin S, Wilson D, McWilliams DF, Walsh DA. Neurovascular invasion at the osteochondral junction and in osteophytes in osteoarthritis. Ann Rheum Dis. 2007;66(11):1423–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Findlay DM, Atkins GJ. Osteoblast-chondrocyte interactions in osteoarthritis. Curr Osteoporos Rep. 2014;12(1):127–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Choi JA, Gold GE. MR imaging of articular cartilage physiology. Magn Reson Imaging Clin N Am. 2011;19(2):249–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Chang EY, Ma Y, Du J. MR Parametric Mapping as a Biomarker of Early Joint Degeneration. Sports Health. 2016;8(5):405–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Gold GE, Thedens DR, Pauly JM, et al. MR imaging of articular cartilage of the knee: new methods using ultrashort TEs. AJR Am J Roentgenol. 1998;170(5):1223–6. [DOI] [PubMed] [Google Scholar]
  • 10.Chang EY, Du J, Chung CB. UTE imaging in the musculoskeletal system. J Magn Reson Imaging. 2015;41(4):870–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Bae WC, Dwek JR, Znamirowski R, et al. Ultrashort echo time MR imaging of osteochondral junction of the knee at 3 T: identification of anatomic structures contributing to signal intensity. Radiology. 2010;254(3):837–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Du J, Takahashi A, Bae WC, Chung CB, Bydder GM. Dual inversion recovery ultrashort echo time imaging: Creating high contrast for short T2 species. Magn Reson Med 2010; 63:447–455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Du J, Carl M, Bae WC, Statum S, Chang EY, Bydder GM, Chung CB. Dual inversion recovery ultrashort echo time (DIR-UTE) imaging and quantification of the zone of calcified cartilage. Osteoarthritis Cartilage 2013; 21:77–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Mackay JW, Low SB, Houston GC, Toms AP. Ultrashort TE evaluation of the osteochondral junction in vivo: a feasibility study. Br J Radiol. 2016;89(1059):20150493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Lu A, Daniel BL, Pauly JM, Pauly KB. Improved slice selection for R2* mapping during cryoablation with eddy current compensation. J Magn Reson Imaging 2008; 28:190–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Lane LB, Bullough PG. Age-related changes in the thickness of the calcified zone and the number of tidemarks in adult human articular cartilage. J Bone Joint Surg Br. 1980; 62:372–375. [DOI] [PubMed] [Google Scholar]
  • 17.Du J, Bydder M, Takahashi AM, M Carl, Chung CB, Bydder GM. Short T2 contrast with three-dimensional ultrashort echo time imaging. Magn Reson Imaging 2011; 29:470–482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Van Dyck P, Vanhevel F, Vanhoenacker FM, et al. Morphological MR imaging of the articular cartilage of the knee at 3 T-comparison of standard and novel 3D sequences. Insights Imaging. 2015;6(3):285–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Williams AA, Titchenal MR, Do BH, Guha A, Chu CR. MRI UTE-T2* Shows High Incidence of Cartilage Subsurface Matrix Changes 2 Years After ACL Reconstruction. J Orthop Res. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Carl M, Bydder GM, Du J. UTE imaging with simultaneous water and fat signal suppression using a time-efficient multi-spoke inversion recovery pulse sequence. Magn Reson Med 2016; 76:577–582. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Zanganeh S, Shao H, Bydder GM, Du J. Ultrashort echo time (UTE) Imaging of Articular Cartilage, Chapter 11 for the book “Biophysics and Biochemistry of Cartilage by NMR and MRI” Edited by Yang Zia and Konstantin Mornot. Royal Society of Chemistry, 2015. [Google Scholar]
  • 22.Larson PE, Conolly SM, Pauly JM, Nishimura DG. Using adiabatic inversion pulses for long-T2 suppression in ultrashort echo time (UTE) imaging. Magn Reson Med 2007; 58:952–961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Ma YJ, Lu X, Carl M, Zhu Y, Szeverenyi N, Bydder GM, Chang E, Du J. Accurate T1 mapping of short T2 tissues using a three-dimensional ultrashort echo time cones actual flip angle – variable TR (3D UTE-Cones AFI-VTR) method. Magn Reson Med 2018; 80:598–608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Ma YJ, Zhao W, Wan L, Guo T, Searleman A, Jang H, Chang EY, Du J. Whole knee joint T1 values measured in vivo at 3T by combined 3D ultrashort echo time cones actual flip angle and variable flip angle methods. Magn Reson Med 2018; doi: 10.1002/mrm.27510. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Manhard MK, Horch RA, Gochberg DF, Nyman JS, Does MD. In vivo quantitative MR imaging of bound and pore water in cortical bone. Radiology 2015; 25;277(1):221–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Stikov N, Keenan KE, Pauly JM, Smith RL, Dougherty RF, Gold GE. Cross-relaxation imaging of human articular cartilage. Magn Reson Med 2011; 66:725–734. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Reiter DA, Lin PC, Fishbein KW, Spencer RG. Multicomponent T2 relaxation analysis in cartilage. Magn Reson Med 2009; 61:803–809. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Pauli C, Bae WC, Lee M, Lotz M, Bydder GM, DLima D, Chung CB, Du J. Ultrashort echo time (UTE) magnetic resonance imaging of the patella with bi-component analysis: correlation with histopathology and polarized light microscopy. Radiology 2012; 264: 484–493. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Kauppinen S, Karhula SS, Thevenot J, Ylitalo T, Rieppo L, Kestilä I, Haapea M, Hadjab I, Finnilä MA, Quenneville E, Garon M. 3D morphometric analysis of calcified cartilage properties using micro-computed tomography. Osteoarthritis and cartilage 2019;27(1):172–80. [DOI] [PubMed] [Google Scholar]
  • 30.Mente PL, Lewis JL. Elastic modulus of calcified cartilage is an order of magnitude less than that of subchondral bone. J Orthopaedic Research 1994; 12:637–647. [DOI] [PubMed] [Google Scholar]
  • 31.Anderson DD, Brown TD, Radin EL. The influence of basal cartilage calcification on dynamic juxtaarticular stress transmission. Basal Cartilage Calcification 1993; 286:298–307. [PubMed] [Google Scholar]
  • 32.Ferguson VL, Bushby AJ, Boyde A. Nanomechanical properties and mineral concentration in articular calcified cartilage and subchondral bone. J Anat 2003; 203:191–199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Oegema TR, Carpenter RJ, Hofmeister F, Thompson RC. The interaction of the zone of calcified cartilage and subchondral bone in osteoarthritis. Microsc Res Tech 1997; 37:324–332. [DOI] [PubMed] [Google Scholar]
  • 34.Revell PA, Pirie C, Amir G, Rashad, Walker F Metabolic activity in the calcified zone of cartilage: observations on tetracycline labeled articular cartilage in human osteoarthritic hips. Rheumatol Int 1990; 10:143–147. [DOI] [PubMed] [Google Scholar]
  • 35.Ma YJ, Carl M, Shao H, Tadros AS, Chang EY, Du J. Three‐dimensional ultrashort echo time cones T1ρ (3D UTE-cones-T1ρ) imaging. NMR in Biomed 2017;30(6):e3709. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Lustig M, Donoho D, Pauly JM. Sparse MRI: The application of compressed sensing for rapid MR imaging. Magn Reson Med 2007;58:1182–1195. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting information

Supplemental Figure 1 Short T2 imaging and quantitative T2* measurement of calcified cartilage and subchondral bone. A high contrast OCJ region image (A) with the T1-weighted IR-UTE-Cones (TR/TI=1200/450ms) shown was used to define the position of the calcified cartilage and subchondral bone. Short T2 images (TE = 0.032, 0.2, 0.4, 0.8 and 2.2 ms) are shown in B to F. These were acquired with a short TR IR-UTE-Cones (TR/TI=133/58ms). The T2* of calcified cartilage (red line in G) and subchondral bone (red line in I) were 0.42 ± 0.01 ms (H) and 0.31 ± 0.03 ms (J), respectively. The bright line in E (arrows) is the same calcified cartilage which is shown in red in G.

Supplemental Figure 2 OCJ region imaging comparison between a long scan (56min) and a clinical feasible scan (10min) in a knee specimen study. The image quality of the 10min scan is comparable with that of the longer scan time image.

NIHMS1670015-supplement-Supporting_information.docx (1.2MB, docx)

RESOURCES